Energy storage has inspired many metaphors to convey the versatility of its uses. Some call it the bacon of the grid, making other resources taste better. But the image of the Swiss Army knife captures the sheer number of different roles this technology can play.
Last week, Storage Plus told the story of frequency regulation, the tool in the toolkit that launched the modern grid storage industry. But frequency regulation markets are shallow, and they saturate quickly. After building a track record in PJM’s fast frequency market, battery developers jumped into capacity, an application with a much larger addressable market.
Capacity provides the essential service of delivering electrons exactly when they’re needed. If energy storage can crack open that application, it will cement its place at the core of the evolving grid.
This outcome was never guaranteed. There’s a world where batteries remained a useful tool for the lightning-fast frequency regulation task, toiling out of sight and in modest numbers around the country. But battery costs fell to the point that bigger, longer-lasting storage plants became not just feasible but attractive. And the stunning success of solar photovoltaics offered a collaboration uniquely tailored for what batteries can do.
As of fall 2020, few batteries in the U.S. primarily serve a capacity function. But a rumble has begun, and soon a torrent of capacity batteries will rush through the pipeline. It starts in California, where the market fundamentals left this clean energy pioneer without enough in-state capacity. Batteries are already filling the gap.
More broadly throughout the West, developers are packaging solar plants with batteries, turning them into sources of clean capacity. Analysts sometimes refer to “renewables integration” or “renewables shifting” as a separate use case for energy storage, but a wily developer would be hard-pressed to get paid for “renewables integration.” Instead, they earn checks for firm capacity, in many cases to replace fossil-fueled capacity that is shutting down.
Frequency regulation remained niche, but capacity has the potential to make battery storage a central player in the future grid. A National Renewable Energy Laboratory study from 2019 found “roughly 28 GW of practical potential for 4-hour storage providing peaking capacity, assuming current grid conditions and demand patterns.” The study notes that such widespread deployment would in turn reduce storage costs, contributing to further growth.
“To me, [capacity] is the biggest use case, especially in regions going to higher shares of renewable energy,” said Jason Burwen, vice president for policy at the U.S. Energy Storage Association.
When asked if capacity would eventually dwarf all the other storage use cases, Wood Mackenzie Energy Storage Director Daniel Finn-Foley said, “Short answer: yes. Long answer: Oh my goodness, yes.”
More quantitatively, WoodMac’s front-of-the-meter forecast predicts that 80 to 90 percent of batteries installed in the next decade will serve capacity applications primarily, Finn-Foley noted.
Expanding the mission
The PJM frequency regulation boom showed power companies that they could make money on batteries as commercial grid assets. But that market cooled within a couple of years, and PJM offered little traction for other uses of storage.
Instead, the action shifted westward to California. In 2013, the state compelled its utilities to procure 1.3 GW of energy storage by 2020. Around that time, Southern California Edison opted to close the San Onofre nuclear power plant, located on the coast between Los Angeles and San Diego. The policy push merged with this capacity loss to create an opportunity.
Southern California Edison called for bids to meet its Local Capacity Requirement and picked the winners in 2014. Most of the 2,200 megawatts, due online by 2022, would come from natural gas. But the utility was required to pick at least 50 MW of “preferred resources,” the California energy bureaucracy’s term for clean or distributed resources, including batteries.
After researching the status of storage technology, SCE picked five times the minimum requirement. As Greentech Media reported at the time:
“No utility has made such a big investment in customer-owned, distributed energy storage assets of this type before, making this a step into the unknown on the part of SCE. The LCR process will set up power-purchase-agreement-type structures with the contract winners, providing them a guaranteed revenue stream.”
It’s notable that this first major foray into batteries-as-capacity chose both large centralized assets (the 100 megawatt/400 megawatt-hour AES Alamitos plant, the largest battery contracted at the time) and distributed batteries, like the networked devices Stem and AMS proposed to install at commercial and industrial sites in the region.
Six years later, it’s clear that not everything went according to plan. Ice Energy went bankrupt; AMS almost did but survived by pivoting to software. It handed its SCE fleet over to Stem to operate. Things changed in a positive way for the storage industry too: Regulators rejected the peaking gas plant NRG was supposed to build in the Moorpark area, and it was replaced by batteries instead.
"With the development of new resources, there is always a risk of project failure," said William Walsh, SCE’s vice president of energy procurement and management. "However, we believe the lessons learned in this solicitation helped pave the way for future procurement of emissions-free resources to support SCE’s Pathway 2045 vision," the utility's decarbonization plan.
AES Alamitos is still moving toward completion, expected later this year, meaning the megabattery that kicked off the trend has yet to see any action. But that’s the funny thing about grid batteries: the assets lag the indicator. The confirmation that a utility would buy power from an unimaginably large battery gave other developers confidence to push ahead. The signal, not the actual delivery of the project, moved the market forward.
Utilities piled on
In the years after SCE’s vote of confidence, a particular form of capacity asset rose to prominence in the West: the solar-plus-storage plant.
Combining cheap and predictable solar generation with the dispatchability of four- or five-hour batteries gave utilities a clean way to meet evening demand. By 2018, it was starting to cost less than fossil fuels.
That was the year Arizona Public Service awarded its first battery peaker to First Solar. In Hawaii, AES constructed a solar peaker for the Kauai Island Utility Cooperative. Soon, utilities like Xcel Energy announced coal plant retirements with new solar and battery facilities to replace the lost capacity. Nevada’s NV Energy signed some of the biggest hybrid plants in the country. NextEra Energy Resources helped pioneer the format with Tucson Electric Power, and even added wind to triple-threat deals in Oregon and Oklahoma.
As different models for valuing and contracting storage solidified, utilities started including storage in the integrated resource plans where they sketched out how their capacity mix would evolve over time. These documents do not guarantee that what they envision actually gets built, but they do reflect how utilities anticipate storage growing.
This WoodMac visualization of storage in utility planning shows broad geographic diversity and interest in several states that have not built much battery capacity to date.
WoodMac’s database counts 16,275 MW of storage in utility integrated resource plans spanning 2021 to 2030. Over the 15-year time horizon, 19 utilities envision 25,056 MW of storage joining their fleets. But that tally is bound to grow, said analyst Gregson Curtin, because IRPs from several rapidly growing storage markets are a few years out of date and will soon be refreshed.
In some cases, the acceleration is already startling. Arizona Public Service had 4 megawatts operating in 2017, but chose 500 MW for its long-range plan, Finn-Foley noted.
“Three years later, their chosen portfolio is 4.8 GW,” Finn-Foley said. “In three years, a utility went from ‘1 GW is too much’ to ‘5 GW is needed.’”
Getting paid is the hard part
Standalone batteries in a peaking capacity role are still quite rare.
The Aliso Canyon procurement delivered a handful to Southern California, led in size by San Diego Gas & Electric’s Escondido system (30 megawatts/120 megawatt-hours). PG&E’s enormous Moss Landing complex is still under construction by Vistra and Tesla. A new round of batteries totaling several gigawatts will hit the California grid in the next two years, filling a gap in the state’s capacity resources.
One of those projects, LS Power’s Gateway battery, is already up and running because the developer didn’t wait for a utility contract to get started. It’s already delivering short-term capacity (resource adequacy, in California parlance) and serving a range of energy contracts, such as a novel structure in which the battery shifts daytime power to peak windows on behalf of a customer.
But Gateway is an outlier in coming to market without a confirmed, long-term utility contract. Much of the storage being built across the West relies on a capacity construct, Burwen noted.
“It’s a lot easier to finance and build something when you have 20-year contracts,” he said.
It gets more challenging in the competitive markets, Burwen added, where developers have to build a business case without long-term certainty. With few examples to look at, developing forward revenue projections is challenging.
A few brave developers have pushed into Texas, willing to bet on a strategy without seeing other people do it first. Broad Reach Power and Key Capture Energy are building out batteries at targeted nodes. They typically choose one hour of energy duration, which keeps costs down but still allows for hitting peak prices in ERCOT. Of course, ERCOT has no capacity market, but a battery discharging at a certain time to serve peak demand counts as a capacity for the purposes of this analysis.
East Coast expansion underway
Farther East, capacity is one of the most hotly anticipated but least actualized uses for storage.
Virginia adopted a massive storage mandate this year, and utility Dominion Energy is examining proposals for storage as a capacity resource. But that process has just begun.
New York has talked about storage for peak capacity longer than Virginia has, but capacity for discussion has not translated into capacity in the field. New York City utility Con Ed held a request for proposals for 300 MW of battery capacity last fall but has yet to announce any contracts from that process. An environmental rule will squeeze out the dirtiest peakers, and New York City will certainly need new capacity to replace them.
But without utility contracts, developers have to take a risk on the unproven prospects of bidding batteries into the NYISO capacity market. Making matters worse, the ISO is implementing a “buyer-side mitigation” rule that would raise the price that storage can bid in the New York City region, neatly disincentivizing battery capacity in the exact place it would be most valuable.
PJM, the birthplace of modern energy storage, has shut batteries out of its capacity market by demanding 10 hours of discharge duration, which kills project economics and is out of step with pretty much anywhere else. But PJM’s stakeholders recently voted to eliminate the 10-hour rule and will file a tariff change to effect that.
“The hope is that instead of drawing an arbitrary line, not based on analysis, as to what we think is needed, we’re replacing this with an analytical approach,” Burwen said.
Massachusetts mobilized policy to create demand for storage where market fundamentals are lacking. It enacted a Clean Peak rule, forcing utilities to obtain an increasing share of peak hour electricity from clean sources. The regional market lacks the clear carbon-intensity binaries of California (sunny hours clean/peak hours dirty), but the carbon value of this rule will become clearer as local policy produces a New England version of California's duck curve. An early battery victory in the New England capacity market came from residential installer Sunrun, which will deliver 20 MW by 2022.