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

Title: Disruption prediction investigations using Machine Learning tools on DIII-D and Alcator C-Mod

Abstract

Using data-driven methodology, we exploit the time series of relevant plasma parameters for a large set of disrupted and non-disrupted discharges to develop a classification algorithm for detecting disruptive phases in shots that eventually disrupt. Comparing the same methodology on different devices is crucial in order to have information on the portability of the developed algorithm and the possible extrapolation to ITER. Therefore, we use data from two very different tokamaks, DIII-D and Alcator C-Mod. We then focus on a subset of disruption predictors, most of which are dimensionless and/or machine-independent parameters, coming from both plasma diagnostics and equilibrium reconstructions, such as the normalized plasma internal inductance ℓ and the n = 1 mode amplitude normalized to the toroidal magnetic field. Using such dimensionless indicators facilitates a more direct comparison between DIII-D and C-Mod. We then choose a shallow Machine Learning technique, called Random Forests, to explore the databases available for the two devices. We show results from the classification task, where we introduce a time dependency through the definition of class labels on the basis of the elapsed time before the disruption (i.e. ‘far from a disruption’ and ‘close to a disruption’). The performances of the different Random Forestmore » classifiers are discussed in terms of several metrics, by showing the number of successfully detected samples, as well as the misclassifications. The overall model accuracies are above 97% when identifying a ‘far from disruption’ and a ‘disruptive’ phase for disrupted discharges. Nevertheless, the Forests are intrinsically different in their capability of predicting a disruptive behavior, with C-Mod predictions comparable to random guesses. Indeed, we show that C-Mod recall index, i.e. the sensitivity to a disruptive behavior, is as low as 0.47, while DIII-D recall is ~0.72. The portability of the developed algorithm is also tested across the two devices, by using DIII-D data for training the forests and C-Mod for testing and vice versa.« less

Authors:
ORCiD logo [1]; ORCiD logo [1]; ORCiD logo [1]; ORCiD logo [1]; ORCiD logo [2];  [3];  [2]
  1. Massachusetts Inst. of Technology (MIT), Cambridge, MA (United States). Plasma Science and Fusion Center (PSFC)
  2. General Atomics, San Diego, CA (United States)
  3. Columbia Univ., New York, NY (United States)
Publication Date:
Research Org.:
Massachusetts Inst. of Technology (MIT), Cambridge, MA (United States)
Sponsoring Org.:
USDOE Office of Science (SC), Fusion Energy Sciences (FES)
OSTI Identifier:
1454268
Grant/Contract Number:  
FC02-04ER54698; SC0014264; FG02-04ER54761
Resource Type:
Accepted Manuscript
Journal Name:
Plasma Physics and Controlled Fusion
Additional Journal Information:
Journal Volume: 60; Journal Issue: 8; Journal ID: ISSN 0741-3335
Publisher:
IOP Science
Country of Publication:
United States
Language:
English
Subject:
70 PLASMA PHYSICS AND FUSION TECHNOLOGY; 43 PARTICLE ACCELERATORS; cross-device study; machine learning; disruptions

Citation Formats

Rea, C., Granetz, R. S., Montes, K., Tinguely, R. A., Eidietis, N., Hanson, J. M., and Sammuli, B. Disruption prediction investigations using Machine Learning tools on DIII-D and Alcator C-Mod. United States: N. p., 2018. Web. doi:10.1088/1361-6587/aac7fe.
Rea, C., Granetz, R. S., Montes, K., Tinguely, R. A., Eidietis, N., Hanson, J. M., & Sammuli, B. Disruption prediction investigations using Machine Learning tools on DIII-D and Alcator C-Mod. United States. doi:10.1088/1361-6587/aac7fe.
Rea, C., Granetz, R. S., Montes, K., Tinguely, R. A., Eidietis, N., Hanson, J. M., and Sammuli, B. Mon . "Disruption prediction investigations using Machine Learning tools on DIII-D and Alcator C-Mod". United States. doi:10.1088/1361-6587/aac7fe. https://www.osti.gov/servlets/purl/1454268.
@article{osti_1454268,
title = {Disruption prediction investigations using Machine Learning tools on DIII-D and Alcator C-Mod},
author = {Rea, C. and Granetz, R. S. and Montes, K. and Tinguely, R. A. and Eidietis, N. and Hanson, J. M. and Sammuli, B.},
abstractNote = {Using data-driven methodology, we exploit the time series of relevant plasma parameters for a large set of disrupted and non-disrupted discharges to develop a classification algorithm for detecting disruptive phases in shots that eventually disrupt. Comparing the same methodology on different devices is crucial in order to have information on the portability of the developed algorithm and the possible extrapolation to ITER. Therefore, we use data from two very different tokamaks, DIII-D and Alcator C-Mod. We then focus on a subset of disruption predictors, most of which are dimensionless and/or machine-independent parameters, coming from both plasma diagnostics and equilibrium reconstructions, such as the normalized plasma internal inductance ℓ and the n = 1 mode amplitude normalized to the toroidal magnetic field. Using such dimensionless indicators facilitates a more direct comparison between DIII-D and C-Mod. We then choose a shallow Machine Learning technique, called Random Forests, to explore the databases available for the two devices. We show results from the classification task, where we introduce a time dependency through the definition of class labels on the basis of the elapsed time before the disruption (i.e. ‘far from a disruption’ and ‘close to a disruption’). The performances of the different Random Forest classifiers are discussed in terms of several metrics, by showing the number of successfully detected samples, as well as the misclassifications. The overall model accuracies are above 97% when identifying a ‘far from disruption’ and a ‘disruptive’ phase for disrupted discharges. Nevertheless, the Forests are intrinsically different in their capability of predicting a disruptive behavior, with C-Mod predictions comparable to random guesses. Indeed, we show that C-Mod recall index, i.e. the sensitivity to a disruptive behavior, is as low as 0.47, while DIII-D recall is ~0.72. The portability of the developed algorithm is also tested across the two devices, by using DIII-D data for training the forests and C-Mod for testing and vice versa.},
doi = {10.1088/1361-6587/aac7fe},
journal = {Plasma Physics and Controlled Fusion},
number = 8,
volume = 60,
place = {United States},
year = {2018},
month = {6}
}

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

Citation Metrics:
Cited by: 4 works
Citation information provided by
Web of Science

Figures / Tables:

Table 1 Table 1: List of signals considered for Machine Learning applications on DⅢ-D and C-Mod.

Save / Share:

Works referenced in this record:

Integrated modeling applications for tokamak experiments with OMFIT
journal, July 2015


Selecting and interpreting measures of thematic classification accuracy
journal, October 1997


Reconstruction of current profile parameters and plasma shapes in tokamaks
journal, November 1985


The MDSplus data acquisition system, current status and future directions
journal, January 1999


Advances in lower hybrid current drive technology on Alcator C-Mod
journal, June 2013


Neoclassical tearing modes
journal, December 2000


Neural network prediction of some classes of tokamak disruptions
journal, August 1996


Survey of disruption causes at JET
journal, April 2011


Statistical analysis of m / n   =  2/1 locked and quasi-stationary modes with rotating precursors at DIII-D
journal, November 2016


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


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


A reduced resistive wall mode kinetic stability model for disruption forecasting
journal, May 2017

  • Berkery, J. W.; Sabbagh, S. A.; Bell, R. E.
  • Physics of Plasmas, Vol. 24, Issue 5
  • DOI: 10.1063/1.4977464

Advanced tokamak research in DIII-D
journal, November 2004


Disruption forecasting at JET using neural networks
journal, December 2003


Achievement of Sustained Net Plasma Heating in a Fusion Experiment with the Optometrist Algorithm
journal, July 2017


Chapter 3: MHD stability, operational limits and disruptions
journal, December 1999

  • Mhd, ITER Physics Expert Group on Disrup; Editors, ITER Physics Basis
  • Nuclear Fusion, Vol. 39, Issue 12
  • DOI: 10.1088/0029-5515/39/12/303

Support vector machines for disruption prediction and novelty detection at JET
journal, October 2007


Magnetic diagnostic system of the DIII-D tokamak
journal, February 2006

  • Strait, E. J.
  • Review of Scientific Instruments, Vol. 77, Issue 2
  • DOI: 10.1063/1.2166493

Random Forests
journal, January 2001


Critical error fields for locked mode instability in tokamaks
journal, March 1992

  • La Haye, R. J.; Fitzpatrick, R.; Hender, T. C.
  • Physics of Fluids B: Plasma Physics, Vol. 4, Issue 7
  • DOI: 10.1063/1.860017

Forecast of TEXT plasma disruptions using soft X rays as input signal in a neural network
journal, February 1999


Neural-net disruption predictor in JT-60U
journal, December 2003


Nonaxisymmetric field effects on Alcator C-Mod
journal, May 2005

  • Wolfe, S. M.; Hutchinson, I. H.; Granetz, R. S.
  • Physics of Plasmas, Vol. 12, Issue 5
  • DOI: 10.1063/1.1883665

Exploratory Machine Learning Studies for Disruption Prediction Using Large Databases on DIII-D
journal, February 2018


Bagging predictors
journal, August 1996


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


On-line prediction and mitigation of disruptions in ASDEX Upgrade
journal, January 2002


A prediction tool for real-time application in the disruption protection system at JET
journal, October 2007


A review of feature selection techniques in bioinformatics
journal, August 2007


TokSearch: A search engine for fusion experimental data
journal, April 2018


Automatic disruption classification at JET: comparison of different pattern recognition techniques
journal, May 2006


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


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


Neural network prediction of some classes of tokamak disruptions
journal, August 1996


Random Forests
journal, January 2001


The MDSplus data acquisition system, current status and future directions
journal, January 1999


Neoclassical tearing modes
journal, December 2000


Bagging predictors
journal, August 1996


Selecting and interpreting measures of thematic classification accuracy
journal, October 1997


Advanced tokamak research in DIII-D
journal, November 2004


    Works referencing / citing this record:

    Machine learning control for disruption and tearing mode avoidance
    journal, February 2020

    • Fu, Yichen; Eldon, David; Erickson, Keith
    • Physics of Plasmas, Vol. 27, Issue 2
    • DOI: 10.1063/1.5125581