I spend ample time talking about the U.S. storage industry, and now I’ve got some data to supply an ex post facto justification.
The U.S. leads the world not just in fast food and starting trade wars, but in deploying batteries for the electrical grid. Australia technically nudged ahead for megawatts deployed in 2017, but the U.S. installed more energy capacity.
That lead will only grow over the next five years, at which point the U.S. will have roughly double the capacity of runner-up China. That intel comes to us from a new global energy storage report that GTM Research put together, which enables a new degree of comparative analysis relative to the previous country-specific research.
The comparisons reveal a pattern in how storage deployments initiate and evolve over time, even in drastically different geopolitical contexts. Understanding this evolutionary process will help other countries tap into this grid technology, and help existing markets grow into more sophisticated ones.
And if the evolution of storage markets interests you, sign up for the live webcast I’m doing exclusively for Squared members, with Stephen Lacey and Julia Pyper, on April 24.
The frequency revolution
A trend emerges when you look at how storage markets develop around the world: It starts with frequency regulation and branches out from there.
In the U.S., that happened in the PJM market, where storage developers started bidding in a decade ago to manage the rapid-fire ups and downs needed to keep the grid’s frequency in safe territory.
That gave storage developers a beachhead; they could make money in the short term while refining their approach to sourcing, integrating and deploying lithium-ion batteries.
The storage market started with frequency regulation in large part because it’s a good fit for the technology. Batteries can respond in the blink of an eye, much faster than it takes gas plants to ramp up or down.
“Storage can achieve the same or better level of performance [compared to conventional resources], given the rapid and efficient response for signals that are rapidly moving up and down,” said Ravi Manghani, energy storage director at GTM Research. “The nature of the signal itself is better suited for storage.”
Furthermore, it’s a cheaper battery use case compared to other ones out there.
Ancillary services typically call for batteries with a half-hour duration or less. For a given power rating, a battery system designed for frequency regulation will be cheaper than a battery geared for peaking capacity, which requires four or more hours duration.
In the early days of lithium-ion mass-production, with prices still up in the pre-gigafactory stratosphere, it made sense to start with a shorter-duration business model.
“Ancillary services are in many ways the first market that any policymaker would look at if they’re looking to adopt storage,” Manghani said.
And that’s exactly what happened all over the world.
The U.K. deployed 117 megawatts in 2017, primarily for frequency regulation. Those projects stem from a competitive auction that National Grid ran in 2016 for Enhanced Frequency Response, which requires sub-second response times from providers.
Germany delivered 135 megawatts in 2017. Its frequency market operates on weekly contracts, which provides much less certainty for developers compared to the four-year contracts of the U.K. EFR signal, but developers persist nonetheless.
South Korea’s first major foray into storage deployment exclusively targeted frequency regulation. Its monopoly transmission and distribution utility Kepco has gotten most of the way through its 500-megawatt roadmap, with the rest coming this year.
Frequency regulation has driven major storage projects in Australia, too. Tesla’s massive Hornsdale battery in Australia delivers quick response ancillary services to keep the grid humming as renewables grow and coal plants shut down.
Australia and Germany have bustling customer-sited storage markets to bolster their grid-scale deployment numbers, but what’s striking is the uniformity of frequency as the initial application of choice. Put another way, most of the world’s storage markets have entered the early-PJM stage of development.
An on-ramp, not a highway
The frequency application won’t sustain the industry for long, though. Finite demand constrains its growth potential.
There’s a reason frequency regulation is known as an “ancillary” service; it’s the side character, there to help out the superstar during unexpected twists in the plot. Those side characters never earn the big bucks.
“After a certain amount of megawatts for ancillary services, storage has diminishing returns in the advantage it can provide,” Manghani said.
He pegs that threshold in the range of 1 percent to 2 percent of a grid’s installed capacity. Beyond that, additional storage development starts to cannibalize the market, lowering returns across the board.
The limits of this application are visible in Kepco’s studious effort to identify exactly how much capacity it needed for the job, and build it in one 500-megawatt campaign.
Built systems, optimized for this particular job, also are vulnerable to changes in market rules or conditions. Such changes in PJM drastically slashed profits for batteries, and “could make future investment in the market untenable,” analyst Daniel Finn-Foley concluded last year.
Countries that want to harness the full potential of energy storage can’t stop at frequency regulation.
“Ancillary services can be the entry drug for storage, and after that some of the other longer-duration applications can open up,” Manghani said.
Storage for local capacity makes for a logical next step. California picked up the baton from PJM and pioneered that use case, with significant early success.
When the Aliso Canyon natural gas leak left Southern California’s fossil-fueled capacity stretched thin, the storage industry responded in six months with roughly 100 megawatts across several sites.
The batteries acted as miniature power plants, slipping quietly into populated areas where a gas plant would face years of permitting challenges. When the summer of 2017 rolled around, the feared blackouts or brownouts failed to materialize.
Now California is pushing the concept further, by shutting down or canceling gas peaker plants in favor of storage. This is cutting-edge stuff, and it carries some risk: Batteries run out of charge, unlike gas plants, so they’re not a one-to-one replacement.
That said, the state’s grid operator has signed off on batteries and distributed energy resources as fully capable of fulfilling the grid duties the Puente plant in Oxnard would have served, and a clean resource procurement round has commenced.
If storage developers succeed in this play, they could capture billions of dollars that would go to new gas peakers in the business-as-usual scenario for the coming decade.
The total U.S. frequency regulation market across all independent system operators is over 2.5 gigawatts. Storage could serve more than 1 gigawatt of that, according to Finn-Foley’s study.
Meanwhile, the U.S. is expected to build 20 gigawatts of new peaking capacity in the next decade. In GTM Research's base-case projection, 4-hour storage can challenge gas plants for 6.4 gigawatts of that capacity. That estimate is likely conservative, because longer-duration storage will be more prevalent and can displace a greater number of peakers, and battery costs have a history of falling faster than expected.
A recent study from NREL concluded that California alone could support a market of 7 gigawatts of 4-hour storage for peak capacity by 2020, based on the state’s renewable generation profile by that time. (That study looked at how much storage could join the grid before it loses operational value, so it's not evaluating the market competitiveness of the asset.)
There’s no reason that a similar demand for peaker batteries can’t arise in other solar-heavy regions around the world.
Arizona, with abundant sunlight and nasty evening peaks of electrical demand, is pioneering a different approach to peaker batteries.
First Solar contracted to build a massive battery to store solar power and release it precisely between 3 p.m. and 8 p.m. in the summer, when utility Arizona Public Service needs it. This simultaneously captures the low-cost energy of abundant solar, mitigates grid stress caused by too much midday solar and tackles the need for quick-ramping evening peak power.
That project beat out conventional gas plants, and every other type of challenger, to get the contract.
The business models around the much-discussed “renewables integration” role are still coming into focus; the values it provides tend to sit outside existing market products or rules. But as renewables advance in places like Australia, China and Germany, the need to time-shift solar from midday to evening peaks will only grow.