Ohio State part of team investigating AI control of nuclear reactors

Researchers from The Ohio State University are part of a consortium focused on developing nuclear power plant control systems that utilize artificial intelligence. The team has received a nearly $3.4 million federal grant from ARPA-E (U.S. Department of Energy's Advanced Research Projects Administration-Energy) to bolster its efforts.

Led by N.C. State University, the consortium includes members from New Mexico State University, The Ohio State University, Oak Ridge National Laboratory, Idaho National Laboratory, TerraPower and Zachry Nuclear Engineering.

The group’s project—Development of a Nearly Autonomous Management and Control System for Advanced Reactors—is the second largest funded by ARPA-E aimed at mobilizing research to develop new, innovative technologies for the next generation of nuclear power plants. Under the agency’s umbrella MEITNER (Modeling-Enhanced Innovations Trailblazing Nuclear Energy Reinvigoration) program, initiatives are funded which enhance advanced nuclear reactors’ commercial viability.

Ohio State’s principal investigator is Carol Smidts, professor and director of the university’s Nuclear Engineering Program. She is overseeing one of the consortium’s four thrust areas, Nearly Autonomous Management and Control (NAMAC) System Technical Components Development.

According to Smidts, “the initial application of the artificial intelligence-based control system is the management of accidents for advanced reactors."

“Ohio State researchers will determine the placement and type of sensors to be used to maximize the ability to detect possible issues and improve system robustness,” she said. “Modeling and assessing reactor operator response within the context of a more autonomous operation is another, equally important target.”

Building on approaches and tools developed by through funding from the Air Force Office of Scientific Research, Battelle Energy Alliance and the U.S. Nuclear Regulatory Commission will be key to their success.

Overall, the consortium seeks to develop a highly-automated management and control system for advanced nuclear reactors. The system will provide recommendations to plant operators and will use artificial intelligence and continuous data monitoring to predict future plant status through machine learning.