Energy Storage

Artificial intelligence project looks to improve energy storage dispatch

Independent power producer Vistra is using artificial intelligence (AI) software developed by a team at the University of Texas at Dallas (UT Dallas) to help it better predict wholesale power market prices in California.

Vistra is using the software to project electricity prices for its soon-to-be 400-megawatt (MW) Moss Landing energy storage facility in Monterey County, Calif.

Vistra’s Moss Landing project is one of four energy storage projects awarded power purchase agreements with Pacific Gas and Electric in 2018 through a solicitation designed to find alternatives to renewing reliability-must-run contracts for gas-fired projects owned by Calpine that serve the South Bay area in California.

The software was developed by researchers from the University of Texas at Dallas who applied statistical and machine-learning methods to build models that Vistra is now using to predict near real-time bid and sell prices in California’s wholesale power market to enable it to buy electricity to charge the Moss Landing batteries at the lowest price and sell the stored energy at the most economically opportune time.

The joint project, which was funded by Vistra, was “crucial to optimizing electricity pricing at the Moss Landing battery farm that came online in early 2021," Rachit Gupta, vice president at Vistra and lead sponsor of the project, said in a statement. “The project was a tremendous success, and we are extremely happy that we availed ourselves of a great source of expertise that is present locally.” The software helps Irving, Texas, based Vistra make more precise pricing projections, Gupta said.

Its work for Vistra was the inaugural project for the Center for Applied AI and Machine Learning (CAIML) at UT Dallas’ Erik Jonsson School of Engineering and Computer Science. The center was established in 2019 to work with industry partners to apply advanced research in AI and machine learning to solve practical problems.

The UT Dallas researchers completed work on the AI project in August 2020 and held classes from December through February to train Vistra employees in the background technologies.

“AI can help a company like Vistra forecast future generation and demand on load, wind and solar energy, and optimize bidding, scheduling and deployment of energy to improve profitability and market participation,” Feng Chen, associate professor of computer science at UT Dallas and the project’s principal investigator, said in a statement.

Power sector increasingly looking at AI

The electric power industry is increasingly looking at AI for ways to solve difficult problems or improve the performance of complex systems. In March, the Electric Power Research Institute (EPRI) held a roundtable on AI in the power sector.

The roundtable was one of several EPRI hosted in its effort to foster collaboration between the power and AI industries through its AI.EPRI project.

Public power utilities, such as CPS Energy, are among the utilities exploring the uses of AI and machine learning. The San Antonio, Texas, utility is machine learning to improve its demand management and is starting to use the technology to improve its vegetation management programs.

In 2019, Salt River Project in Arizona signed a deal to use AI to improve its information technology operations. And in New York, the New York Power Authority (NYPA) is working with software vendor C3 IoT to use AI to help meet its energy efficiency targets.

In April, APPA received its third patent related to its efforts to help ensure that public power utilities have long-term access to advanced analytical technologies for business-related decision making.