The Department of Energy’s (DOE's) SunShot Initiative worked with IBM and its supercomputing capacity to improve solar forecasting accuracy by as much as 30 percent. Solar forecasting can help utilities and solar owners better predict how a system will perform and operate over time.
The tool uses machine-learning technology and big data to better forecast weather and part of the SunShot Initiative’s Improving Accuracy of Solar Forecasting funding program. “Because solar production can be hampered by cloud coverage and bad weather, solar generation levels can be difficult to gauge on a day-to-day basis,” said DOE.
“Utilities and electric grid operators are responsible for meeting consumer electricity demands, and when solar system production varies, an accurate solar forecast is needed in order to maintain an efficient supply of energy reserves for solar customers,” the department said. “When utilities and electric system operators better understand generation patterns, they’re able to maximize solar resources, operate more efficiently, and improve solar energy’s economic competitiveness.”
At IBM’s Thomas Watson Research Center a research team led by Dr. Hendrik Hamann, created the self-learning weather model and renewable forecasting technology that it calls SMT. The platform synthesizes data from numerous sources. The sources include historical weather data as well as real-time measurements from local weather stations, sensor networks, satellites and sky imagers. Watt Sun uses the Department of Energy’s computing facilities to pinpoint weather prediction models. The models can forecast weather more accurately up to months in advance.
They can use that information to anticipate solar generation levels and better match consumer electricity demands. That information can also help with other forms of renewable energy like wind and hydropower. That can help utilities understand how to integrate more solar and renewable energy into the electric grid. IBM will make the forecasts available to government agencies and other organizations to help speed adoption of solar power.
When the SunShot Initiative launched the forecasting program it also supported two other forecasting efforts. One with the National Oceanic and Atmospheric Administration, the Solar Forecast Improvement Project and another with the University Corporation for Atmospheric Research, the A Public-Private-Academic Partnership to Advance Solar Power Forecasting. Thanks to their successes the SunShot Initiative added a second round of funding to the forecasting tools last year.
IBM also is doing other work to improve solar power in a number of ways. For instance, it’s partnered with Airlight Energy, ETH Zurich and Interstate University of Applied Sciences Buchs NTB to develop low-cost high-concentration photovoltaic thermal (HCPVT) systems.Tweet