skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Predicting disruptive instabilities in controlled fusion plasmas through deep learning

Abstract

Nuclear fusion power delivered by magnetic-confinement tokamak reactors holds the promise of sustainable and clean energy1. The avoidance of large-scale plasma instabilities called disruptions within these reactors is one of the most pressing challenges, because disruptions can halt power production and damage key components. Disruptions are particularly harmful for large burning-plasma systems such as the multibillion-dollar International Thermonuclear Experimental Reactor (ITER) project currently under construction, which aims to be the first reactor that produces more power from fusion than is injected to heat the plasma. Here we present a method based on deep learning for forecasting disruptions. Our method extends considerably the capabilities of previous strategies such as first-principles-based and classical machine-learning approaches. In particular, it delivers reliable predictions for machines other than the one on which it was trained—a crucial requirement for future large reactors that cannot afford training disruptions. Our approach takes advantage of high-dimensional training data to boost predictive performance while also engaging supercomputing resources at the largest scale to improve accuracy and speed. Trained on experimental data from the largest tokamaks in the United States (DIII-D) and the world (Joint European Torus, JET), our method can also be applied to specific tasks such as prediction withmore » long warning times: this opens up the possibility of moving from passive disruption prediction to active reactor control and optimization. These initial results illustrate the potential for deep learning to accelerate progress in fusion-energy science and, more generally, in the understanding and prediction of complex physical systems.« less

Authors:
 [1];  [2];  [3]
  1. Harvard Univ., Cambridge, MA (United States); Princeton Plasma Physics Lab. (PPPL), Princeton, NJ (United States)
  2. Princeton Univ., NJ (United States)
  3. Princeton Plasma Physics Lab. (PPPL), Princeton, NJ (United States); Princeton Univ., NJ (United States)
Publication Date:
Research Org.:
Princeton Plasma Physics Lab. (PPPL), Princeton, NJ (United States); General Atomics, San Diego, CA (United States)
Sponsoring Org.:
USDOE Office of Science (SC), Fusion Energy Sciences (FES) (SC-24)
OSTI Identifier:
1507546
Alternate Identifier(s):
OSTI ID: 1566845
Grant/Contract Number:  
FG02-97ER25308; FC02-04ER54698
Resource Type:
Journal Article: Accepted Manuscript
Journal Name:
Nature (London)
Additional Journal Information:
Journal Volume: 568; Journal Issue: 7753; Journal ID: ISSN 0028-0836
Publisher:
Nature Publishing Group
Country of Publication:
United States
Language:
English
Subject:
70 PLASMA PHYSICS AND FUSION TECHNOLOGY

Citation Formats

Kates-Harbeck, Julian, Svyatkovskiy, Alexey, and Tang, William. Predicting disruptive instabilities in controlled fusion plasmas through deep learning. United States: N. p., 2019. Web. doi:10.1038/s41586-019-1116-4.
Kates-Harbeck, Julian, Svyatkovskiy, Alexey, & Tang, William. Predicting disruptive instabilities in controlled fusion plasmas through deep learning. United States. doi:10.1038/s41586-019-1116-4.
Kates-Harbeck, Julian, Svyatkovskiy, Alexey, and Tang, William. Wed . "Predicting disruptive instabilities in controlled fusion plasmas through deep learning". United States. doi:10.1038/s41586-019-1116-4. https://www.osti.gov/servlets/purl/1507546.
@article{osti_1507546,
title = {Predicting disruptive instabilities in controlled fusion plasmas through deep learning},
author = {Kates-Harbeck, Julian and Svyatkovskiy, Alexey and Tang, William},
abstractNote = {Nuclear fusion power delivered by magnetic-confinement tokamak reactors holds the promise of sustainable and clean energy1. The avoidance of large-scale plasma instabilities called disruptions within these reactors is one of the most pressing challenges, because disruptions can halt power production and damage key components. Disruptions are particularly harmful for large burning-plasma systems such as the multibillion-dollar International Thermonuclear Experimental Reactor (ITER) project currently under construction, which aims to be the first reactor that produces more power from fusion than is injected to heat the plasma. Here we present a method based on deep learning for forecasting disruptions. Our method extends considerably the capabilities of previous strategies such as first-principles-based and classical machine-learning approaches. In particular, it delivers reliable predictions for machines other than the one on which it was trained—a crucial requirement for future large reactors that cannot afford training disruptions. Our approach takes advantage of high-dimensional training data to boost predictive performance while also engaging supercomputing resources at the largest scale to improve accuracy and speed. Trained on experimental data from the largest tokamaks in the United States (DIII-D) and the world (Joint European Torus, JET), our method can also be applied to specific tasks such as prediction with long warning times: this opens up the possibility of moving from passive disruption prediction to active reactor control and optimization. These initial results illustrate the potential for deep learning to accelerate progress in fusion-energy science and, more generally, in the understanding and prediction of complex physical systems.},
doi = {10.1038/s41586-019-1116-4},
journal = {Nature (London)},
issn = {0028-0836},
number = 7753,
volume = 568,
place = {United States},
year = {2019},
month = {4}
}

Journal Article:
Free Publicly Available Full Text
Publisher's Version of Record

Citation Metrics:
Cited by: 10 works
Citation information provided by
Web of Science

Save / Share:

Works referenced in this record:

Commissioning of electron cyclotron emission imaging instrument on the DIII-D tokamak and first data
journal, October 2010

  • Tobias, B.; Domier, C. W.; Liang, T.
  • Review of Scientific Instruments, Vol. 81, Issue 10
  • DOI: 10.1063/1.3460456

Statistical analysis of disruptions in JET
journal, April 2009


Deep learning
journal, May 2015

  • LeCun, Yann; Bengio, Yoshua; Hinton, Geoffrey
  • Nature, Vol. 521, Issue 7553
  • DOI: 10.1038/nature14539

Fast visible camera installation and operation in JET
conference, January 2008

  • Alonso, J. A.; Andrew, P.; Neto, A.
  • 2007, AIP Conference Proceedings
  • DOI: 10.1063/1.2905063

Advances and challenges in computational plasma science
journal, January 2005


Human-level control through deep reinforcement learning
journal, February 2015

  • Mnih, Volodymyr; Kavukcuoglu, Koray; Silver, David
  • Nature, Vol. 518, Issue 7540
  • DOI: 10.1038/nature14236

Accelerating recurrent neural network training using sequence bucketing and multi-GPU data parallelization
conference, August 2016

  • Khomenko, Viacheslav; Shyshkov, Oleg; Radyvonenko, Olga
  • 2016 IEEE First International Conference on Data Stream Mining & Processing (DSMP)
  • DOI: 10.1109/DSMP.2016.7583516

Survey of disruption causes at JET
journal, April 2011


Cost-sensitive learning by cost-proportionate example weighting
conference, January 2003


Simultaneous closed-loop control of the current profile and the electron temperature profile in the TCV tokamak
conference, July 2015

  • Barton, Justin E.; Wehner, William P.; Schuster, Eugenio
  • 2015 American Control Conference (ACC)
  • DOI: 10.1109/ACC.2015.7171844

Disruption Prediction on JET during the ILW Experimental Campaigns
journal, April 2016

  • Moreno, R.; Vega, J.; Dormido-Canto, S.
  • Fusion Science and Technology, Vol. 69, Issue 2
  • DOI: 10.13182/FST15-167

Disruption mitigation studies in DIII-D
journal, May 1999

  • Taylor, P. L.; Kellman, A. G.; Evans, T. E.
  • Physics of Plasmas, Vol. 6, Issue 5
  • DOI: 10.1063/1.873445

Real time equilibrium reconstruction for tokamak discharge control
journal, July 1998


JET ITER-like wall—overview and experimental programme
journal, December 2011


Overview of the JET results in support to ITER
journal, June 2017


Prototype of an adaptive disruption predictor for JET based on fuzzy logic and regression trees
journal, February 2008


Training distributed deep recurrent neural networks with mixed precision on GPU clusters
conference, January 2017

  • Svyatkovskiy, Alexey; Kates-Harbeck, Julian; Tang, William
  • Proceedings of the Machine Learning on HPC Environments - MLHPC'17
  • DOI: 10.1145/3146347.3146358

Disruption forecasting at JET using neural networks
journal, December 2003


Improved feature selection based on genetic algorithms for real time disruption prediction on JET
journal, September 2012


A design retrospective of the DIII-D tokamak
journal, May 2002


ImageNet: A large-scale hierarchical image database
conference, June 2009

  • Deng, Jia; Dong, Wei; Socher, Richard
  • 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops (CVPR Workshops), 2009 IEEE Conference on Computer Vision and Pattern Recognition
  • DOI: 10.1109/CVPR.2009.5206848

The use of the area under the ROC curve in the evaluation of machine learning algorithms
journal, July 1997


Disruptions in tokamaks
journal, November 1995


The ITER design
journal, April 2002


XGBoost: A Scalable Tree Boosting System
conference, January 2016

  • Chen, Tianqi; Guestrin, Carlos
  • Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining - KDD '16
  • DOI: 10.1145/2939672.2939785

Cyclical Learning Rates for Training Neural Networks
conference, March 2017

  • Smith, Leslie N.
  • 2017 IEEE Winter Conference on Applications of Computer Vision (WACV)
  • DOI: 10.1109/WACV.2017.58

Results of the JET real-time disruption predictor in the ITER-like wall campaigns
journal, October 2013


Disruptions in ITER and strategies for their control and mitigation
journal, August 2015


Requirements for Triggering the ITER Disruption Mitigation System
journal, April 2016

  • de Vries, P. C.; Pautasso, G.; Humphreys, D.
  • Fusion Science and Technology, Vol. 69, Issue 2
  • DOI: 10.13182/FST15-176

Tokamak disruption alarm based on a neural network model of the high- beta limit
journal, June 1997


Feature extraction for improved disruption prediction analysis at JET
journal, October 2008

  • Rattá, G. A.; Vega, J.; Murari, A.
  • Review of Scientific Instruments, Vol. 79, Issue 10
  • DOI: 10.1063/1.2965020

The Power of an Idea: The International Impacts of the Grand Challenges for Engineering
journal, March 2016


Solving the quantum many-body problem with artificial neural networks
journal, February 2017


Machine learning phases of matter
journal, February 2017

  • Carrasquilla, Juan; Melko, Roger G.
  • Nature Physics, Vol. 13, Issue 5
  • DOI: 10.1038/nphys4035

A cross-tokamak neural network disruption predictor for the JET and ASDEX Upgrade tokamaks
journal, April 2005