Special Issue:
SPECIAL TOPIC — Plasma disruption
|
SPECIAL TOPIC—Plasma disruption |
Prev
Next
|
|
|
Disruption prediction based on fusion feature extractor on J-TEXT |
Wei Zheng(郑玮)1, Fengming Xue(薛凤鸣)1,2, Zhongyong Chen(陈忠勇)1,†, Chengshuo Shen(沈呈硕)1, Xinkun Ai(艾鑫坤)1, Yu Zhong(钟昱)1, Nengchao Wang(王能超)1, Ming Zhang(张明)1, Yonghua Ding(丁永华)1, Zhipeng Chen(陈志鹏)1, Zhoujun Yang(杨州军)1, and Yuan Pan(潘垣)1 |
1 International Joint Research Laboratory of Magnetic Confinement Fusion and Plasma Physics, State Key Laboratory of Advanced Electromagnetic Engineering and Technology, School of Electrical and Electronic Engineering, Huazhong University of Science and Technology, Wuhan 430074, China; 2 Institute of Artificial Intelligence, Huazhong University of Science and Technology, Wuhan 430074, China |
|
|
Abstract Predicting disruptions across different tokamaks is necessary for next generation device. Future large-scale tokamaks can hardly tolerate disruptions at high performance discharge, which makes it difficult for current data-driven methods to obtain an acceptable result. A machine learning method capable of transferring a disruption prediction model trained on one tokamak to another is required to solve the problem. The key is a feature extractor which is able to extract common disruption precursor traces in tokamak diagnostic data, and can be easily transferred to other tokamaks. Based on the concerns above, this paper presents a deep feature extractor, namely, the fusion feature extractor (FFE), which is designed specifically for extracting disruption precursor features from common diagnostics on tokamaks. Furthermore, an FFE-based disruption predictor on J-TEXT is demonstrated. The feature extractor is aimed to extracting disruption-related precursors and is designed according to the precursors of disruption and their representations in common tokamak diagnostics. Strong inductive bias on tokamak diagnostics data is introduced. The paper presents the evolution of the neural network feature extractor and its comparison against general deep neural networks, as well as a physics-based feature extraction with a traditional machine learning method. Results demonstrate that the FFE may reach a similar effect with physics-guided manual feature extraction, and obtain a better result compared with other deep learning methods.
|
Received: 29 December 2022
Revised: 24 March 2023
Accepted manuscript online: 28 March 2023
|
PACS:
|
52.55.Fa
|
(Tokamaks, spherical tokamaks)
|
|
07.05.Mh
|
(Neural networks, fuzzy logic, artificial intelligence)
|
|
Fund: Project supported by the National Key R&D Program of China (Grant No. 2022YFE03040004) and the National Natural Science Foundation of China (Grant No. 51821005). |
Corresponding Authors:
Zhongyong Chen
E-mail: zychen@hust.edu.cn
|
Cite this article:
Wei Zheng(郑玮), Fengming Xue(薛凤鸣), Zhongyong Chen(陈忠勇), Chengshuo Shen(沈呈硕), Xinkun Ai(艾鑫坤), Yu Zhong(钟昱), Nengchao Wang(王能超), Ming Zhang(张明),Yonghua Ding(丁永华), Zhipeng Chen(陈志鹏), Zhoujun Yang(杨州军), and Yuan Pan(潘垣) Disruption prediction based on fusion feature extractor on J-TEXT 2023 Chin. Phys. B 32 075203
|
[1] ITER Physics Expert Group on Disruptions, Plasma Control and MHD and ITER Physics Basis Editors 1999 Nucl. Fusion 39 2251 [2] Hender T C, Wesley J C, Bialek J, et al. 2007 Nucl. Fusion 47 S128 [3] Boozer A H 2012 Phys. Plasmas 19 058101 [4] Sugihara M, Shimada M, Fujieda H, Gribov Yu, Ioki K, Kawano Y, Khayrutdinov R, Lukash V and Ohmori J 2007 Nucl. Fusion 47 337 [5] Putvinski S V, et al. 2010 Disruption mitigation in ITER Proc. 23rd IAEA Fusion Energy Conf. (Daejeon, S. Korea) (IAEA), pp. ITR/1-6 [6] Luo Y H, Chen Z Y, Tang Y, Wang S Y, Ba W G, Wei Y N, Ma T K, Huang D W, Tong R H, Yan W, Geng P, Shao J and Zhuang G 2014 Review Sci. Instruments 85 083504 [7] Li Y, Chen Z Y, Wei Y N, Tong R H, Yan W, Lin Z F, Yang Z J and Jiang Z H 2018 Rev. Sci. Instrum. 89 10K116 [8] Aymerich E, Fanni A, Sias G, Carcangiu S, Cannas B, Murari A, Pau A and the JET contributors 2021 Nucl. Fusion 61 036013 [9] Lungaroni M, Murari A, Peluso E, Vega J, Farias G, Gelfusa M and JET contributors 2018 Fusion Eng. Design 130 62 [10] Rattá G A, Vega J, Murari A, Vagliasindi G, Johnson M F, de Vries P C and JET EFDA contributors 2010 Nucl. Fusion 50 025005 [11] Rea C, Montes K J, Erickson K G, Granetz R S and Tinguely R A 2019 Nucl. Fusion 59 096016 [12] Yang Z Y, Xia F, Song X M, Gao Z, Huang Y and Wang S 2020 Nucl. Fusion 60 016017 [13] Guo B H, Shen B, Chen D L, Rea C, Granetz R S, Huang Y, Zeng L, Zhang H, Qian J P and Sun Y W 2021 Plasma Phys. Control. Fusion 63 025008 [14] Guo B H, Chen D L, Shen B, Rea C, Granetz R S, Zeng L, Hu W H, Qian J P, Sun Y W and Xiao B J 2021 Plasma Phys. Control. Fusion 63 115007 [15] Kates-Harbeck J, Svyatkovskiy A and Tang W 2019 Nature 568 526 [16] Voulodimos A, Doulamis N, Doulamis A and Protopapadakis E 2018 Comput. Intel Neuroscience 2018 7068349 [17] Young T, Hazarika D, Poria S and Cambria E 2018 IEEE Comput. Intell. Magazine 13 55 [18] Zhu J X, Rea C, Montes K, Granetz R S, Sweeney R and Tinguely R A 2021 Nucl. Fusion 61 026007 [19] Ding Y H, Chen Z Y and Chen Z P, et al. 2018 Plasma Sci. Technology 20 125101 [20] Liang Y F, Wang N C and Ding Y H, et al. 2019 Nucl. Fusion 59 112016 [21] Zhang M, Wu Q Q, Zheng W, Shang Y X and Wang Y X 2020 Fusion Eng. Design 160 111981 [22] Hochreiter S and Schmidhuber J 1997 Neural Computation 9 1735 [23] Zheng W, Hu F R, Zhang M, Chen Z Y, Zhao X Q, Wang X L, Shi P, Zhang X L, Zhang X Q, Zhou Y N, Wei Y N, Pan Y and J-TEXT team 2018 Nucl. Fusion 58 056016 [24] Nave M F F, Ali-Arshad S and Alper B, et al. 1995 Nucl. Fusion 35 409 [25] Cortes C and Vapnik V 1995 Mach. Learning 20 273 [26] Li D, Yu Q Q and Ding Y D, et al. 2020 Nucl. Fusion 60 056022 [27] Shen C S, Cai Z M, Ren T, Zhang X T, Hu Q M, Wang N C, Huang Z, Zhou S, Li J C, Li M, Li D, Han D L and Ding Y H 2019 Rev. Sci. Instruments 90 123506 [28] Caruana R 1997 Mach. Learning 28 41 |
No Suggested Reading articles found! |
|
|
Viewed |
|
|
|
Full text
|
|
|
|
|
Abstract
|
|
|
|
|
Cited |
|
|
|
|
Altmetric
|
blogs
Facebook pages
Wikipedia page
Google+ users
|
Online attention
Altmetric calculates a score based on the online attention an article receives. Each coloured thread in the circle represents a different type of online attention. The number in the centre is the Altmetric score. Social media and mainstream news media are the main sources that calculate the score. Reference managers such as Mendeley are also tracked but do not contribute to the score. Older articles often score higher because they have had more time to get noticed. To account for this, Altmetric has included the context data for other articles of a similar age.
View more on Altmetrics
|
|
|