Energy Storage to Count On

In separate studies, EPRI and NERC highlight areas of concern regarding long-term battery performance and reliability

By Chris Warren

At the end of 2020, 12 states had goals to achieve 100% renewable or net-zero emission electricity generation. The states vary in geography, climate, and population, ranging from Maine in the northeast to Louisiana in the southeast to California in the west and the Hawaiian Islands in the Pacific. And it’s not only states making deep decarbonization commitments. Nearly 300 large corporations have made a pledge to go 100% renewable, and a number of utilities, like Xcel Energy, Dominion Energy, and the Sacramento Municipal Utility District (SMUD), have committed to net-zero electricity.

The blueprint for achieving these ambitious clean energy targets varies from state to state and organization to organization. But they all share a common feature: energy storage will play a prominent role. With so much intermittent wind and solar generation being added to the grid, storage is critical for maintaining reliability.

Grid-scale batteries are being increasingly deployed. According to Wood Mackenzie, 3.5 gigawatt-hours of battery energy storage were installed across the United States in 2020. For context, in the prior six years, a total of 3.1 gigawatt-hours of battery storage went into operation. Most of this growth was due to large-scale systems deployed by utilities.

The increasingly important role battery storage is expected to play in the power system of the near- and longer-term future raises important questions about the technology’s performance and reliability. Utilities, regulators, and other industry stakeholders are looking for assurances that they can count on these assets to perform when they’re needed, especially in the absence of a significant track record.

“We anticipate storage is going to be a bigger and bigger part of our operations,” said Steven Baxley, research and development manager for Southern Company. “Battery storage projects have so far been mostly pilots, but we are on the cusp of beginning commercial projects on the regulated and wholesale sides of our company. We want to understand whether the systems perform according to our specifications long term, whether they will be reliable, and what O&M (operations and maintenance) issues may be of concern over time.”

There may very well be reasons for concern. An initial EPRI analysis of battery performance metrics reported by manufacturers has revealed some discrepancies. While these discrepancies don’t immediately affect the ability of storage systems to deliver vital services, they may have significant impacts on their long-term performance.

EPRI Delves into Performance and Reliability

For the past three years, EPRI has been working to investigate the performance of batteries deployed in the field as part of its Energy Storage Performance and Reliability Data Initiative. During the recently completed first phase of research, EPRI monitored and collected data on numerous battery storage systems deployed in different locations, ranging in power capacity from 6 kilowatts to 1 megawatt. The systems were intended for various applications, such as residential and utility substation support. All the systems—with the exception of one vanadium redox flow battery system—are lithium ion batteries, the predominant technology being installed across the world today.

To assess energy storage performance and reliability, EPRI tracked two critical metrics and compared them with values reported by the manufacturers. One metric, state of charge, is the remaining capacity in a battery expressed as a percentage of its fully charged capacity. An accurate measure of a battery’s state of charge is essential, because it indicates whether the battery can deliver the energy, flexibility, and reserves needed by grid operators. For example, the state of charge measurement tells an operator if a storage system has enough energy to compensate for potential drops in wind and solar generation—a common and expected scenario in power systems with a high penetration of renewables.

Accurate measurements are also needed so that storage systems can be reliable participants in wholesale electricity markets, which have high performance standards. In addition, they can inform grid planners on the extent to which storage systems may help reduce peak demand and thereby avoid investments in powerlines, transformers, and other traditional grid infrastructure.

The second energy storage metric, state of health, is a measurement of system degradation. It is commonly expressed as the maximum state of charge available in a system at a particular time relative to its original capacity. As lithium ion battery storage systems are charged and discharged over time, their available energy capacity decreases, and their efficiency declines. Storage operators need accurate state of health data to determine whether their assets are performing to their specifications, associated performance guarantees, and warranty commitments. Accurate data also informs planning for maintenance, battery replacements, and other downtime.

“When manufacturers sell storage systems, they provide performance and lifetime expectations, but these may differ from actual performance because of unexpected degradation in real-world conditions,” said Steve Willard, an EPRI technical executive leading the Energy Storage Performance and Reliability Data Initiative. “Accurate state of health measurements over time can help storage owners and operators get ahead of unexpected changes in performance.”

Discrepancies in Battery Performance Metrics

On the surface, the state of health and state of charge metrics may appear straightforward, similar to a fuel gauge or odometer on a vehicle. But they are not as simple. The state of charge of a lithium ion battery can’t be measured directly and instead must be estimated by measuring parameters like voltage, temperature, and current. Similarly, the state of health is rarely measured directly. Standard testing approaches take the battery out of operation, fully discharge it, and then recharge it. By measuring the energy that enters and leaves the battery during the discharge and charge cycle, it’s possible to quantify the battery’s remaining energy capacity. Many owners and operators may not be able to collect such measurements at regular intervals, because they need their storage systems to deliver certain benefits. EPRI plans to develop and test methods that owners and operators can use to estimate state of charge “on the fly.”

Battery operating systems continuously report state of health and state of charge values to owners, but the vendors of these systems often don’t explain how they are calculated. “They are coming up with these values within a proprietary black box, and we don’t know what they are assuming,” said Willard.

After monitoring the storage systems deployed in the field for three years, EPRI found some notable discrepancies between the values it calculated and the values that the vendors were reporting. For example, one of the systems under investigation operated for a year and a half while regularly reporting that the state of health was 99.5%. EPRI’s measure was 96.5%, which translates into a degradation rate that is seven times faster and a lifetime that is potentially many years shorter. Additionally, EPRI research has revealed that battery systems have displayed state of charge values that fluctuated faster than is physically possible.

A Move Toward Standards

There is consensus among utilities, regulators, and other power industry stakeholders that such discrepancies are problematic, given the prominent role storage is expected to play in grid operations and electricity markets. Indeed, a recent report on grid-scale energy storage by the North American Electric Reliability Corporation (NERC) highlighted a lack of uniformity in storage performance data as a key challenge. As an important first step toward addressing this challenge, NERC and the Institute of Electrical and Electronics Engineers (IEEE) are developing standards and guidelines for reporting storage metrics.

EPRI’s research findings, field data, and approaches to assess state of health and state of charge are informing these standards. An expanded EPRI database of information on the performance of storage systems operating in the field can inform the standards as well. EPRI also worked with Sandia National Laboratory to develop a guide about the performance and reliability data that could inform relevant standards and procurement specifications.

Standards for reporting performance data are not new in the electricity sector. For decades, NERC has required operators of traditional electric generating equipment such as transformers, circuit breakers, and fossil fuel-powered generators to report performance data through a program called Generating Availability Data Systems (GADS). “A working group develops reporting standards, and NERC has a database where asset owners and operators have to report metrics—such as forced outage rates—that show how well the asset is running,” said Willard. “NERC recently developed reporting requirements for wind generators, and they’re about to finish developing requirements for big solar PV farms.”

EPRI is expanding its storage performance database by incorporating data from national laboratories, universities, utilities, and other institutions. “We need a lot of data for robust independent verifications of the metrics reported by vendors,” said Willard.

EPRI researchers are also developing tools that storage owners can use to assess system degradation and performance as well as report equipment failures and the amount of time it takes for repairs. The idea is to enable owners to conduct these assessments without having to take their systems out of normal operation. The tools draw on the latest research and powerful computational techniques. Last fall, the Energy Storage Integration Council, EPRI’s open industry collaborative forum, released a publicly available tool that supports uniform data collection for tracking storage operational activities.

“When a system fails, operators often have to wait weeks for the vendor to come and repair it,” said Willard. “Deploying a megawatt-scale storage system can cost millions of dollars, and taking it offline for a few days can mean a significant loss in value. These delays are reflective of a nascent market.”

As utilities deploy more storage around the world, they need more real-world performance data.

“We are eager to learn how these systems will operate and perform in terms of degradation and reliability in different conditions,” said Baxley. “Our wholesale business will operate some systems in the desert of California, and our regulated business is planning systems in the Southeast. We want to understand any differences in performance in these regions with different conditions. As we move toward the renewable future everyone is envisioning, it’s critical to know if storage is reliable.”

Artwork by Craig Diskowski/Edge Design

---