Skip to main content
OSTI.GOVU.S. Department of Energy Office of Scientific and Technical Information
U.S. Department of Energy
Office of Scientific and Technical Information
 

Low latency optical-based mode tracking with machine learning deployed on FPGAs on a tokamak

Journal Article · · Review of Scientific Instruments
DOI:https://doi.org/10.1063/5.0190354· OSTI ID:2283683
 [1];  [2];  [1];  [1];  [3];  [4];  [5];  [1];  [1]
  1. Columbia Univ., New York, NY (United States)
  2. Fermi National Accelerator Laboratory (FNAL), Batavia, IL (United States); Lehigh Univ., Bethlehem, PA (United States)
  3. Fermi National Accelerator Laboratory (FNAL), Batavia, IL (United States); Northwestern Univ., Evanston, IL (United States)
  4. Drexel Univ., Philadelphia, PA (United States)
  5. Northwestern Univ., Evanston, IL (United States); Fermi National Accelerator Laboratory (FNAL), Batavia, IL (United States)
+ Show Author Affiliations
Active feedback control in magnetic confinement fusion devices is desirable to mitigate plasma instabilities and enable robust operation. Optical high-speed cameras provide a powerful, non-invasive diagnostic and can be suitable for these applications. Here, in this study, we process high-speed camera data, at rates exceeding 100 kfps, on in situ field-programmable gate array (FPGA) hardware to track magnetohydrodynamic (MHD) mode evolution and generate control signals in real time. Our system utilizes a convolutional neural network (CNN) model, which predicts the n = 1 MHD mode amplitude and phase using camera images with better accuracy than other tested non-deep-learning-based methods. By implementing this model directly within the standard FPGA readout hardware of the high-speed camera diagnostic, our mode tracking system achieves a total trigger-to-output latency of 17.6 μs and a throughput of up to 120 kfps. This study at the High Beta Tokamak-Extended Pulse (HBT-EP) experiment demonstrates an FPGA-based high-speed camera data acquisition and processing system, enabling application in real-time machine-learning-based tokamak diagnostic and control as well as potential applications in other scientific domains.
Research Organization:
Fermi National Accelerator Laboratory (FNAL), Batavia, IL (United States)
Sponsoring Organization:
National Science Foundation (NSF); USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR); USDOE Office of Science (SC), Fusion Energy Sciences (FES); USDOE Office of Science (SC), High Energy Physics (HEP)
Grant/Contract Number:
AC02-07CH11359; FG02-86ER53222; SC0021325; SC0022234
OSTI ID:
2283683
Report Number(s):
FERMILAB-PUB--23-655-CSAID; arXiv:2312.00128; oai:inspirehep.net:2729258
Journal Information:
Review of Scientific Instruments, Journal Name: Review of Scientific Instruments Journal Issue: 7 Vol. 95; ISSN 0034-6748
Publisher:
American Institute of Physics (AIP)Copyright Statement
Country of Publication:
United States
Language:
English

References (25)

Intelligent Image-Activated Cell Sorting journal September 2018
A Deep Reinforcement Learning approach for Vertical Stabilization of tokamak plasmas journal September 2023
Development of machine learning model for automatic ELM-burst detection without hyperparameter adjustment in KSTAR tokamak journal January 2023
A general infrastructure for data-driven control design and implementation in tokamaks journal January 2023
Deep learning journal May 2015
Predicting disruptive instabilities in controlled fusion plasmas through deep learning journal April 2019
Magnetic control of tokamak plasmas through deep reinforcement learning journal February 2022
Suppression of rotating external kink instabilities using optimized mode control feedback journal April 2005
A digital control system for external magnetohydrodynamic modes in tokamak plasmas journal April 2009
Stabilization of the external kink and the resistive wall mode journal October 2010
The high beta tokamak-extended pulse magnetohydrodynamic mode control research program journal May 2011
Adaptive control of rotating magnetic perturbations in HBT-EP using GPU processing journal June 2013
High-Speed imaging of the plasma response to resonant magnetic perturbations in HBT-EP journal March 2015
Improved feedback control of wall stabilized kink modes with different plasma–wall couplings and mode rotation journal January 2016
MHD mode tracking using high-speed cameras and deep learning journal May 2023
Prediction of electron density and pressure profile shapes on NSTX-U using neural networks journal March 2021
A dimensionality reduction algorithm for mapping tokamak operational regimes using a variational autoencoder (VAE) neural network journal November 2021
Preemptive RMP-driven ELM crash suppression automated by a real-time machine-learning classifier in KSTAR journal January 2022
Halo current rotation scaling in post-disruption plasmas journal January 2022
Predicting resistive wall mode stability in NSTX through balanced random forests and counterfactual explanations journal January 2022
Development of an operation trajectory design algorithm for control of multiple 0D parameters using deep reinforcement learning in KSTAR journal July 2022
Integrated deep learning framework for unstable event identification and disruption prediction of tokamak plasmas journal March 2023
Fast convolutional neural networks on FPGAs with hls4ml journal July 2021
Real-Time and Adaptive Reservoir Computing With Application to Profile Prediction in Fusion Plasma journal June 2022
Applications and Techniques for Fast Machine Learning in Science journal April 2022

Similar Records

Fast, multi-channel real-time processing of signals with microsecond latency using graphics processing units
Journal Article · Tue Apr 15 00:00:00 EDT 2014 · Review of Scientific Instruments · OSTI ID:22254934
MHD mode tracking using high-speed cameras and deep learning
Journal Article · Mon May 29 20:00:00 EDT 2023 · Plasma Physics and Controlled Fusion · OSTI ID:2325440

Related Subjects

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND TECHNOLOGY