Using machine learning, DeepMind, an artificial intelligence firm and Google affiliate, increased the value of wind farms by approximately 20 percent, according to the firm.
Last year, DeepMind and Google started applying machine learning algorithms to 700 megawatts of wind in the central U.S., researchers at the artificial intelligence firm said in a blog post late last month. The wind farms are part of Google’s global fleet of renewable energy projects.
DeepMind used a neural network trained on widely available weather forecasts and historical turbine data to predict wind power output 36 hours ahead of time, according to Carl Elkin, DeepMind research engineer, Sims Witherspoon, DeepMind program manager, and Will Fadrhonc, Google Carbon Free Energy program lead.
The DeepMind model recommended day-ahead hourly delivery commitments to the power grid, the company said, noting that scheduled deliveries are often more valuable to the grid.
“To date, machine learning has boosted the value of our wind energy by roughly 20 percent, compared to the baseline scenario of no time-based commitments to the grid,” DeepMind said.
Machine learning — a branch of artificial intelligence centered on computers analyzing data to find patterns and learn from the information — can help better predict wind farm output and electric demand while improving wind farm operations to save money, according to DeepMind.
Machine learning can make wind power more predictable and valuable, according to the company. Machine learning can also help wind farm operators make smarter, faster and more data-driven assessments of how their power output can meet electricity demand, the DeepMind researchers said.
Researchers and practitioners are developing new ways to make the most of variable power sources like solar and wind, the researchers said.
“We’re eager to join them in exploring general availability of these cloud-based machine learning strategies,” they said.
DeepMind and Google are owned by Alphabet. Google, the largest corporate renewable energy buyer in the world, has contracted for more than 3,000 MW of renewables, mainly from wind farms.
Google bought London-based DeepMind in 2014.
Brookings Institution expert weighs in
The announcement comes as artificial intelligence is increasingly being used to improve renewable energy output and in other energy-related areas.
The Brookings Institution, a Washington, DC-based think tank, expects that artificial intelligence will have a growing effect on the renewables sector.
Wind turbine heads, for example, can be shifted in real-time to capture a greater fraction of the incoming wind, something that has been possible for a long time but can be done more efficiently with machine learning, David Victor, co-chair of the Brookings Institution’s energy and climate initiative.
Machine learning can improve solar forecasting, leading to better day-ahead and hour-ahead predictions, making it easier and more lucrative for solar generators to participate in electricity markets, Victor said in a post earlier this year.
Artificial intelligence can also be used to help electricity customers respond to price changes where there is time-of-use pricing, according to Victor.
In California, wholesale prices sometimes become negative when solar output soars, but then jumps when solar generation falls off at the end of the day, Victor said, noting the pattern is one of the reasons the state adopted time-of-use pricing this year.
“AI can allow even small consumers to automatically adjust their power consumption in real time with prevailing prices—something that ordinary people won’t do unless they like sitting at home staring at real-time data from power markets,” Victor said.
NYPA to deploy artificial intelligence-based application
The New York Power Authority last year selected C3 IoT to provide a software platform to help the Power Authority and the state implement and meet its energy efficiency targets.
Under a multi-year, software-as-a-service agreement, NYPA will deploy C3 Energy Management, C3 IoT’s, artificial intelligence-based application, as part of New York Energy Manager, NYPA's advanced, secure energy management center, headquartered in Albany, N.Y. It provides public and private facility operators across New York State with timely data on energy use.