FES Research: Theory, Simulation, and Artificial Intelligence
To develop a fusion energy source, scientists and engineers need predictive capability based in theory and computer simulations. FES contributes to this work by supporting theory and computational research, including efforts in artificial intelligence / machine learning.
Theory research is focused on advancing our understanding of the fundamental physical processes that govern the behavior of magnetically confined plasmas. These projects range from small single-investigator grants, mainly at universities, to large, coordinated teams engaging researchers from National Laboratories, universities, and private industry. The focus of the research spans fundamental analytic theory to mid- and large-scale computational work. Computational efforts within this program consist of high-fidelity simulations of plasma dynamics, AI-assisted digital twins of experimental facilities, multiscale modeling of structural materials, and foundation models of plasma science.
In addition to its scientific discovery mission, theory research provides grounding for the physics models used in computational simulations. These high-fidelity physical models are important for designing future fusion power plants. Theory also helps validate major experiments. This research leverages Office of Science investments in leadership-class computing systems and advancements in computational science.
Researchers are transforming how fusion reactors are modeled and designed. By combining experimental datasets with the power of high-performance computing, one project has developed models driven by artificial intelligence that can rapidly predict how particles move in tokamaks. Another team has integrated high-fidelity plasma simulation codes with commercial engineering tools. This effort is enabling scientists to streamline calculations of plasma instabilities and how they impact liquid metal blankets and structural components.
Together, these accomplishments signal a decisive shift toward an artificial intelligence-powered fusion ecosystem that can speed progress from experiment to deployment.
Learn more about research supported by the Fusion Energy Sciences program:
