Multi-Timescale Dynamics Model Bayesian Optimization for Plasma Stabilization in Tokamaks

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Sonker, Rohit, Capone, Alexandre, Rothstein, Andrew, Kaga, Hiro Josep Farre, Kolemen, Egemen, Schneider, Jeff
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911001642795008
author Sonker, Rohit
Capone, Alexandre
Rothstein, Andrew
Kaga, Hiro Josep Farre
Kolemen, Egemen
Schneider, Jeff
author_facet Sonker, Rohit
Capone, Alexandre
Rothstein, Andrew
Kaga, Hiro Josep Farre
Kolemen, Egemen
Schneider, Jeff
contents Machine learning algorithms often struggle to control complex real-world systems. In the case of nuclear fusion, these challenges are exacerbated, as the dynamics are notoriously complex, data is poor, hardware is subject to failures, and experiments often affect dynamics beyond the experiment's duration. Existing tools like reinforcement learning, supervised learning, and Bayesian optimization address some of these challenges but fail to provide a comprehensive solution. To overcome these limitations, we present a multi-scale Bayesian optimization approach that integrates a high-frequency data-driven dynamics model with a low-frequency Gaussian process. By updating the Gaussian process between experiments, the method rapidly adapts to new data, refining the predictions of the less reliable dynamical model. We validate our approach by controlling tearing instabilities in the DIII-D nuclear fusion plant. Offline testing on historical data shows that our method significantly outperforms several baselines. Results on live experiments on the DIII-D tokamak, conducted under high-performance plasma scenarios prone to instabilities, shows a 50% success rate, marking a 117% improvement over historical outcomes.
format Preprint
id arxiv_https___arxiv_org_abs_2506_10287
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Multi-Timescale Dynamics Model Bayesian Optimization for Plasma Stabilization in Tokamaks
Sonker, Rohit
Capone, Alexandre
Rothstein, Andrew
Kaga, Hiro Josep Farre
Kolemen, Egemen
Schneider, Jeff
Robotics
Machine learning algorithms often struggle to control complex real-world systems. In the case of nuclear fusion, these challenges are exacerbated, as the dynamics are notoriously complex, data is poor, hardware is subject to failures, and experiments often affect dynamics beyond the experiment's duration. Existing tools like reinforcement learning, supervised learning, and Bayesian optimization address some of these challenges but fail to provide a comprehensive solution. To overcome these limitations, we present a multi-scale Bayesian optimization approach that integrates a high-frequency data-driven dynamics model with a low-frequency Gaussian process. By updating the Gaussian process between experiments, the method rapidly adapts to new data, refining the predictions of the less reliable dynamical model. We validate our approach by controlling tearing instabilities in the DIII-D nuclear fusion plant. Offline testing on historical data shows that our method significantly outperforms several baselines. Results on live experiments on the DIII-D tokamak, conducted under high-performance plasma scenarios prone to instabilities, shows a 50% success rate, marking a 117% improvement over historical outcomes.
title Multi-Timescale Dynamics Model Bayesian Optimization for Plasma Stabilization in Tokamaks
topic Robotics
url https://arxiv.org/abs/2506.10287