Filament structures and average radial scrape-off layer profiles in Alcator C-Mod
ORAL
Abstract
Time-averaged radial profiles of electron temperature and density and their fluctuations are suspected to be set by plasma filaments moving through the scrape-off layer (SOL) [1; 2; 3]. The statistical properties of the plasma filaments (their sizes, velocities, amplitudes and frequency of occurrence) determine both the magnitude and the scale length of the average radial profile.
In this contribution, gas puff imaging (GPI) data from Alcator C-Mod are analysed. The two-dimensional measurements are recorded utilizing a fast-framing camera with a frame rate of 396 kHz and high pixel density (64×64). Plasma filaments are identified and tracked over time utilizing Recurrent All-Pairs Field Transforms (RAFT), a deep neural network architecture developed for studying optical flow [4; 5].
The average SOL profiles are inferred from the measured distributions of filament parameters via a stochastic model, describing SOL profiles and fluctuations as a super-position of uncorrelated structures [6]. It is investigated how the filament parameters depend on experimental control parameters such as line-averaged density, plasma current and heating power. The work will aid in developing predictive capabilities for SOL profiles of future experiments such as SPARC.
References
[1] N. R. Walkden, et al., Plasma Phys. Contr. Fusion, 59.8, 085009, (2017).
[2] F. Militello, et al., Nucl. Fusion, 56.10, 104004, (2016).
[3] O. E. Garcia, Phys. Rev., 108.26, 265001, (2012).
[4] Z. Teed, et al., Europ. conf. on computer vision, 402-419 (2020)
[5] W. Han, et al. arXiv preprint, 2111.08570 (2021).
[6] O. E. Garcia, et al., Phys. Plasmas, 23.5, 052308, (2016).
In this contribution, gas puff imaging (GPI) data from Alcator C-Mod are analysed. The two-dimensional measurements are recorded utilizing a fast-framing camera with a frame rate of 396 kHz and high pixel density (64×64). Plasma filaments are identified and tracked over time utilizing Recurrent All-Pairs Field Transforms (RAFT), a deep neural network architecture developed for studying optical flow [4; 5].
The average SOL profiles are inferred from the measured distributions of filament parameters via a stochastic model, describing SOL profiles and fluctuations as a super-position of uncorrelated structures [6]. It is investigated how the filament parameters depend on experimental control parameters such as line-averaged density, plasma current and heating power. The work will aid in developing predictive capabilities for SOL profiles of future experiments such as SPARC.
References
[1] N. R. Walkden, et al., Plasma Phys. Contr. Fusion, 59.8, 085009, (2017).
[2] F. Militello, et al., Nucl. Fusion, 56.10, 104004, (2016).
[3] O. E. Garcia, Phys. Rev., 108.26, 265001, (2012).
[4] Z. Teed, et al., Europ. conf. on computer vision, 402-419 (2020)
[5] W. Han, et al. arXiv preprint, 2111.08570 (2021).
[6] O. E. Garcia, et al., Phys. Plasmas, 23.5, 052308, (2016).
–
Presenters
-
Gregor Decristoforo
UiT The Arctic University of Norway, UiT - The Arctic University of Norway
Authors
-
Gregor Decristoforo
UiT The Arctic University of Norway, UiT - The Arctic University of Norway
-
Woonghee Han
Massachusetts Institute of Technology MI, Massachusetts Institute of Technology MIT
-
James L Terry
MIT PSFC, Massachusetts Institute of Technology MI, MIT-PSFC, MIT Plasma Science Fusion Center, Massachusetts Institute of Technology MIT
-
Odd Erik Garcia
UiT The Arctic University of Norway, UiT - The Arctic University of Norway