- New DOE-supported project aims to predict plasma behavior inside fusion reactor.
- Digital twin will connect to physical SPARC device to improve prediction accuracy.
- Project will also use surrogate models to accelerate computation.
- Researchers must verify prediction accuracy under actual operating conditions.
A new project supported by the US Department of Energy aims to use a digital replica of the SPARC device from Commonwealth Fusion Systems (CFS) to predict plasma behavior inside a fusion reactor in under 20 minutes. The initiative is led by a US fusion energy company and brings together researchers from four DOE national laboratories, NVIDIA, Google DeepMind, and multiple universities. The project will develop a digital twin that predicts SPARC's operation through physics simulations, experimental data, and advanced computing.
The initiative is part of the Department of Energy's Genesis mission, which announced its second-phase awards on October 8.
How the Digital Twin Works
The Princeton Plasma Physics Laboratory (PPPL) will provide simulation software, experimental data, and computing resources for the project. Unlike traditional simulations, the planned digital twin will connect to the physical SPARC device currently under construction. As experimental data becomes available, researchers intend to update the virtual model to improve its predictive accuracy and guide future experiments.
SPARC is a tokamak that uses powerful magnetic fields to confine extremely hot plasma. Researchers must understand plasma behavior under different operating conditions to improve reactor performance. The digital twin aims to predict a full plasma pulse, or an experimental operating cycle, within 20 minutes. This will enable scientists to analyze results and prepare adjustments between consecutive experiments. PPPL will provide three established physics simulation codes: M3D-C1, TRANSP, and XGC.
These tools simulate different aspects of plasma behavior and will be combined with other software to generate predictions.
The project will also use faster computational models trained on large simulation datasets. Known as surrogate models, these can approximate the results of more complex calculations while requiring less compute time. "PPPL will play a central role in this project, contributing foundational software engineering and SI models, core physics models, historical tokamak data, and high-performance computing resources," said Michael Churchill, head of PPPL's AI Science Group. The planned 20-minute turnaround is a development target, not a proven performance result.
Researchers still need to verify the accuracy of predictions under actual operating conditions.
Before applying the technology to SPARC, researchers will use PPPL's National Spherical Torus Experiment Upgrade (NSTX-U) to test and validate the digital twin workflow. The experimental platform will allow scientists to assess the accuracy of their models in reproducing plasma behavior and whether they can reliably support decision-making in fusion experiments. Data from NSTX-U and earlier PPPL experiments will also help train and evaluate these models. Some of these datasets will be made available to the broader fusion research community.
Partners and Long-Term Goals
In addition to PPPL, the collaboration includes Oak Ridge National Laboratory, Lawrence Livermore National Laboratory, and SLAC National Accelerator Laboratory. MIT, Carnegie Mellon University, General Atomics, NVIDIA, Google DeepMind, and Synopsys will also participate. The project builds on CFS's existing digital twin development work, expanding it through DOE-supported research and collaboration. Its long-term goal is to support ARC, the commercial fusion power plant CFS plans to develop after SPARC.
Researchers hope the digital twin will help them assess operating conditions, refine reactor designs, and reduce the time needed to develop future fusion systems.
However, SPARC remains under construction, and the new platform has yet to demonstrate its intended prediction speed or accuracy. Its ultimate contribution to commercial fusion energy will depend on successful testing and validation.
Potential Impact of Digital Twin on Fusion Research
This digital twin project marks a significant advance in fusion energy research because it can provide predictions of plasma behavior in a short timeframe, which is critical for improving reactor performance. Through connection with the physical device, researchers can adjust experimental designs more quickly, thereby accelerating fusion technology development. Although the accuracy of predictions still needs verification, the project's success could transform the future commercialization of fusion energy and have far-reaching implications for meeting global energy demand.