Hybrid deep learning architecture for general disruption prediction across tokamaks

Zhu, Jinxiang; Rea, Cristina; Montes, Kevin Joseph; Granetz, Robert; Sweeney, Ryan; Tinguely, Roy Alexander

doi:10.1088/1741-4326/abc664

Hybrid deep learning architecture for general disruption prediction across tokamaks

Journal Article · Fri Oct 30 00:00:00 EDT 2020 · Nuclear Fusion

DOI:https://doi.org/10.1088/1741-4326/abc664· OSTI ID:1735947

Zhu, Jinxiang ^[1]; ^[2]; ^[2]; Granetz, Robert ^[2]; ^[2]; ^[2]

Massachusetts Inst. of Technology (MIT), Cambridge, MA (United States); General Atomics, Energy & Advanced Concepts, DIII-D
Massachusetts Inst. of Technology (MIT), Cambridge, MA (United States)

In this paper, we present a new deep learning disruption prediction algorithm based on important findings from explorative data analysis which effectively allows knowledge transfer from existing devices to new ones, thereby predicting disruptions using very limited disruptive data from the new devices. Here, the explorative data analysis conducted via unsupervised clustering techniques confirms that time-sequence data are much better separators of disruptive and non-disruptive behavior than the instantaneous plasma state data with further advantageous implications for a sequence-based predictor. Based on such important findings, we have designed a new algorithm for multi-machine disruption prediction that achieves high predictive accuracy on the C-Mod (AUC=0.801), DIII-D (AUC=0.947) and EAST (AUC=0.973) tokamaks with limited hyperparameter tuning. Through numerical experiments, we show that boosted accuracy (AUC=0.959) is achieved on EAST predictions by including in the training only 20 disruptive discharges, thousands of non-disruptive discharges from EAST, and combining this with more than a thousand discharges from DIII-D and C-Mod. The improvement of predictive ability obtained by combining disruptive data from other devices is found to be true for all permutations of the three devices. Furthermore, by comparing the predictive performance of each individual numerical experiment, we find that non-disruptive data are machine-specific while disruptive data from multiple devices contain device-independent knowledge that can be used to inform predictions for disruptions occurring on a new device.

View Accepted Manuscript (DOE)

Research Organization:: General Atomics, San Diego, CA (United States)

Sponsoring Organization:: USDOE Office of Science (SC), Fusion Energy Sciences (FES)

Grant/Contract Number:: FC02-04ER54698; SC0014264

OSTI ID:: 1735947

Alternate ID(s):: OSTI ID: 23129597

Journal Information:: Nuclear Fusion, Journal Name: Nuclear Fusion Journal Issue: 2 Vol. 61; ISSN 0029-5515

Publisher:: IOP ScienceCopyright Statement

Country of Publication:: United States

Language:: English

References (30)

Results of the JET real-time disruption predictor in the ITER-like wall campaigns Vega, Jesús; Dormido-Canto, Sebastián; López, Juan M. Fusion Engineering and Design, Vol. 88, Issue 6-8 https://doi.org/10.1016/j.fusengdes.2013.03.003	journal	October 2013
The use of the area under the ROC curve in the evaluation of machine learning algorithms Bradley, Andrew P. Pattern Recognition, Vol. 30, Issue 7, p. 1145-1159 https://doi.org/10.1016/s0031-3203(96)00142-2	journal	July 1997
Predicting disruptive instabilities in controlled fusion plasmas through deep learning Kates-Harbeck, Julian; Svyatkovskiy, Alexey; Tang, William Nature, Vol. 568, Issue 7753 https://doi.org/10.1038/s41586-019-1116-4	journal	April 2019
Status of research toward the ITER disruption mitigation system Hollmann, E. M.; Aleynikov, P. B.; Fülöp, T. Physics of Plasmas, Vol. 22, Issue 2 https://doi.org/10.1063/1.4901251	journal	November 2014
A reduced resistive wall mode kinetic stability model for disruption forecasting Berkery, J. W.; Sabbagh, S. A.; Bell, R. E. Physics of Plasmas, Vol. 24, Issue 5 https://doi.org/10.1063/1.4977464	journal	May 2017
Viability Assessment of a Cross-Tokamak AUG-JET Disruption Predictor Rattá, G. A.; Vega, J.; Murari, A. Fusion Science and Technology, Vol. 74, Issue 1-2 https://doi.org/10.1080/15361055.2017.1390390	journal	February 2018
Tokamak disruption alarm based on a neural network model of the high- beta limit Wroblewski, D.; Jahns, G. L.; Leuer, J. A. Nuclear Fusion, Vol. 37, Issue 6 https://doi.org/10.1088/0029-5515/37/6/i02	journal	June 1997
Chapter 7: Measurement of plasma parameters Diagnostics, ITER Physics Expert Group on; Editors, ITER Physics Basis Nuclear Fusion, Vol. 39, Issue 12 https://doi.org/10.1088/0029-5515/39/12/307	journal	December 1999
Disruption forecasting at JET using neural networks Cannas, B.; Fanni, A.; Marongiu, E. Nuclear Fusion, Vol. 44, Issue 1 https://doi.org/10.1088/0029-5515/44/1/008	journal	December 2003
Status report on fusion research Burkart, Werner Nuclear Fusion, Vol. 45, Issue 10A https://doi.org/10.1088/0029-5515/45/10a/e01	journal	September 2005
A prediction tool for real-time application in the disruption protection system at JET Cannas, B.; Fanni, A.; Sonato, P. Nuclear Fusion, Vol. 47, Issue 11 https://doi.org/10.1088/0029-5515/47/11/018	journal	October 2007
Progress in the ITER Physics Basis Ikeda, K. Nuclear Fusion, Vol. 47, Issue 6 https://doi.org/10.1088/0029-5515/47/6/E01	journal	June 2007
Prototype of an adaptive disruption predictor for JET based on fuzzy logic and regression trees Murari, A.; Vagliasindi, G.; Arena, P. Nuclear Fusion, Vol. 48, Issue 3 https://doi.org/10.1088/0029-5515/48/3/035010	journal	February 2008
Survey of disruption causes at JET de Vries, P. C.; Johnson, M. F.; Alper, B. Nuclear Fusion, Vol. 51, Issue 5 https://doi.org/10.1088/0029-5515/51/5/053018	journal	April 2011
Dynamics of the disruption halo current toroidal asymmetry in NSTX Gerhardt, S. P. Nuclear Fusion, Vol. 53, Issue 2 https://doi.org/10.1088/0029-5515/53/2/023005	journal	January 2013
Integrated modeling applications for tokamak experiments with OMFIT Meneghini, O.; Smith, S. P.; Lao, L. L. Nuclear Fusion, Vol. 55, Issue 8 https://doi.org/10.1088/0029-5515/55/8/083008	journal	July 2015
Statistical analysis of m / n = 2/1 locked and quasi-stationary modes with rotating precursors at DIII-D Sweeney, R.; Choi, W.; La Haye, R. J. Nuclear Fusion, Vol. 57, Issue 1 https://doi.org/10.1088/0029-5515/57/1/016019	journal	November 2016
Preliminary R&D on flat-type W/Cu plasma-facing materials and components for Experimental Advanced Superconducting Tokamak Xu, Y.; Xie, C. Y.; Qin, S. G. Physica Scripta, Vol. T159 https://doi.org/10.1088/0031-8949/2014/T159/014008	journal	April 2014
Path-oriented early reaction to approaching disruptions in ASDEX Upgrade and TCV in view of the future needs for ITER and DEMO Maraschek, M.; Gude, A.; Igochine, V. Plasma Physics and Controlled Fusion, Vol. 60, Issue 1 https://doi.org/10.1088/1361-6587/aa8d05	journal	November 2017
Disruption prediction investigations using Machine Learning tools on DIII-D and Alcator C-Mod Rea, C.; Granetz, R. S.; Montes, K. Plasma Physics and Controlled Fusion, Vol. 60, Issue 8 https://doi.org/10.1088/1361-6587/aac7fe	journal	June 2018
Adaptive predictors based on probabilistic SVM for real time disruption mitigation on JET Murari, A.; Lungaroni, M.; Peluso, E. Nuclear Fusion, Vol. 58, Issue 5 https://doi.org/10.1088/1741-4326/aaaf9c	journal	March 2018
The causes of the disruptive tearing instabilities of the ITER Baseline Scenario in DIII-D Turco, F.; Luce, T. C.; Solomon, W. Nuclear Fusion, Vol. 58, Issue 10 https://doi.org/10.1088/1741-4326/aadbb5	journal	September 2018
Machine learning for disruption warnings on Alcator C-Mod, DIII-D, and EAST Montes, K. J.; Rea, C.; Granetz, R. S. Nuclear Fusion, Vol. 59, Issue 9 https://doi.org/10.1088/1741-4326/ab1df4	journal	July 2019
A real-time machine learning-based disruption predictor in DIII-D Rea, C.; Montes, K. J.; Erickson, K. G. Nuclear Fusion, Vol. 59, Issue 9 https://doi.org/10.1088/1741-4326/ab28bf	journal	July 2019
A machine learning approach based on generative topographic mapping for disruption prevention and avoidance at JET Pau, A.; Fanni, A.; Carcangiu, S. Nuclear Fusion, Vol. 59, Issue 10 https://doi.org/10.1088/1741-4326/ab2ea9	journal	August 2019
Physics-guided machine learning approaches to predict the ideal stability properties of fusion plasmas Piccione, A.; Berkery, J. W.; Sabbagh, S. A. Nuclear Fusion, Vol. 60, Issue 4 https://doi.org/10.1088/1741-4326/ab7597	journal	March 2020
Neural network ensembles Hansen, L. K.; Salamon, P. IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 12, Issue 10 https://doi.org/10.1109/34.58871	journal	January 1990
Fully Character-Level Neural Machine Translation without Explicit Segmentation Lee, Jason; Cho, Kyunghyun; Hofmann, Thomas Transactions of the Association for Computational Linguistics, Vol. 5 https://doi.org/10.1162/tacl_a_00067	journal	December 2017
Requirements for Triggering the ITER Disruption Mitigation System de Vries, P. C.; Pautasso, G.; Humphreys, D. Fusion Science and Technology, Vol. 69, Issue 2 https://doi.org/10.13182/fst15-176	journal	April 2016
On the Properties of Neural Machine Translation: Encoder–Decoder Approaches Cho, Kyunghyun; van Merrienboer, Bart; Bahdanau, Dzmitry Proceedings of SSST-8, Eighth Workshop on Syntax, Semantics and Structure in Statistical Translation https://doi.org/10.3115/v1/W14-4012	conference	January 2014

Similar Records

Hybrid deep learning architecture for general disruption prediction across tokamaks

Dataset · Fri Jun 25 00:00:00 EDT 2021 · OSTI ID:1886671

Integrated deep learning framework for unstable event identification and disruption prediction of tokamak plasmas

Journal Article · Tue Feb 28 23:00:00 EST 2023 · Nuclear Fusion · OSTI ID:1959076

Related Subjects

70 PLASMA PHYSICS AND FUSION TECHNOLOGY

Hybrid deep learning architecture for general disruption prediction across tokamaks

Citation Formats

References (30)

Similar Records

Related Subjects