DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: A real-time machine learning-based disruption predictor in DIII-D

Journal Article · · Nuclear Fusion

A disruption prediction algorithm, named DPRF (Disruption Prediction using Random Forests), has run in real-time in the DIII-D plasma control system (PCS) over more than 900 discharges. DPRF naturally provides a probability mapping associated to its predictions, i.e. the disruptivity signal, now incorporated in the DIII-D PCS. Here, we report on disruption prediction accomplishments in terms of shot-by-shot performances, by simulating alarms on each discharge as in the PCS framework. Due to the optimised performance metric chosen to evaluate DPRF, we find that almost all disruptive discharges are detected in average with few hundred of milliseconds of warning time, but this comes at a high cost of false alarms produced. Performances are not satisfying ITER requirements, where the success rate has to be higher than 95%, but this is not completely unexpected. DPRF is trained on many years of major disruptions occurring during the flattop phase of the plasma current on DIII-D, but without any differentiation by cause. Furthermore, we find that DPRF is affected by a relatively high fraction of false alarms occurring during the first 500 milliseconds from the flattop onset. This subtle effect, more evident on the discharges where DPRF ran in real-time, can be marginalized by taking specific precautions on the validity range of the predictions, and performances do improve. Even if presently burdened by some limitations, DPRF provides an incredible and novel advantage. Thanks to the feature contribution analysis (e.g., the identification of which signals contributed to triggering an alarm), it is possible to interpret and explain DPRF predictions. It is the first time that such interpretability features is exploited by a disruption predictor: by uncovering the disruption events' causes, a better understanding of disruption dynamics is achieved, and a clear path toward the design of disruption avoidance strategies can be provided.

Research Organization:
General Atomics, San Diego, CA (United States); Massachusetts Inst. of Technology (MIT), Cambridge, MA (United States); Princeton Univ., NJ (United States)
Sponsoring Organization:
USDOE Office of Science (SC), Fusion Energy Sciences (FES) (SC-24)
Grant/Contract Number:
FC02-04ER54698; SC0014264; AC02-09CH11466
OSTI ID:
1562300
Alternate ID(s):
OSTI ID: 22929898
Report Number(s):
DOE-GA--54698
Journal Information:
Nuclear Fusion, Journal Name: Nuclear Fusion Journal Issue: 9 Vol. 59; ISSN 0029-5515
Publisher:
IOP ScienceCopyright Statement
Country of Publication:
United States
Language:
English

References (26)

Random Forests journal January 2001
Real time equilibrium reconstruction for tokamak discharge control journal July 1998
Development of an efficient real-time disruption predictor from scratch on JET and implications for ITER journal September 2013
Unbiased and non-supervised learning methods for disruption prediction at JET journal April 2009
A cross-tokamak neural network disruption predictor for the JET and ASDEX Upgrade tokamaks journal April 2005
A reduced resistive wall mode kinetic stability model for disruption forecasting journal May 2017
Integrated modeling applications for tokamak experiments with OMFIT journal July 2015
The Elements of Statistical Learning book January 2009
The Elements of Statistical Learning book January 2009
Magnetic diagnostic system of the DIII-D tokamak journal February 2006
Exploratory Machine Learning Studies for Disruption Prediction Using Large Databases on DIII-D journal February 2018
A First Analysis of JET Plasma Profile-Based Indicators for Disruption Prediction and Avoidance journal July 2018
miRNALoc: predicting miRNA subcellular localizations based on principal component scores of physico-chemical properties and pseudo compositions of di-nucleotides journal September 2020
Automatic disruption classification in JET with the ITER-like wall journal October 2015
Multivariate statistical models for disruption prediction at ASDEX Upgrade journal October 2013
Machine learning for disruption warnings on Alcator C-Mod, DIII-D, and EAST journal July 2019
Advanced Disruption Predictor Based On The Locked Mode Signal: Application To Jet conference October 2016
Survey of disruption causes at JET journal April 2011
Results of the JET real-time disruption predictor in the ITER-like wall campaigns journal October 2013
Implementing a finite-state off-normal and fault response system for disruption avoidance in tokamaks journal March 2018
The Elements of Statistical Learning book January 2001
Optimization of probiotic therapeutics using machine learning in an artificial human gastrointestinal tract journal January 2021
Improvements in disruption prediction at ASDEX Upgrade journal October 2015
An advanced disruption predictor for JET tested in a simulated real-time environment journal January 2010
Tokamak disruption alarm based on a neural network model of the high- beta limit journal June 1997
Advanced tokamak research in DIII-D journal November 2004