-
利用机器学习技术，开发了一种全新的实验诊断方法，纯粹基于单颗粒的位置涨落信息，实现了对二维尘埃等离子体屏蔽参数κ和耦合参数Γ等全局性质信息的准确诊断，并通过模拟和实验数据有效验证.为了训练、验证和测试神经网络模型，针对二维尘埃等离子体系统，本文实施了不同κ和Γ数值下数百组独立的朗之万动力学模拟，以获取大量的单颗粒动力学数据.为了验证该诊断方法的可行性，设计了三种不同的卷积神经网络模型，用于实现对该系统屏蔽参数κ的诊断.分析结果显示，这三种模型对κ诊断结果和设定值几乎一致，其均方根误差分别为0.081、0.279和0.155,表现达到预期.而对实验数据，其诊断出的κ数值分布呈单峰分布，且峰值位置与传统方法诊断出的κ数值高度一致.在此基础上，对该诊断方法进行了进一步地优化改进，使其能同时确定二维尘埃等离子体系统的屏蔽参数κ和耦合参数Γ,并通过模拟和实验数据确认其准确性.本文设计的卷积神经网络，其优异表现清楚地表明，通过机器学习，能够仅根据单颗粒动力学信息准确诊断尘埃等离子体系统的全局性质信息.Currently, it is a great challenge to accurately diagnose global properties of dusty plasmas from limited data. Based on machine learning, a novel diagnostic of various global properties in dusty plasma experiments is developed from individual particle dynamics. It is found that, for both two-dimensional (2D) dusty plasma simulations and experiments, the global properties of the screening parameters κ and the coupling parameter Γ can be accurately determined purely from the position fluctuations of individual particles. Hundreds of independent Langevin dynamical simulations are performed with various specified κ and Γ values, leading to abundant individual particle position fluctuation data, which can be used for training, validating, and testing various convolutional neural network (CNN) models. To confirm the feasibility of this diagnostic, three different CNN models are designed to determine the κ value. For the simulation data, all these CNN models have excellent performance in determining the κ value, i.e., the averaged determined κ value is nearly the same as the specified κ value. For the experiment data, the distributions of the determined κ values always exhibit one prominent peak, whose locations well agree with the determined κ value from the widely accepted phonon spectra fitting method. Furthermore, this diagnostic method is further developed to determine the κ and Γ values simultaneously, with the satisfactory outcome using the input 2D dusty plasma data from both simulations and experiments. The excellent performance of the CNN models developed here clearly indicates that, using machine learning, all information of global properties of 2D dusty plasmas can be obtained purely from individual particle dynamics.
-
Keywords:
- Machine learning /
- Convolutional neural network /
- Dusty plasma /
- Diagnositc
-
[1] Pathria R K, Beale P D 2021 Statistical Mechanics (London:Academic)
[2] Feng Y, Goree J, Liu B 2007 Rev. Sci. Instrum. 78 053704
[3] Feng Y, Goree J, Liu B 2011 Rev. Sci. Instrum. 82 053707
[4] He Y F, Ai B Q, Dai C X, Song C, Wang R Q, Sun W T, Liu F C, Feng Y 2020 Phys. Rev. Lett. 124 075001
[5] Beckers J, Berndt J, Block D, Bonitz M, Bruggeman P J, Couëdel L, Delzanno G L, Feng Y, Gopalakrishnan R, Greiner F, Hartmann P, Horányi M, Kersten H, Knapek C A, Konopka U, Kortshagen U, Kostadinova E G, Kovačević E, Krasheninnikov S I, Mann I, Mariotti D, Matthews L S, Melzer A, Mikikian M, Nosenko V, Pustylnik M Y, Ratynskaia S, Sankaran R M, Schneider V, Thimsen E J, Thomas E, Thomas H M, Tolias P, van de Kerkhof M 2023 Phys. Plasmas 30 120601
[6] Goree J 1994 Plasma Sources Sci. Technol. 3 400
[7] Feng Y, Goree J, Liu B 2008 Phys. Rev. Lett. 100 205007
[8] Feng Y, Goree J, Liu B 2010 Phys. Rev. Lett. 105 025002
[9] Lu S, Huang D, Feng Y 2021 Phys. Rev. E 103 063214
[10] Huang D, Lu S, Shi X Q, Goree J, Feng Y 2021 Phys. Rev. E 104 035207
[11] Konopka U, Morfill G, Ratke L 2000 Phys. Rev. Lett. 84 891
[12] Ichimaru S 1982 Rev. Mod. Phys. 54 1017
[13] Khrapak S, Couëdel L 2020 Phys. Rev. E 102 033207
[14] Bajaj P, Khrapak S, Yaroshenko V, Schwabe M 2022 Phys. Rev. E 105 025202
[15] Nunomura S, Goree J, Hu S, Wang X, Bhattacharjee A, Avinash K 2002 Phys. Rev. Lett. 89 035001
[16] Nunomura S, Zhdanov S, Morfill G E, Goree J 2003 Phys. Rev. E 68 026407
[17] Nosenko V, Goree J 2004 Phys. Rev. Lett. 93 155004
[18] Melzer A, Homann A, Piel A 1996 Phys. Rev. E 53 2757
[19] Zhang S, Wang S, Liu X, Wang X, Liu F, He Y F 2025 Acta Phys. Sin. 74 075202(in Chinese)[张 顺欣,王硕,刘雪,王新占,刘富成,贺亚峰2025物理学报74 075202]
[20] Tian M, Yao T Y, Cai Z M, Liu F C, He Y F 2024 Acta Phys. Sin. 73 115201(in Chinese)[田淼,姚 廷昱,才志民,刘富成,贺亚峰2024物理学报73 115201]
[21] Huang Y F, Jia W Z, Zhang Y Y, Song Y H 2024 Acta Phys. Sin. 73 085202(in Chinese)[黄渝峰, 贾文柱,张莹莹,宋远红2024物理学报73 085202]
[22] Kalman G J, Hartmann P, Donkó Z, Rosenberg M 2004 Phys. Rev. Lett. 92 065001
[23] Nosenko V, Goree J, Ma Z W, Piel A 2002 Phys. Rev. Lett. 88 135001
[24] Brunton S L, Noack B R, Koumoutsakos P 2020 Annu. Rev. Fluid Mech. 52 477
[25] Butler K T, Davies D W, Cartwright H, Isayev O, Walsh A 2018 Nature 559 547
[26] Degrave J, Felici F, Buchli J, Neunert M, Tracey B, Carpanese F, Ewalds T, Hafner R, Abdolmaleki A, de Las Casas D, et al. 2022 Nature 602 414
[27] Huang H, Nosenko V, Huang-Fu H X, Thomas H M, Du C R 2022 Phys. Plasmas 29 073702
[28] Huang H, Schwabe M, Du C R 2019 J. Imaging 5 36
[29] Wang Z, Xu J, Kovach Y E, Wolfe B T, Thomas E, Guo H, Foster J E, Shen H W 2020 Phys. Plasmas 27 033703
[30] Dormagen N, Klein M, Schmitz A S, Thoma M H, Schwarz M 2024 J. Imaging 10 40
[31] Ding Z, Yao J, Wang Y, Yuan C, Zhou Z, Kudryavtsev A A, Gao R, Jia J 2021 Plasma Sci. Technol. 23 095403
[32] Yu W, Cho J, Burton J C 2022 Phys. Rev. E 106 035303
[33] Liang C, Huang D, Lu S, Feng Y 2023 Phys. Rev. Res. 5 033086
[34] Liang C, Huang D, Lu S, Feng Y 2024 Phys. Plasmas 31 113702
[35] Liu B, Avinash K, Goree J 2003 Phys. Rev. Lett. 91 255003
[36] Feng Y, Liu B, Goree J 2008 Phys. Rev. E 78 026415
[37] LeCun Y, Bengio Y, Hinton G 2015 Nature 521 436
[38] Kingma D P, Ba J 2014 arXiv:1412.6980
计量
- 文章访问数: 25
- PDF下载量: 2
- 被引次数: 0
下载:
