New ML framework predicts shifts between shots at DIII-D

[Marlin]

New ML framework predicts shifts between shots at DIII-D

https://www.ans.org/news/2026-08-18/article-8302/new-ml-framework-predicts-shifts-between-shots-at-diiid/

[Core Content]

Researchers at the DIII-D National Fusion Facility developed a machine learning framework that continuously adapts to shifting data from toroidal field coils. By utilizing deep neural networks to account for hardware drift between plasma shots, the team reduced prediction errors by 80 percent. This diagnostic tool aims to improve operational reliability and maintenance scheduling by identifying potential hardware issues before they occur during fusion experiments.

An artist's sketch and a cross-section view of the DIII-D tokamak. (Images: General Atomics)

At the DIII-D National Fusion Facility near San Diego, Calif., home to the largest operating tokamak in North America, researchers from Thomas Jefferson National Accelerator Facility worked to develop a machine learning framework capable of adaptively predicting changes in a tokamak's hardware.

The team's research was recently published in the journal Machine Learning with Applications.

Read the full article at ans.org:
https://www.ans.org/news/2026-08-18/article-8302/new-ml-framework-predicts-shifts-between-shots-at-diiid/