Cross‑facility agentic workflows that connect data, tools, and expert knowledge
Today’s particle accelerators rely on vast data and decades of experience. MOAT‑Core, led at ORNL by Alexander Zhukov in partnership with Lawrence Berkeley National Laboratory, will expand a shared AI platform that helps accelerator operators diagnose issues faster, plan machine setups, and improve beam quality and uptime regardless of location. This shared platform will help teams collaborate and apply lessons quickly across the DOE accelerator ecosystem.
Internally, the platform organizes control‑system archives, electronic logbooks, simulations, manuals, and procedures into common ontologies and AI‑ready formats. Agentic AI workflows pair advanced reasoning models with accelerator software to assist in routine operations and design tasks. The approach builds on cross‑lab deployments and allows gains at one facility to be reused at others, supporting the accelerator lifecycle from design through long‑term operation.
