Massive energy storage system from Tesla and PG&E will use Fluence AI to optimize bidding and battery health

Pacific Gas and Electric’s 182.5-megawatt/730-megawatt-hour battery energy storage system in Moss Landing, California is using Fluence’s trading platform to optimize market bidding performance in the wholesale market of the state’s grid operator, CAISO. Fluence contends that using AI for advanced price forecasting can improve revenue for the utility and make the grid more efficient. Fluence has already tested the software in PG&E’s bidding and scheduling systems for the 4 MW Yerba Buena and 2 MW Vaca Dixon energy storage projects.

When we interviewed Seyed Madaeni, chief digital officer of Fluence (and former CEO of AMS, which was acquired by Fluence), he described the software as ​“technology-agnostic” with the ability to optimize not only the Moss Landing project but potentially ​“entire portfolios of generation and storage resources.” Fluence claims that the trading platform can increase revenue and operational efficiency for battery-based energy storage by 40 to 50 percent.

The Moss Landing battery system will comprise 256 Tesla Megapack batteries, with energization and AI-powered bidding scheduled for later this year. An expansion was approved in August that could increase the Vistra-Energy-backed battery to a spectacularly sized 1,500 MW/6,000 MWh — the biggest battery in the world by far.

The holy grail: Stacking energy revenue

CAISO allows an energy source to bid into a number of wholesale markets, including ancillary services such as frequency regulation or spinning reserve, as well as real-time and day-ahead energy markets.

The holy grail for extracting value from energy storage has always been the ability to ​“stack” these revenue streams. However, few battery operators have been able to pull off this feat, despite a bevy of press releases making claims to the contrary.

Fluence’s Madaeni said the company is ​“putting our money where our mouth is” in its capture of multiple stacked revenue streams, while still accounting for battery lifetime and degradation.

According to Madaeni, traditional energy trading conducted via spreadsheets and manual inputs is suitable for thermal assets such as coal and gas plants, with a calculus based on marginal cost. ​“But renewables and energy storage makes this very complicated,” he said, adding, ​“Batteries won’t solve the problem; it [depends on] how we use the battery.”

Fluence is a Siemens and AES company with energy storage software deployed or awarded at 2.4 gigawatts’ worth of projects.