Skip to content

Latest commit

 

History

History
11 lines (9 loc) · 726 Bytes

README.md

File metadata and controls

11 lines (9 loc) · 726 Bytes

Tearing Avoidance Control

  • Tearing mode instability is the leading cause of plasma disruption in a tokamak.
  • This repository provides the Python scripts for training the ML models:
    (1) Tearing mode prediction model using deep learning
    (2) Tearing mode avoidance model using deep reinforcement learning

Note

  • Some scripts need experimental data from DIII-D, which are not available from this repository. To access DIII-D data, you should become a DIII-D user, per the instructions at https://d3dfusion.org/become-a-user/ .

References

  • J. Seo et al., "Avoiding fusion plasma tearing instability with deep reinforcement learning." Nature 626 (2024): 746.