-
对于旨在实现高参数和长脉冲运行的磁约束聚变装置而言，基于离子温度实时测量的等离子体反馈控制至关重要，而电荷交换复合光谱是等离子体离子温度的基本测量手段。本文提出了一种基于神经网络的电荷交换复合光谱诊断数据快速分析方法，并对其跨参数区间的外推能力进行研究。该研究使用了中国环流器二号A装置HL-2A的12.2万个光谱数据及离线解谱获得的离子温度标签值构成数据集。模型基于卷积神经网络，相对于标签值实现了拟合优度R2~0.92的效果，在推理阶段单光谱耗时小于1 ms，相比传统方法加速了100至1000倍。在外推能力方面，提出基于低温度实验数据生成高温度的合成光谱数据的方法，并通过在只包含离子温度2 keV以下的训练集中添加大约5%的合成数据，大幅增加了模型在外推参数区间（2至4 keV）分析的准确性，并将模型在3-4 keV区间测试的误差降低了约60%。该研究证明了在磁约束核聚变领域利用合成数据提升人工智能算法性能的可行性。Real-time measurement and feedback control of key plasma parameters are critical for future fusion reactor operation, with ion temperature being a vital control target as part of the triple product for fusion ignition. However, plasma diagnostics tends to require complex data analysis. To acquire ion temperature Ti from charge exchange recombination spectroscopy (CXRS), a widely used method is through iterative spectral fitting, which is time-consuming and calls for expert intervention during data analysis. On top of that, frequent human expert intervention is needed in the conventional iterative fitting. Therefore, the conventional method cannot meet the meet the demand for real-time Ti measurement. Neural Networks (NN), which is capable of learning the underlying relationships between the measured spectra and Ti, is a promising approach to cope with this problem. In fact, NN approaches have been widely adopted in the field of magnetic confined plasma. Previous study in JET has achieved a satisfactory accuracy for inferring Ti from CXRS spectra compared to the conventional fitting results. Recently the study of disruption prediction has achieved great progress with the help of deep neural networks. However, these researches are conducted in steadily-operating devices, where for NN models, the data distribution is similar in training set and test set. This is not the case for newly-built tokamak like HL-3, or for future fusion reactors such as ITER. For new devices, there will be a period for the plasma parameters to raise from low to high ranges. In this case, investigating the extrapolation capability of NN models based on low parameter training data is of paramount importance.
A Convolutional Neural Network (CNN)-based model is proposed to accelerate the analysis of spectral data of CXRS, with a focus on investigating the model’s extrapolation capability to much higher Ti ranges. The dataset consists of about 122 thousand pieces of spectral data, along with their corresponding inferred Ti from offline iterative process. The results demonstrate that the CNN-based model provides excellent Ti analysis and reduces the inference time for analyzing a single spectrum to less than 1 ms, which is 100-1000 times faster compared to traditional spectral fitting method. However, the performance of the data-driven neural network model is limited by challenges such as insufficient data and imbalanced data distribution, which further deteriorates the extrapolation capability. Generally, data with higher Ti constitute a small percentage of the total dataset. In the case of our study, only about 5% of the spectra correspond to Ti > 2 keV (among 2-4 keV). Yet they reflect the temperature of central plasma, which is more important for assessing the performance of plasma. To overcome this limitation, the study synthesizes high-temperature data based on experimental data from discharges with Ti in low-temperature range. By incorporating 5% synthetic data into the training set only consisting of data with Ti<2 keV, the model’s extrapolation capability is extended to cover the whole range of Ti < 4 keV. The mean relative error of the mode in 3 keV < Ti < 4 keV is reduced from 35% to below 15%. This approach demonstrates the feasibility of using synthetic data to enhance the performance of artificial intelligence algorithms in the field of magnetic confinement fusion. The findings provide valuable insights for the development of real-time ion temperature measurement and feedback control for future high-parameter fusion devices. Furthermore, the study lays a foundation for research in areas that require high-performance across-device characteristic, such as machine learning-based disruption prediction and tearing mode control.-
Keywords:
- plasma /
- neural network /
- extrapolation capability /
- spectral diagnostic
-
[1] Lister J B, Bruzzone P L, Costley A E, Fukuda T, Yu G, Mertens V, Moreau D, Oikawa T, Pitts R A, Portone A, Vayakis G, Wesley J,Yoshino R 2000 Nucl. Fusion 40 1167
[2] Gormezano C, Sips A C C, Luce T C, Ide S, Becoulet A, Litaudon X, Isayama A, Hobirk J, Wade M R, Oikawa T, Prater R, Zvonkov A, Lloyd B, Suzuki T, Barbato E, Bonoli P, Phillips C K, Vdovin V, Joffrin E, Casper T, Ferron J, Mazon D, Moreau D, Bundy R, Kessel C, Fukuyama A, Hayashi N, Imbeaux F, Murakami M, Polevoi A R,St John H E 2007 Nucl. Fusion 47 S285
[3] Hellermann M G V, Mandl W, Summers H P, Weisen H,Wolf R 1990 Rev. Sci. Instrum. 61 3479
[4] Seraydarian R P,Burrell K H 1986 Rev. Sci. Instrum. 57 2012
[5] He X X, Chen W J, Wei Y L, Liu L, Wang S Q, Yu D L 2024 Nuclear Fusion and Plasma Physics 44 477（何小雪, 陈文锦, 魏彦玲, 刘亮, 王诗琴,余德良 2024 核聚变与等离子体物理 44 477）
[6] Zhang Y 2015 M.S. Thesis (Hefei: University of Science and Technology of China) （张镱 2015 硕士学位论文（合肥：中国科学技术大学））
[7] Svensson J, Hellermann M v,König R W T 1999 Plasma Phys. Control. Fusion 41 315
[8] Bishop C M,Roach C M 1992 Rev. Sci. Instrum. 63 4450
[9] Abbate J, Conlin R,Kolemen E 2021 Nucl. Fusion 61 046027
[10] Zheng G H, Yang Z Y, Liu S F, Ma R, Gong X W, Wang A, Wang S,Zhong W L 2024 Nucl. Fusion 64 126041
[11] Yang Z, Xia F, Song X, Gao Z, Wang S,Dong Y 2021 Nucl. Fusion 61 126042
[12] Yang Z, Xia F, Song X, Gao Z, Li Y, Gong X, Dong Y, Zhang Y, Chen C, Luo C, Li B, Zhu X, Ji X, Li Y, Liu L, Gao J,Liu Y 2022 Fusion Eng. Des. 182 113223
[13] Seo J, Kim S K, Jalalvand A, Conlin R, Rothstein A, Abbate J, Erickson K, Wai J, Shousha R,Kolemen E 2024 Nature 626
[14] Duan X R, Xu M, Zhong W L, Ji X Q, Chen W, Shi Z B, Liu X L, Lu B, Li B, Wang Y Q, Li J Q, Zheng G Y, Liu Y, Yang Q W, Yan L W, Cai L J, Li Q, Liu Y, Bai X Y, Cao Z, Chen X, Chen H T, Chen Y H, Dong G Q, Du H L, Fan D M, Gao J M, Geng S F, Hao G Z, He H M, Huang M, Jiang M, Ke R, Liang A S, Li J X, Li Q, Li Y, Li L C, Li H J, Li W B, Liu D Q, Long T, Lu L F, Nie L, Shi P W, Peng J F, Sun A P, Sun T F, Tong R H, Wei H L, Wang S, Xiao G L, Xiao X P, Xue L, Xu H B, Yang Z Y, Yu D L, Yu L M, Zhang Y P, Zheng X, Zhang L, Zhang Y, Zhang F, Zhang X L, Team H L,Collaborators 2024 Nucl. Fusion 64 112021
[15] Duan X R, Xu M, Zhong W L, Liu Y, Song X M, Liu D Q, Wang Y Q, Lu B, Shi Z B, Zheng G Y, Liu Y, Yang Q W, Mao W C, Li Q, Cai L J, Ji X Q, Liu X L, Li L C, Li B, Dong J Q, Ding X T, Yan L W, Artaud J F, Bai X Y, Cao J Y, Cao Z, Chen L, Chen W, Delpech L, Du H L, Ekedahl A, Gao Z, Garcia J, Han M K, Hao G Z, He H M, Hoang G T, Huang M, Isobe M, Jiang M, Liang A S, Liu Y Q, Li D, Li H J, Li J Q, Li J X, Li Q, Li Y, Long T, Mazon D, McKee G R, Qiu Z Y, Peng J F, Peysson Y, Rao J, Song X, Sun T F, Wang Z X, Wei H L, Wen J, Wu N, Xu Y H, Xiao G L, Xiao X P, Xue L, Yan Z, Yang Z Y, Yu D L, Yu L M, Yu Y, Zang L G, Zhang J H, Zhang N, Zhang Y P, Zonca F, Zou X L,Team H-A H-M 2022 Nucl. Fusion 62 042020
[16] Donné A J H, Costley A E, Barnsley R, Bindslev H, Boivin R, Conway G, Fisher R, Giannella R, Hartfuss H, von Hellermann M G, Hodgson E, Ingesson L C, Itami K, Johnson D, Kawano Y, Kondoh T, Krasilnikov A, Kusama Y, Litnovsky A, Lotte P, Nielsen P, Nishitani T, Orsitto F, Peterson B J, Razdobarin G, Sanchez J, Sasao M, Sugie T, Vayakis G, Voitsenya V, Vukolov K, Walker C, Young K,the I T G o D 2007 Nucl. Fusion 47 S337
[17] Wei Y L, Yu D L, Liu L, Ida K, von Hellermann M, Cao J Y, Sun A P, Ma Q, Chen W J, Liu Y, Yan L W, Yang Q W, Duan X R,Liu Y 2014 Rev. Sci. Instrum. 85 103503
[18] He X X, Yu D L, Yan L W, Liu L, Chen W J, Wei Y L, He X F, Ma Q, Shi Z B, Liu Y, Yang Q W, Xu M,Duan X R 2020 Rev. Sci. Instrum. 91 053504
[19] Wang Y F, Zhu X M, Zhang M Z, Meng S F, Chai H, Wang Y, Ning Z X 2021 Acta Phys. Sin., 70, 095211（王彦飞, 朱悉铭, 张明志, 孟圣峰, 贾军伟, 柴昊, 王旸,宁中喜 2021 70 095211-1）
[20] Fei N, Gao Y, Lu Z,Xiang T 2021 IEEE International Conference on Computer Vision 142-151
[21] IEEE Conference on Computer Vision and Pattern Recognition He K, Zhang X, Ren S,Sun J 2015 10.48550/arXiv.1512.03385[2015-12-10]
[22] He M, Yang Z, Liu S, Xia F,Zhong W 2024 Plasma Phys. Control. Fusion 66 105019
[23] Hendrycks D,Gimpel K 2016 10.48550/arXiv.1606.08415[2016-7-8]
[24] Loshchilov I,Hutter F 2017 10.48550/arXiv.1711.05101[2017-11-14]
[25] Yushmanov P N, Takizuka T, Riedel K S, Kardaun O J W F, Cordey J G, Kaye S M,Post D E 1990 Nucl. Fusion 30 1999
[26] Sundararajan M, Taly A,Yan Q 2017 10.48550/arXiv.1703.01365[2017-3-4]
计量
- 文章访问数: 80
- PDF下载量: 2
- 被引次数: 0
下载:
