Energy storage has become this year’s cause célèbre.
The main selling point for storage, forged into a meme by story after story in the press, is that storage is needed to compensate for the variable output of wind and solar. “Today, the power grid isn’t able to easily handle the rapid fluctuations in the production of wind and solar power,” is just a random sample quote.
But this claim is not true now, and will not be true for quite some time. The electricity grid is by far the most cost-effective and reliable way to deal with the variability of wind and solar, just like it deals with the variability of demand.
Storage may have a business case, but backing up wind and solar is not it for now. And a number of recent studies are raising uncomfortable questions.
To understand why, we need to start with a few fundamental facts about the grid.
First, everything backs up everything else. There is no need to have one thing dedicated to backing up one other thing (e.g., battery to solar panel, gas turbine to wind turbine) because every generator and every source of load is connected.
Second, and related, the electricity network takes advantage of the awesome power of statistics. The bigger the sample size, the less likely variations will correlate with each other, meaning their variability will be smoothed out. Also, a bigger pool reduces the consequences of any one thing failing. A more interconnected grid is itself a solution to variable generation.
Third, from the perspective of the grid operator, supply and demand are increasingly interchangeable. The grid operator’s task is to match supply and demand at every microsecond. Controllable demand is just as effective as a dispatchable power plant in this regard. In other words, storage will have to compete not just with generation but with increasingly interactive demand.
From these facts flow a host of tools and tactics that grid operators use to balance supply and demand. The easiest solutions are simply refinements or extensions of existing tactics.
The most common is, of course, power plants that are ramped up and down to meet demand. The energy in fossil fuels is the original storage, converted to heat and power as needed. Large amounts of wind and solar will increase the need for these plants to become more flexible, including capabilities such as faster start-up, lower minimum operating levels (called "turn-down levels"), and more nimble ramping up and down.
(Another meme to be addressed at another time is the notion that natural gas is the only kind of power plant that can follow load. This is also not true, as evidenced by a region like the Upper Midwest that was 88 percent coal and nuclear in 2011 and only 5 percent gas. Coal is now and always has been used for load following -- although there is an emissions penalty with this approach.)
Larger control areas are another easy solution to variability. Wind has constituted more than 60 percent of Xcel’s Colorado power supply at times, and more wind and solar is on the way. The quickest and easiest solution for Xcel is to combine its balancing operations with neighbors. The energy imbalance market (EIM) that starts operations this month is aggregating the balancing functions of California with the six-state region of PacifiCorp, and Nevada’s NV Energy will be included next year.
NREL has estimated that adoption of the EIM across the Western U.S. would save $150 million to $300 million per year. Similarly, the Midwest grid operator MISO found that adding Entergy to its system will save an estimated $1.4 billion over ten years, simply by pooling operations.
Better forecasting, both in the day-ahead and hour-ahead time periods, along with the simple change of dispatching power plants every five or fifteen minutes instead of every hour, has greatly increased the accuracy and reduced the cost of integrating wind and solar. NREL estimates that moving from hourly to ten-minute intervals for dispatch would save $1.3 billion a year if it was implemented across the West as part of the EIM.
Demand response is another tough competitor to battery storage, offering many of the same grid and customer services, at a fraction of the cost. It is much cheaper to put a switch and wireless control on an appliance than to install a lithium-ion battery system.
These basic facts are borne out by recent studies that have called into question the need for energy storage to integrate very large amounts of renewable energy.
The cost of many integration options were quantified in a paper for the 21st Century Power Partnership, a project of the Clean Energy Ministerial, which is a forum of the energy departments of twenty-three countries. Creating a sample supply curve, the report cites changes to system operations and market design as the least-cost options. These are followed by demand response and flexible generation, then by “advanced network management” and transmission.
Last on the list is storage. Even within storage, “chemical storage” in the form of batteries falls well behind thermal and pumped hydro. While the authors take care to point out that “option costs are system-dependent and evolving over time,” batteries clearly have a long way to go to be a competitive option against the twenty-two others on the list.
Source: 21st Century Power Partnership
Another study, released by NREL last November, found that storage wasn’t cost-effective until variable generation reached at least 40 percent of annual energy. The study points out that only surplus renewable energy would be used to charge storage, since selling it as it is produced would have a higher value. As long as there is demand for the power, it will displace other generation rather than be put into storage.
The study also found that the value of energy arbitrage -- shifting production from low-value times to high-value times -- “is relatively low and by itself is unlikely to yield a positive cost-benefit ratio for most existing storage technologies.”
Another value of storage, providing operating reserves like regulation and contingency reserves, is worth more, but the overall market size is small. And perversely, even a small amount of storage can collapse the prices for those services, undercutting its own value. A common analytical mistake is to use historic market data when evaluating the value of storage and renewables, the study authors say, “because the introduction of [variable generation] would fundamentally affect market prices and system operations.”
One shortcoming of the NREL study is that it does not place a value on avoided generation, transmission and distribution, or other services, including the ability of storage to address forecast errors and the increased deployment of operating reserves.
No major region of the U.S. is close to 40 percent wind and solar, but Germany hit 17 percent in the first three quarters of this year (out of 31 percent total renewables). Still, Agora Energiewende, the Berlin-based think tank devoted to the German energy transition, recently commissioned a report from four research institutes that found storage to be unnecessary in Germany until renewables hit at least 60 percent of the power supply. When renewables reach 90 percent, they calculate Germany will need 10 gigawatts of storage capacity.
“There are cheaper ways of adding flexibility than storage,” Daniel Fuerstenwerth, project manager at Agora Energiewende, told PV Tech. “For example, you can just turn off coal- or gas-fired power plants; you can use electricity for power-to-heat; you can exchange with neighboring countries. [...] There are other opportunities that are cheaper than storage, from the overall system perspective.”
But storage may thrive for other reasons, according to Agora, such as at the distribution level, for the benefit of customers.
They point out that feed-in tariff payments last twenty years, while PV systems may last longer. As customers reach the end of their tariff, a battery system will help them use all of their own PV output, rather than losing it to the grid. Batteries on the low-voltage distribution grid can also help smooth out production peaks” from distributed renewables, solving the back-feed problem that Hawaii is grappling with. And like NREL, they point out the cost-effective use of storage for ancillary services, though this application has a limited market size.
Then there is an increasing list of niche applications where storage makes sense today, or will soon, like off-grid, village power and islands. Developers are also combining different services to make storage pencil out, using the same system to reduce demand charges through peak shaving, provide ancillary services, deal with load pockets, and provide backup power for grid failures.
If batteries get as cheap as optimists say they will, even more opportunities will open up, such as competing head-to-head with gas-fired peaker plants. These combustion turbines are typically used only a fraction of the time, for meeting peak demand and providing some ancillary services, with capacity factors in the 5 percent to 10 percent range. When not deployed for these uses, they just sit there. A battery system that can provide supply at peak times, but also provide other services throughout the year, would have more opportunities to earn revenues.
Of course, the best-case scenario is if the batteries are already paid for. The “vehicle-to-grid” concept uses electric cars to both buy and sell power to the grid. It is being tested now by the University of Delaware and NRG Energy.
Even so, battery storage will have to compete with other emerging uses for the surplus renewable power. Agora points out the opportunity of “power-to-X,” converting surplus electricity to heat, chemicals and synthetic natural gas. While power-to-X is still in the research phase, it may be pushed by climate policies outside the electricity sector, creating low-GHG fertilizers, for example.
So while there may be a rationale for energy storage, integrating wind and solar into a modern power grid is not one of them.
There is one exception: to have a truly zero-carbon power system, we will have limited ability to rely on even the cleanest and most efficient gas-fired power plants. In a look at achieving California’s 80 percent carbon reduction goal, the California Council for Science and Technology found that the state would have to eliminate carbon emissions from the power sector completely -- while doubling in size -- to compensate for the inability of other sectors to reduce emissions as much.
Even if high-efficiency natural gas plants were used to follow load only, it would bust the carbon budget.
The report identifies the need for “zero-emission load balancing” (ZELB), where load balancing should be required to reduce emissions just like any other energy technology. “We will need major advances in energy storage and the construction and management of a smart grid to do this,” the authors conclude.