FRIB-led collaboration named U.S. Genesis Mission awardee for AI-powered accelerator research

Story originally published by Michigan State University

EAST LANSING, Mich. —  FRIB, in collaboration with Argonne National Laboratory, was chosen to receive a Genesis Mission Phase I award from the U.S. Department of Energy. The project will develop artificial intelligence, or AI, tools—specifically, physics-informed “digital twins”—that could transform how heavy-ion accelerators and isotope separators are operated. In addition to the FRIB-led award, FRIB will partner with other institutions on three other Genesis Mission Phase 1 projects.

The project, titled “Towards Self-Evolving, Physics-Informed Digital Twins of Ion Accelerators and Isotope Separators,” is led by Peter Ostroumov, associate director of the Accelerator Systems Division and professor at FRIB, and Brahim Mustapha, accelerator physicist at Argonne National Laboratory.

The award is among the inaugural national portfolio of projects the DOE announced 22 July under the Genesis Mission Request for Applications, or RFA, as part of the Genesis Mission, designed to double America’s scientific productivity. 

The Genesis Mission is a historic national initiative led by the DOE, which is building the world’s most powerful integrated science discovery platform. By uniting government, industry, academia and philanthropy, it is accelerating breakthroughs in energy, scientific discovery and national security through a new platform that combines AI, supercomputing, quantum systems and advanced scientific instruments.

Out of the largest funding response in DOE history — resulting in 278 awards across 342 participating institutions — the FRIB-Argonne collaboration places Michigan State University (MSU) at the leading edge of AI-driven scientific innovation.

The goal of the Phase I RFA awards is to identify promising pathways toward transformative scientific capabilities and establish a foundation for future investment and scale. Project teams will design and demonstrate research workflows that integrate AI with scientific investigation, while rigorously evaluating whether those approaches can accelerate discovery, improve predictive capabilities, enhance experimentation or generate new scientific insights. 

“At MSU, our land-grant mission has always been about tackling huge challenges that make a real difference. Receiving this Genesis Mission award alongside Argonne National Laboratory underscores the power of pairing top university talent with national lab resources,” said Shashank Priya, vice president for Research and Innovation at MSU. “By integrating physics-informed AI directly into FRIB’s operations, we’re cutting down machine setup times — helping scientists spend less time tweaking equipment and more time making discoveries in nuclear medicine, energy and fundamental science that serve the public good.”

Building intelligent models of complex accelerator facilities

FRIB and the Argonne Tandem Linac Accelerator System, or ATLAS, are DOE Office of Science user facilities that provide scientists with beams of rare isotopes to study the properties of atomic nuclei and their applications. Because experiments require different ion species and beam characteristics, operators must frequently adjust thousands of accelerator parameters to deliver beams with the required energy, intensity and purity. The process is highly complex and requires significant expertise.

To reduce the need for manual optimization, the project will develop physics-informed digital twins to streamline accelerator and isotope separator operations. These digital twins will combine established accelerator physics with AI algorithms trained on data from operating facilities. By integrating physics-based modeling with AI, this approach moves beyond conventional simulations and data-driven machine-learning models. The team will develop surrogate physics models for the recoil separator, led by Fernando Montes, senior research physicist at FRIB, and the fragment separator, led by Deuk Soon Ahn, manager of the fragment separation department at FRIB. Together, these models will extend the digital twin to include the full beam delivery process.

Unlike current approaches, which require optimization and tuning to be performed directly on the accelerator, the digital twins will provide a virtual environment for testing machine settings. Researchers can evaluate and optimize those settings before applying them to the physical machine. This approach reduces commissioning time, minimizes operational risk, and improves machine availability. 

During the Phase I project, the researchers will develop surrogate physics models for the accelerator and isotope separator systems. The team will demonstrate the digital twin using a 120-meter section of the FRIB linear accelerator. Testing the technology on a section of the FRIB accelerator will allow the team to evaluate its performance before expanding it to larger portions of the facility. 

Over time, the digital twins will become more accurate as they learn from real-world operations. The long-term goal is to create an integrated digital representation of the accelerator and isotope separator systems rather than optimizing each component independently. Demonstrating success in Phase I establishes the framework to extend the digital twin architecture across both the entire FRIB and ATLAS facilities under a potential Phase II award.

“It’s an honor for the FRIB team to be part of the historic Genesis Mission,” Ostroumov said. “Transforming a complex facility like FRIB into a more efficient, autonomous and productive scientific user facility requires tightly integrating AI into every stage of rare-isotope production. This project is an important step toward that vision.”

FRIB researchers will also collaborate on several other awarded projects led by partner institutions:

• From Data to Equation of State: AI-Driven Discovery of Optimal Observables in Heavy-Ion Collisions
  • Lead institution: Duke University
• Scalable Agentic Digital Twins for Autonomous Precision Facilities
  • Lead institution: Texas A&M University
• AI/ML Resonance Control for High-Reliability, Low-Cost Accelerator Operations
  • Lead institution: Fermi National Accelerator Laboratory

The mission of the U.S. Department of Energy is to ensure America’s security and prosperity by addressing its energy, environmental and nuclear challenges through transformative science and technology solutions. Learn more at energy.gov.

The U.S. Department of Energy Office of Science is the single largest supporter of basic research in the physical sciences in the United States and is working to address some of today’s most pressing challenges. For more information, visit energy.gov/science.

Michigan State University (MSU) operates the Facility for Rare Isotope Beams (FRIB) as a user facility for the U.S. Department of Energy Office of Science (DOE-SC), with financial support from and furthering the mission of the DOE-SC Office of Nuclear Physics.