The debate over energy storage replacing gas-fired peakers has raged for years, but a new approach that shifts the terms of the argument could lead to an acceleration of storage deployments.

Rather than looking at peak demand as a single mountainous peak, some analysts now advocate a layered approach that allows energy storage to better match peak needs. The idea is beginning to gain traction with some states and utilities.


“You don’t have to have batteries that run to infinity.”

Ray Hohenstein

Market applications director, Fluence


Some developers of solar-plus-storage projects say they can already compete head-to-head with gas-fired peakers. “I can beat a gas peaker anywhere in the country today with a solar-plus-storage power plant,” Tom Buttgenbach, president and CEO of developer 8minutenergy Renewables, recently told S&P Global.

Others disagree. Storage is not disruptive for generation, but will be disruptive for transmission and distribution, Kris Zadlo, executive vice president and chief development officer at Invenergy, told the audience at a Bloomberg New Energy Finance conference last spring. Invenergy develops generation, energy storage and transmission projects.

But there is another path that avoids the pitfalls of positions on either end of the all-or-none approach. “Do the analysis of the need itself,” Ray Hohenstein, market applications director at Fluence, told Utility Dive. If the need is only two hours in duration, it may be best served by a two-hour battery. “You don’t have to have batteries that run to infinity.”

Storage vs. fossil fuel peakers

Energy storage has several benefits over traditional fossil fuel peaking plants, Hohenstein said. It is instantaneous, it has no emissions and requires no fuel, and has limited infrastructure needs. It can also help the grid absorb higher levels of renewable generation by soaking up excess output, such as solar power at noon. But the one thing energy storage cannot do, he said, is provide limitless energy.

So, instead of looking at replacing an individual peaker, Hohenstein advocated a “duration portfolio” approach that uses energy storage to shave peak load.

If the need is for 150 MW of resources that will never need to run for more than two hours at a time, then a battery is “quite cheap,” significantly less than a four or eight-hour battery, said Hohenstein. “If you fill up your peak by duration layer, it could be more cost effective.”

NREL research driver

Fluence’s approach is informed by research by Paul Denholm and Robert Margolis at the National Renewable Energy Laboratory (NREL), released last spring.

The NREL researchers looked at the California market where they said 11 GW of fossil fuel capacity is expected to be retired by 2029 because of new once-through-cooling requirements that are taking effect. A lot of that capacity is peaking capacity and, according to NREL’s analysis, a large fraction could be replaced with four-hour energy storage, assuming continued storage cost reductions and growth in solar installations.

The key in NREL’s research was the level of solar power penetration. There is a “synergistic” relationship between solar penetration and storage deployment, the researchers wrote.

When solar penetration is lower than about 11%, the potential of four-hour storage is lower than it would be with zero solar deployment because solar penetration of 11% or less flattens the load curve. Above 11% penetration, however, NREL found that solar power creates a “peakier” load curve that increases the potential of four-hour storage.

As Hohenstein explained in an October blog post, peakers typically run 10% or less of the year, and many never run for longer than four hours at a time. If that is the case, then a four-hour energy storage system could do the same job.

The value of two, four or six-hour storage is a difficult one to answer because “it’s not often investigated,” Hohenstein said, but “it’s an important question as the cost of energy storage is highly dependent on the number of hours of duration needed.”

How much storage to procure?

That changes the terms of the conversation, Hohenstein said. Instead of focusing broadly on the cost competitiveness of peakers vs. storage, system planners are wrestling with a different question, he said, “How much energy storage should we procure?”

In many instances, however, that analysis is still waiting to be done. “We haven’t seen a thorough understanding of peak needs in most territories in the U.S.,” Hohenstein said. “No utility has literally taken my approach yet, but it is a clear direction they are going in. Utilities are being more and more clear about when resources will be needed.”


“Time shifting with energy storage is the natural conclusion.”

Daniel Finn-Foley

Senior energy storage analyst, Wood Mackenzie Power & Renewables


“I think that is the future” — storage providing not just peaking power, but clean peaking power, Daniel Finn-Foley, senior energy storage analyst at Wood Mackenzie Power & Renewables, told Utility Dive. “This is a natural conclusion to several different policies.”

The evolution of that policy began when system planners realized that not all electrons are created equal. That resulted in the creation of the renewable portfolio standard (RPS), Finn-Foley said. Thirty-eight states and the District of Columbia now have some form of RPS. The next step, Finn-Foley said is the recognition that not all hours are equal.

If, for whatever reason, a state says an electron from a wind turbine has a different value at different times of day, “that is how you identify value,” Finn-Foley said. “Time shifting with energy storage is the natural conclusion.”

Clean peak growth

Massachusetts is leading the charge in that effort. The state in July passed a “minibus” energy bill that included the creation of a clean peak standard. The law follows the RPS framework, but recognizes the time value of a resource’s contribution toward a cleaner environment. The law requires every retail electric supplier to provide a “minimum percentage of kilowatt-hours sales to end use customers from clean peak resources.”

Though Massachusetts is first, it is going to take time before the law has a practical effect. The state’s Department of Energy Resources earlier this month set the 2019 clean peak standard at zero while it works out other details of the program.

New York is also looking at establishing a clean peak program, and the concept is under discussion in Arizona. “2019 is going to be the year clean peak energy is recognized for providing value,” Finn-Foley said. “It is a natural bridge to 100% renewable energy.”


“Leading utilities are beginning to carefully measure the shape of their peak need, and assemble portfolios of resources that can provide a range of durations.”

Lon Huber

Director, Navigant


A clean peak standard is not the only way to encourage the creation of duration portfolios. The influence can also run in the other direction. “Duration portfolios accelerate the adoption of storage, and ultimately promote clean peaks, which is critical to reducing emissions,” Lon Huber, a director at Navigant that helped develop the concept of a clean peak standard, told Utility Dive via email.

Duration portfolios are “the natural extension of some of the more advanced procurement practices out there today,” Huber said. “Leading utilities are beginning to carefully measure the shape of their peak need, and assemble portfolios of resources that can provide a range of durations.” Huber pointed to the contract Arizona Public Service signed with First Solar last February for a 50 MW, 135 MWh battery to help shift the output of a 65 MW solar farm, as well as solicitations by Public Service Company of New Mexico and Salt River Project that call for combining energy storage with renewable resources.

“We need to seek to move beyond having single capacity curves that apply equally to all storage resources,” Huber said. Instead, he recommends filling need in tranches of duration based on the shape of a utility’s net peak.

Looking more deeply into the issue, in addition to creating layers of duration, Huber also recommends coming up with different net load shapes to stress test the “unlayering” of peak load. “I would run a loss of load probability analysis to figure out what load shape leads to an outage. I would also look at yearly extremes in the near past as well as in the future as more renewable energy resources come online,” Huber said.

“Peaker dispatch patterns are a popular proxy for storage duration needs, but they paint an incomplete picture,” he said.