PG&E demonstrates battery storage participation in California’s electricity market

Frequency regulation represents the best use of battery energy storage in California’s electricity market.
Published: Wed 16 Nov 2016

Storage is envisaged as playing a key role in California in helping the state to meet its renewables targets. Since as far back as 2010 the opportunity to harness its anticipated benefits have been in the sightlines. A mandate was issued in late 2014 for 1,325MW of storage procurement by 2020 – a first in the United States – and a roadmap was developed to overcome the barriers to deployment. [Engerati-California Seeks To Advance Energy Storage and California's Energy Storage Mandate-Will Others Follow?]

In order to help address the barriers and gain ‘real world’ experience, a project was conceived to use Pacific Gas & Electric (PG&E) batteries for the provision of ancillary services in the California Independent System Operator (CAISO) markets.

The project, funded by California's Electric Program Investment Charge (EPIC) programme, began in 2014 and utilized PG&E's first and then recently deployed 2MW Vaca-Dixon and 4MW Yerba Buena sodium-sulphur battery storage systems. The former is at a substation of the same name, while the latter is at the end of a distribution feeder on Yerba Buena Road in San Jose.

"When we started this project, there were no batteries participating in the CAISO markets, so the only information available about battery market performance came from economic models,” says Manho Yeung, senior director, Transmission Asset Management, PG&E. “So much goes into deploying a completely new type of technology in the CAISO market, and PG&E developed new workflows and procedures, and a new IT technology platform to make it all work.”

Key market participation findings

While the primary goal of the project wasn’t to optimise for highest revenues, a key finding was that frequency regulation represented the best financial use of the battery storage systems in this market. Frequency regulation provided significantly higher revenues than day-ahead or real-time market participation. For the month of August 2015, when the Vaca-Dixon system was bid into the market every day for its full frequency regulation range, total revenues were approximately $2,000/MW. This remained the highest revenue month for frequency regulation until early 2016 when revenues were significantly higher, primarily driven by higher daily procurement by CAISO. For the month of March 2016, the system operated exclusively for frequency regulation generated approximately $7,000/MW. Revenues began to return to lower levels in May 2016 as the daily procurement decreased.

The day-ahead and real-time energy revenues were found to be currently not conducive to energy arbitrage. The differentials between electricity prices at which the battery is charged versus those at which the battery is discharged were not large enough on a consistent basis to offset the inherent round trip efficiency of the systems, which averaged about 75%.

On the other hand, spinning reserve revenues can very modestly add to resource revenues with a value of approximately $4/hour/MW.

These findings indicate that current market dynamics do not favour long-duration batteries. As the most significant revenues are from frequency regulation, and that frequency regulation is a power rather than energy product, a 30-minute system might be able to provide the same capabilities as a 7-hour system, with presumably less capital investment.

The location also was found to be important as the Locational Marginal Price (LMP), which drives the economics of when the resource will charge or discharge, differs based on the congestion and losses at different locations in the CAISO system. While two different geographic locations might have equivalent amounts of renewable generation at a given time, there is a possibility that only one location has a negative LMP. If the storage resources are geographically outside the grid conditions driving negative LMPs, it may not make economic sense to charge. For example, even if there is over generation in Southern California, such a condition may not incentivize energy storage in Northern California to charge.

Next steps for storage

According to the final report, the project successfully achieved all of its key objectives. During the course of the project, various challenges were identified and many of them resolved. Recommendations were also made for enhancements to the CAISO market model for storage.

“We've successfully tested how batteries actually perform in the markets, substantially helping to inform the conversation about energy storage in California. Now that we have put these elements in place, it will be much easier to bring future battery systems online,” says Yeung, commenting on the findings.

PG&E will continue to maintain the Automated Dispatch System that was developed as a platform to automate the response of current and future battery storage resources to CAISO market awards. In addition, PG&E plans to provide a new flexible ramping product that CAISO plans to introduce in late 2016. This product is designed to provide additional revenues for fast ramping resources such as batteries.  

Finally, in 2017, PG&E will utilize the Yerba Buena battery for another technology demonstration involving the coordination of 3rd  party distributed energy resources – such as residential and commercial solar – using smart inverters and battery storage controlled through a distributed energy resource management system (DERMS).

Further reading

Related Webinar