Shadow masks predictions in SPARC tokamak plasma-facing components using HEAT code and machine learning methods

D. Corona; M. Scotto d'Abusco; M. Churchill; S. Munaretto; A. Kleiner; A. Wingen; T. Looby

doi:10.1016/j.fusengdes.2025.115010

Shadow masks predictions in SPARC tokamak plasma-facing components using HEAT code and machine learning methods

D. Corona
, M. Scotto d'Abusco
, M. Churchill
, S. Munaretto
, A. Kleiner
, A. Wingen
, T. Looby

Advanced Tokamak Physics

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

2 Scopus citations

Abstract

This work uses machine learning (ML) to complement HEAT (Heat flux Engineering Analysis Toolkit) by developing 3-D footprint surrogate models for fast and accurate heat load calculations in the divertor of the SPARC tokamak. The focus is on shadowed regions, or magnetic shadows, caused by the 3-D geometry of plasma-facing components (PFCs). ML classifiers are employed to create a surrogate model for HEAT generated shadow masks, predicting these shadow masks and divertor heat flux profiles based on a diverse range of equilibria and only the plasma current, safety factor(q95) at the edge, and magnetic flux angles as input parameters. The ultimate goal is to integrate the model for real-time control and future operational decisions.

Original language	English
Article number	115010
Journal	Fusion Engineering and Design
Volume	217
DOIs	https://doi.org/10.1016/j.fusengdes.2025.115010
State	Published - Aug 2025

Funding

This work was supported by the U.S. Department of Energy under contract number DE-AC02-09CH11466 , DE-AC05-00OR22725 and the support from Commonwealth Fusion Systems . The United States Government retains a non-exclusive, paid-up, irrevocable, world-wide license to publish or reproduce the published form of this manuscript, or allow others to do so, for United States Government purposes.

Keywords

CAD
Equilibrium
Heat flux
Neuronal Network
Plasma facing component
Shadow mask
Surrogate model

Access to Document

10.1016/j.fusengdes.2025.115010

Cite this

@article{a4d2cd9d8b164994af12fd1445376845,

title = "Shadow masks predictions in SPARC tokamak plasma-facing components using HEAT code and machine learning methods",

abstract = "This work uses machine learning (ML) to complement HEAT (Heat flux Engineering Analysis Toolkit) by developing 3-D footprint surrogate models for fast and accurate heat load calculations in the divertor of the SPARC tokamak. The focus is on shadowed regions, or magnetic shadows, caused by the 3-D geometry of plasma-facing components (PFCs). ML classifiers are employed to create a surrogate model for HEAT generated shadow masks, predicting these shadow masks and divertor heat flux profiles based on a diverse range of equilibria and only the plasma current, safety factor(q95) at the edge, and magnetic flux angles as input parameters. The ultimate goal is to integrate the model for real-time control and future operational decisions.",

keywords = "CAD, Equilibrium, Heat flux, Neuronal Network, Plasma facing component, Shadow mask, Surrogate model",

author = "D. Corona and d'Abusco, \{M. Scotto\} and M. Churchill and S. Munaretto and A. Kleiner and A. Wingen and T. Looby",

note = "Publisher Copyright: {\textcopyright} 2025",

year = "2025",

month = aug,

doi = "10.1016/j.fusengdes.2025.115010",

language = "English",

volume = "217",

journal = "Fusion Engineering and Design",

issn = "0920-3796",

publisher = "Elsevier",

}

TY - JOUR

T1 - Shadow masks predictions in SPARC tokamak plasma-facing components using HEAT code and machine learning methods

AU - Corona, D.

AU - d'Abusco, M. Scotto

AU - Churchill, M.

AU - Munaretto, S.

AU - Kleiner, A.

AU - Wingen, A.

AU - Looby, T.

PY - 2025/8

Y1 - 2025/8

N2 - This work uses machine learning (ML) to complement HEAT (Heat flux Engineering Analysis Toolkit) by developing 3-D footprint surrogate models for fast and accurate heat load calculations in the divertor of the SPARC tokamak. The focus is on shadowed regions, or magnetic shadows, caused by the 3-D geometry of plasma-facing components (PFCs). ML classifiers are employed to create a surrogate model for HEAT generated shadow masks, predicting these shadow masks and divertor heat flux profiles based on a diverse range of equilibria and only the plasma current, safety factor(q95) at the edge, and magnetic flux angles as input parameters. The ultimate goal is to integrate the model for real-time control and future operational decisions.

AB - This work uses machine learning (ML) to complement HEAT (Heat flux Engineering Analysis Toolkit) by developing 3-D footprint surrogate models for fast and accurate heat load calculations in the divertor of the SPARC tokamak. The focus is on shadowed regions, or magnetic shadows, caused by the 3-D geometry of plasma-facing components (PFCs). ML classifiers are employed to create a surrogate model for HEAT generated shadow masks, predicting these shadow masks and divertor heat flux profiles based on a diverse range of equilibria and only the plasma current, safety factor(q95) at the edge, and magnetic flux angles as input parameters. The ultimate goal is to integrate the model for real-time control and future operational decisions.

KW - CAD

KW - Equilibrium

KW - Heat flux

KW - Neuronal Network

KW - Plasma facing component

KW - Shadow mask

KW - Surrogate model

UR - https://www.scopus.com/pages/publications/105003933713

U2 - 10.1016/j.fusengdes.2025.115010

DO - 10.1016/j.fusengdes.2025.115010

M3 - Article

AN - SCOPUS:105003933713

SN - 0920-3796

VL - 217

JO - Fusion Engineering and Design

JF - Fusion Engineering and Design

M1 - 115010

ER -

Shadow masks predictions in SPARC tokamak plasma-facing components using HEAT code and machine learning methods

Abstract

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this