FADO: Fusion Agentic Design Orchestrator
The fusion pilot plant (FPP) vision cannot be realized without high-fidelity design simulations to reduce the risk associated with the first-of-a-kind device. A reliable digital representation of the design is needed to ensure an understanding of the facility’s performance, to optimize control systems, and to identify no-go designs.
However, code complexity necessitates a team of experts to perform the analysis, posing a substantial barrier to entry for new users without access to these powerful but highly complex simulation frameworks. Additionally, high-fidelity modeling and simulation often focus on the device design without considering maintenance, robotics, or construction. These considerations should be at the forefront of fusion design because Reliability, Availability, Maintainability, and Inspectability, known as the RAMI principles, directly impact economic viability.
This proposal will address these gaps by developing an agentic workflow to demonstrate the AI advantage in automating the time-consuming and difficult process of setting up and executing high-fidelity simulations for fusion systems. The Fusion Agentic Design Orchestrator will support fusion design by integrating high-performance computing modeling tools, RAMI principles, and robotics considerations, within a unified virtual environment.
Partners
- Georgia Institute of Technology
- Microsoft
