Using convolutional neural networks to detect edge localized modes in DIII-D from Doppler backscattering measurements
In H-mode tokamak plasmas, the plasma is sometimes ejected beyond the edge transport barrier. These events are known as edge localized modes (ELMs). ELMs cause a loss of energy and damage the vessel walls. Understanding the physics of ELMs, and by extension, how to detect and mitigate them, is an im...
| Main Authors: | , , , , , , , , |
|---|---|
| Other Authors: | |
| Format: | Journal Article |
| Language: | English |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/181600 |
| _version_ | 1826115967893110784 |
|---|---|
| author | Teo, Nigel Qun Xuan Hall-Chen, V. H. Barada, K. Ng, R. J. H. Gu, L. Yeoh, A. K. Pratt, Q. T. Garbet, X. Rhodes, T. L. |
| author2 | School of Physical and Mathematical Sciences |
| author_facet | School of Physical and Mathematical Sciences Teo, Nigel Qun Xuan Hall-Chen, V. H. Barada, K. Ng, R. J. H. Gu, L. Yeoh, A. K. Pratt, Q. T. Garbet, X. Rhodes, T. L. |
| author_sort | Teo, Nigel Qun Xuan |
| collection | NTU |
| description | In H-mode tokamak plasmas, the plasma is sometimes ejected beyond the edge transport barrier. These events are known as edge localized modes (ELMs). ELMs cause a loss of energy and damage the vessel walls. Understanding the physics of ELMs, and by extension, how to detect and mitigate them, is an important challenge. In this paper, we focus on two diagnostic methods-deuterium-alpha (Dα) spectroscopy and Doppler backscattering (DBS). The former detects ELMs by measuring Balmer alpha emission, while the latter uses microwave radiation to probe the plasma. DBS has the advantages of having a higher temporal resolution and robustness to damage. These advantages of DBS diagnostic may be beneficial for future operational tokamaks, and thus, data processing techniques for DBS should be developed in preparation. In sight of this, we explore the training of neural networks to detect ELMs from DBS data, using Dα data as the ground truth. With shots found in the DIII-D database, the model is trained to classify each time step based on the occurrence of an ELM event. The results are promising. When tested on shots similar to those used for training, the model is capable of consistently achieving a high f1-score of 0.93. This score is a performance metric for imbalanced datasets that ranges between 0 and 1. We evaluate the performance of our neural network on a variety of ELMs in different high confinement regimes (grassy ELM, RMP mitigated, and wide-pedestal), finding broad applicability. Beyond ELMs, our work demonstrates the wider feasibility of applying neural networks to data from DBS diagnostic. |
| first_indexed | 2025-03-09T11:31:37Z |
| format | Journal Article |
| id | ntu-10356/181600 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2025-03-09T11:31:37Z |
| publishDate | 2024 |
| record_format | dspace |
| spelling | ntu-10356/1816002024-12-16T15:35:57Z Using convolutional neural networks to detect edge localized modes in DIII-D from Doppler backscattering measurements Teo, Nigel Qun Xuan Hall-Chen, V. H. Barada, K. Ng, R. J. H. Gu, L. Yeoh, A. K. Pratt, Q. T. Garbet, X. Rhodes, T. L. School of Physical and Mathematical Sciences Institute of High Performance Computing, A*STAR Physics Backscattering measurement Convolutional neural network In H-mode tokamak plasmas, the plasma is sometimes ejected beyond the edge transport barrier. These events are known as edge localized modes (ELMs). ELMs cause a loss of energy and damage the vessel walls. Understanding the physics of ELMs, and by extension, how to detect and mitigate them, is an important challenge. In this paper, we focus on two diagnostic methods-deuterium-alpha (Dα) spectroscopy and Doppler backscattering (DBS). The former detects ELMs by measuring Balmer alpha emission, while the latter uses microwave radiation to probe the plasma. DBS has the advantages of having a higher temporal resolution and robustness to damage. These advantages of DBS diagnostic may be beneficial for future operational tokamaks, and thus, data processing techniques for DBS should be developed in preparation. In sight of this, we explore the training of neural networks to detect ELMs from DBS data, using Dα data as the ground truth. With shots found in the DIII-D database, the model is trained to classify each time step based on the occurrence of an ELM event. The results are promising. When tested on shots similar to those used for training, the model is capable of consistently achieving a high f1-score of 0.93. This score is a performance metric for imbalanced datasets that ranges between 0 and 1. We evaluate the performance of our neural network on a variety of ELMs in different high confinement regimes (grassy ELM, RMP mitigated, and wide-pedestal), finding broad applicability. Beyond ELMs, our work demonstrates the wider feasibility of applying neural networks to data from DBS diagnostic. Agency for Science, Technology and Research (A*STAR) Nanyang Technological University National Research Foundation (NRF) Published version This work was funded by the Urban and Green Tech Office, A ∗ STAR, Green Seed Fund (Grant No. C231718014), the National Research Foundation, Singapore; the Nanyang Technological University Overseas Travel (Grant No. 03INS001464C230OST01); and JST Moonshot R&D (Grant No. JPMJMS2011). This material is also based upon work supported by the U.S. Department of Energy, Office of Science, Office of Fusion Energy Sciences, using the DIIID National Fusion Facility, a DOE Office of Science user facility, under Grant Nos. DEFC02-04ER54698, DE-SC0019005, and DESC0019007. N.Q.X. Teo thanks the Nanyang Technological University (NTU) Singapore, CN Yang Scholars Program, for financial support. 2024-12-10T06:13:21Z 2024-12-10T06:13:21Z 2024 Journal Article Teo, N. Q. X., Hall-Chen, V. H., Barada, K., Ng, R. J. H., Gu, L., Yeoh, A. K., Pratt, Q. T., Garbet, X. & Rhodes, T. L. (2024). Using convolutional neural networks to detect edge localized modes in DIII-D from Doppler backscattering measurements. Review of Scientific Instruments, 95(7), 073528-. https://dx.doi.org/10.1063/5.0215748 0034-6748 https://hdl.handle.net/10356/181600 10.1063/5.0215748 39037296 2-s2.0-85199360855 7 95 073528 en C231718014 03INS001464C230OST01 JPMJMS2011 Review of Scientific Instruments © 2024 Author(s). Published under an exclusive license by AIP Publishing. All rights reserved. This article may be downloaded for personal use only. Any other use requires prior permission of the copyright holder. The Version of Record is available online at http://doi.org/10.1063/5.0215748 application/pdf |
| spellingShingle | Physics Backscattering measurement Convolutional neural network Teo, Nigel Qun Xuan Hall-Chen, V. H. Barada, K. Ng, R. J. H. Gu, L. Yeoh, A. K. Pratt, Q. T. Garbet, X. Rhodes, T. L. Using convolutional neural networks to detect edge localized modes in DIII-D from Doppler backscattering measurements |
| title | Using convolutional neural networks to detect edge localized modes in DIII-D from Doppler backscattering measurements |
| title_full | Using convolutional neural networks to detect edge localized modes in DIII-D from Doppler backscattering measurements |
| title_fullStr | Using convolutional neural networks to detect edge localized modes in DIII-D from Doppler backscattering measurements |
| title_full_unstemmed | Using convolutional neural networks to detect edge localized modes in DIII-D from Doppler backscattering measurements |
| title_short | Using convolutional neural networks to detect edge localized modes in DIII-D from Doppler backscattering measurements |
| title_sort | using convolutional neural networks to detect edge localized modes in diii d from doppler backscattering measurements |
| topic | Physics Backscattering measurement Convolutional neural network |
| url | https://hdl.handle.net/10356/181600 |
| work_keys_str_mv | AT teonigelqunxuan usingconvolutionalneuralnetworkstodetectedgelocalizedmodesindiiidfromdopplerbackscatteringmeasurements AT hallchenvh usingconvolutionalneuralnetworkstodetectedgelocalizedmodesindiiidfromdopplerbackscatteringmeasurements AT baradak usingconvolutionalneuralnetworkstodetectedgelocalizedmodesindiiidfromdopplerbackscatteringmeasurements AT ngrjh usingconvolutionalneuralnetworkstodetectedgelocalizedmodesindiiidfromdopplerbackscatteringmeasurements AT gul usingconvolutionalneuralnetworkstodetectedgelocalizedmodesindiiidfromdopplerbackscatteringmeasurements AT yeohak usingconvolutionalneuralnetworkstodetectedgelocalizedmodesindiiidfromdopplerbackscatteringmeasurements AT prattqt usingconvolutionalneuralnetworkstodetectedgelocalizedmodesindiiidfromdopplerbackscatteringmeasurements AT garbetx usingconvolutionalneuralnetworkstodetectedgelocalizedmodesindiiidfromdopplerbackscatteringmeasurements AT rhodestl usingconvolutionalneuralnetworkstodetectedgelocalizedmodesindiiidfromdopplerbackscatteringmeasurements |