TorbeamNN: machine learning-based steering of ECH mirrors on KSTAR

Andrew Rothstein; Minseok Kim; Minho Woo; Minsoo Cha; Cheolsik Byun; Sangkyeun Kim; Keith Erickson; Youngho Lee; Josh Josephy-Zack; Jalal Butt; Ricardo Shousha; Mi Joung; June Woo Juhn; Kyu Dong Lee; Egemen Kolemen

doi:10.1088/1361-6587/add08b

TorbeamNN: machine learning-based steering of ECH mirrors on KSTAR

Andrew Rothstein, Minseok Kim, Minho Woo, Minsoo Cha, Cheolsik Byun, Sangkyeun Kim, Keith Erickson, Youngho Lee, Josh Josephy-Zack, Jalal Butt, Ricardo Shousha, Mi Joung, June Woo Juhn, Kyu Dong Lee, Egemen Kolemen

Research output: Contribution to journal › Article › peer-review

Abstract

We have developed TorbeamNN: a machine learning surrogate model for the TORBEAM ray tracing code to predict electron cyclotron heating (ECH) and current drive locations in tokamak plasmas. TorbeamNN provides more than a 100 times speed-up compared to the highly optimized and simplified real-time implementation of TORBEAM without any reduction in accuracy compared to the offline, full fidelity TORBEAM code. The model was trained using KSTAR ECH mirror geometries and works for both O-mode and X-mode absorption. The TorbeamNN predictions have been validated both offline and real-time in experiment. TorbeamNN has been utilized to track an ECH absorption vertical position target in dynamic KSTAR plasmas as well as under varying toroidal mirror angles and with a minimal average tracking error of 0.5 cm.

Original language	English (US)
Article number	055036
Journal	Plasma Physics and Controlled Fusion
Volume	67
Issue number	5
DOIs	https://doi.org/10.1088/1361-6587/add08b
State	Published - May 31 2025

All Science Journal Classification (ASJC) codes

Nuclear Energy and Engineering
Condensed Matter Physics

Keywords

electron cyclotron heating
machine learning surrogate
nuclear fusion
Tokamak

Access to Document

10.1088/1361-6587/add08b

Cite this

@article{d5046402b3914ecd9753c5e2242004de,

title = "TorbeamNN: machine learning-based steering of ECH mirrors on KSTAR",

abstract = "We have developed TorbeamNN: a machine learning surrogate model for the TORBEAM ray tracing code to predict electron cyclotron heating (ECH) and current drive locations in tokamak plasmas. TorbeamNN provides more than a 100 times speed-up compared to the highly optimized and simplified real-time implementation of TORBEAM without any reduction in accuracy compared to the offline, full fidelity TORBEAM code. The model was trained using KSTAR ECH mirror geometries and works for both O-mode and X-mode absorption. The TorbeamNN predictions have been validated both offline and real-time in experiment. TorbeamNN has been utilized to track an ECH absorption vertical position target in dynamic KSTAR plasmas as well as under varying toroidal mirror angles and with a minimal average tracking error of 0.5 cm.",

keywords = "electron cyclotron heating, machine learning surrogate, nuclear fusion, Tokamak",

author = "Andrew Rothstein and Minseok Kim and Minho Woo and Minsoo Cha and Cheolsik Byun and Sangkyeun Kim and Keith Erickson and Youngho Lee and Josh Josephy-Zack and Jalal Butt and Ricardo Shousha and Mi Joung and Juhn, {June Woo} and Lee, {Kyu Dong} and Egemen Kolemen",

note = "Publisher Copyright: {\textcopyright} 2025 The Author(s). Published by IOP Publishing Ltd.",

year = "2025",

month = may,

day = "31",

doi = "10.1088/1361-6587/add08b",

language = "English (US)",

volume = "67",

journal = "Plasma Physics and Controlled Fusion",

issn = "0741-3335",

publisher = "IOP Publishing Ltd.",

number = "5",

}

TY - JOUR

T1 - TorbeamNN

T2 - machine learning-based steering of ECH mirrors on KSTAR

AU - Rothstein, Andrew

AU - Kim, Minseok

AU - Woo, Minho

AU - Cha, Minsoo

AU - Byun, Cheolsik

AU - Kim, Sangkyeun

AU - Erickson, Keith

AU - Lee, Youngho

AU - Josephy-Zack, Josh

AU - Butt, Jalal

AU - Shousha, Ricardo

AU - Joung, Mi

AU - Juhn, June Woo

AU - Lee, Kyu Dong

AU - Kolemen, Egemen

PY - 2025/5/31

Y1 - 2025/5/31

N2 - We have developed TorbeamNN: a machine learning surrogate model for the TORBEAM ray tracing code to predict electron cyclotron heating (ECH) and current drive locations in tokamak plasmas. TorbeamNN provides more than a 100 times speed-up compared to the highly optimized and simplified real-time implementation of TORBEAM without any reduction in accuracy compared to the offline, full fidelity TORBEAM code. The model was trained using KSTAR ECH mirror geometries and works for both O-mode and X-mode absorption. The TorbeamNN predictions have been validated both offline and real-time in experiment. TorbeamNN has been utilized to track an ECH absorption vertical position target in dynamic KSTAR plasmas as well as under varying toroidal mirror angles and with a minimal average tracking error of 0.5 cm.

AB - We have developed TorbeamNN: a machine learning surrogate model for the TORBEAM ray tracing code to predict electron cyclotron heating (ECH) and current drive locations in tokamak plasmas. TorbeamNN provides more than a 100 times speed-up compared to the highly optimized and simplified real-time implementation of TORBEAM without any reduction in accuracy compared to the offline, full fidelity TORBEAM code. The model was trained using KSTAR ECH mirror geometries and works for both O-mode and X-mode absorption. The TorbeamNN predictions have been validated both offline and real-time in experiment. TorbeamNN has been utilized to track an ECH absorption vertical position target in dynamic KSTAR plasmas as well as under varying toroidal mirror angles and with a minimal average tracking error of 0.5 cm.

KW - electron cyclotron heating

KW - machine learning surrogate

KW - nuclear fusion

KW - Tokamak

UR - http://www.scopus.com/inward/record.url?scp=105005154260&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=105005154260&partnerID=8YFLogxK

U2 - 10.1088/1361-6587/add08b

DO - 10.1088/1361-6587/add08b

M3 - Article

AN - SCOPUS:105005154260

SN - 0741-3335

VL - 67

JO - Plasma Physics and Controlled Fusion

JF - Plasma Physics and Controlled Fusion

IS - 5

M1 - 055036

ER -

TorbeamNN: machine learning-based steering of ECH mirrors on KSTAR

Abstract

All Science Journal Classification (ASJC) codes

Keywords

Access to Document

Other files and links

Fingerprint

Cite this