A team of institutions that includes Argonne has been awarded $3.8 million by the U.S. Department of Energy to combine new methods in cloud physics and modeling, big-data integration and processing, and deep machine learning to make solar energy more viable in the energy marketplace.
IBM leads the multisector, multidisciplinary team, which includes Argonne, the National Renewable Energy Laboratory, Northrop Grumman Corp., Northeastern University and the University of Arizona. Other team members include electric grid operators and solar power companies.
Current attempts to predict hour-ahead, let alone day-ahead, cloud cover and solar photovoltaic power generation have been notoriously inaccurate, one of the major barriers to the widespread use of solar energy. The team will attempt to lower that barrier by applying big-data processing and machine learning and combining them with better cloud physics modeling to vastly improve the accuracy of solar forecasting.
IBM’s contribution includes members from the Watson project, which built the computer system famous for winning the television quiz show Jeopardy! in a human-versus-machine competition. Information about the initiative is available online.
Posted December 10, 2012