结合机器学习的大气压介质阻挡放电数值模拟研究

艾飞; 刘志兵; 张远涛

doi:10.7498/aps.71.20221555

摘要

大气压介质阻挡放电是应用中常用的放电形式, 通常使用等离子体流体模型进行理论描述. 本文针对大气压均匀介质阻挡放电每半个电压周期出现一次或多次电流脉冲的特性, 基于机器学习方法构造一个全连接多层神经网络, 采用误差反向传播算法, 并设计了一个通用的隐藏层结构, 将计算数据或实验数据作为训练集, 借助于人工神经网络程序研究大气压介质阻挡放电的电流密度、电子密度、离子密度和电场强度等宏观与微观放电特性. 通过分析计算结果可知, 在给定合适训练集的条件下, 构造的机器学习程序与流体模型能以近乎相同的计算精度(误差小于2%)来描述大气压介质阻挡放电等离子体性质, 同时计算效率远高于求解流体模型, 并能极大地拓展放电参数的遍历范围. 本文的算例表明, 将机器学习程序与现有的流体模型或动理学模型结合起来, 将极大地提高大气压放电等离子体的模拟效率与效果, 深化对放电等离子体的认识 .

关键词:

Abstract

In recent years, with the development of gas discharge technology at atmospheric pressure, the application of low temperature plasma has received widespread attention in pollution prevention, disinfection, sterilization, energy conversion and other fields. Atmospheric dielectric barrier discharge is widely used to produce low temperature plasma in various applications, which is usually numerically investigated by using fluid models. The unique advantages of machine learning in various branches of physics have been discovered with the advancement of big data processing technology. Recent studies have shown that artificial neural networks with multiple hidden layers have a pivotal role in the simulation of complex datasets. In this work, a fully connected multilayer BP (back propagation) network together with a universal hidden layer structure is developed to explore the characteristics of one or more current pulses per half voltage cycle of atmospheric dielectric barrier discharge. The calculated data are used as training sets, and the discharge characteristics such as current density, electron density, ion density, and electric field of atmospheric dielectric barrier discharge can be quickly predicted by using artificial neural network program. The computational results show that for a given training set, the constructed machine learning program can describe the properties of atmospheric dielectric barrier discharge with almost the same accuracy as the fluid model. Also, the computational efficiency of the machine learning is much higher than that of the fluid model. In addition, the use of machine learning programs can also greatly extend the calculation range of parameters. Limiting discharge parameter range is considered as a major challenge for numerical calculation. By substituting a relatively limited set of training data obtained from the fluid model into the machine learning, the discharge characteristics can be accurately predicted within a given range of discharge parameters, leading an almost infinite set of data to be generated, which is of great significance for studying the influence of discharge parameters on discharge evolution. The examples in this paper show that the combination of machine learning and fluid models can greatly improve the computational efficiency, which can enhance the understanding of discharge plasmas.

Keywords:

作者及机构信息

山东大学电气工程学院, 济南　250014

通信作者: 张远涛, ytzhang@sdu.edu.cn

基金项目: 国家自然科学基金(批准号: 11975142)资助的课题

Authors and contacts

School of Electrical Engineering, Shandong University, Jinan 250014, China

Corresponding author: Zhang Yuan-Tao, ytzhang@sdu.edu.cn

Funds: Project supported by the National Natural Science Foundation of China (Grant No. 11975142)

文章全文 : translate this paragraph

参考文献

[1]	Von Woedtke T, Metelmann H R, Weltmann K D 2014 Contrib. Plasma Phys. 54 104 Google Scholar
[2]	Agarwal P, Girshick S L 2014 Plasma Chem. Plasma Process. 34 489 Google Scholar
[3]	Chen Q, Li J, Li Y 2015 J. Phys. D Appl. Phys. 48 424005 Google Scholar
[4]	Von Keudell A, Schulz-Von Der Gathen V 2017 Plasma Sources Sci. Technol. 26 113001 Google Scholar
[5]	Bruggeman P J, Iza F, Brandenburg R 2017 Plasma Sources Sci. Technol. 26 123002 Google Scholar
[6]	Zhang Y T, Chi Y Y, He J 2014 Plasma Process. Polym. 11 639 Google Scholar
[7]	Wang X C, Bai J X, Zhang T H, Sun Y, Zhang Y T 2022 Vacuum 203 111200 Google Scholar
[8]	张泰恒, 王绪成, 张远涛 2021 物理学报 70 215201 Google Scholar Zhang T H, Wang X C, Zhang Y T 2021 Acta Phys. Sin. 70 215201 Google Scholar
[9]	Iqbal M M, Turner M M 2015 Plasma Process. Polym. 12 1104 Google Scholar
[10]	Zhang Y T, Wang D Z, Kong M G 2005 J. Appl. Phys. 98 113308 Google Scholar
[11]	Alves L, Bogaerts A, Guerra V, Turner M 2018 Plasma Sources Sci. Technol. 27 023002 Google Scholar
[12]	Wang G, Kuang Y, Zhang Y T 2019 Plasma Sci. Technol. 22 015404 Google Scholar
[13]	Brodtkorb A R, Hagen T R, Sætra M L 2013 J. Parallel Distrib. Comput. 73 4 Google Scholar
[14]	Pandey M, Fernandez M, Gentile F, Isayev O, Tropsha A, Stern A C, Cherkasov A 2022 Nat. Mach. Intell. 4 211 Google Scholar
[15]	Eklund A, Dufort P, Forsberg D, LaConte S M 2013 Med. Image Anal. 17 1073 Google Scholar
[16]	Carleo G, Cirac I, Cranmer K, Daudet L, Schuld M, Tishby N, Vogt-Maranto L, Zdeborová L 2019 Rev. Mod. Phys. 91 045002 Google Scholar
[17]	Piccione A, Berkery J, Sabbagh S, Andreopoulos Y 2020 Nucl. Fusion 60 046033 Google Scholar
[18]	Fu Y, Eldon D, Erickson K, Kleijwegt K, Lupin-Jimenez L, Boyer M D, Eidietis N, Barbour N, Izacard O, Kolemen E 2020 Phys. Plasma 27 022501 Google Scholar
[19]	Mesbah A, Graves D B 2019 J. Phys. D Appl. Phys. 52 30LT02 Google Scholar
[20]	Jordan M I, Mitchell T M 2015 Science 349 255 Google Scholar
[21]	Wang H, Lei Z, Zhang X, Zhou B, Peng J 2019 Energy Convers. Manage. 198 111799 Google Scholar
[22]	Bonzanini A D, Shao K, Stancampiano A, Graves D B, Mesbah A 2021 IEEE Trans. Radiat. Plasma Med. Sci. 6 16 Google Scholar
[23]	Hong Y, Hou B, Jiang H, Zhang J 2020 Wiley Interdiscip. Rev. Comput. Mol. Sci. 10 e1450 Google Scholar
[24]	Abiodun O I, Jantan A, Omolara A E, Dada K V, Mohamed N A, Arshad H 2018 Heliyon 4 e00938 Google Scholar
[25]	Dongare A, Kharde R, Kachare A D 2012 IJEIT 2 189
[26]	Kukreja H, Bharath N, Siddesh C, Kuldeep S 2016 Int. J. Adv. Res. Innov. Ideas Educ. 1 27
[27]	Han J, Jentzen A, Weinan E 2018 Proc. Natl. Acad. Sci. 115 8505 Google Scholar
[28]	Zhong L, Gu Q, Wu B 2020 Comput. Phys. Commun. 257 107496 Google Scholar
[29]	张远涛, 王德真, 王艳辉 2005 物理学报 54 4808 Google Scholar Zhang Y T, Wang D Z, Wang Y H 2005 Acta Phys. Sin. 54 4808 Google Scholar
[30]	王艳辉, 王德真 2003 物理学报 52 1694 Google Scholar Wang Y H, Wang D Z 2003 Acta Phys. Sin. 52 1694 Google Scholar
[31]	Massines F, Rabehi A, Decomps P, Gadri R B, Ségur P, Mayoux C 1998 J. Appl. Phys. 83 2950 Google Scholar
[32]	Zhang Y T, Wang D Z, Kong M G 2006 J. Appl. Phys. 100 063304 Google Scholar
[33]	He J, Zhang Y T 2012 Plasma Process. Polym. 9 919 Google Scholar
[34]	Wang Y, Zhang Y, Wang D Z, Kong M G 2007 Appl. Phys. Lett. 90 071501 Google Scholar
[35]	Yuan X, Raja L L 2003 IEEE Trans. Plasma Sci. 31 495 Google Scholar
[36]	Song S, Guo Y, Choe W, Zhang J, Zhang J, Shi J 2012 Phys. Plasma 19 123508 Google Scholar
[37]	Zhang Y T, Wang Y H 2018 Phys. Plasma 25 023509 Google Scholar
[38]	Simeni M S, Zheng Y, Barnat E V, Bruggeman P J 2021 Plasma Sources Sci. Technol. 30 055004 Google Scholar
[39]	王艳辉 2006 博士学位论文 (大连: 大连理工大学) Wang Y H 2006 Ph. D. Dissertation (Dalian: Dalian University of Technology) (in Chinese)
[40]	Vanraes P, Nikiforov A, Bogaerts A, Leys C 2018 Sci. Rep. 8 1 Google Scholar
[41]	Massines F, Segur P, Gherardi N, Khamphan C, Ricard A 2003 Surf. Coat. Tech. 174 8 Google Scholar
[42]	徐学基, 诸定昌 1996 气体放电物理 (上海: 复旦大学出版社) 第277页 Xu X J, Zhu D C 1996 Discharge Physics of Gas (Shanghai: Fudan University Press) p277 (in Chinese)
[43]	刘勇, 何湘宁, 马飞 2005 高电压技术 31 55 Google Scholar Liu Y, He X N, Ma F 2005 High Volt. Eng. 31 55 Google Scholar
[44]	Sadeghi B 2000 J. Mater. Process. Technol. 103 411 Google Scholar
[45]	Gawehn E, Hiss J A, Brown J B, Schneider G 2018 Expert Opin. Drug Discovery 13 579 Google Scholar
[46]	He J, Hu J T, Liu D W, Zhang Y T 2013 Plasma Sources Sci. Technol. 22 035008 Google Scholar
[47]	Golubovskii Y B, Maiorov V, Behnke J, Behnke J 2002 J. Phys. D Appl. Phys. 36 39 Google Scholar

施引文献

图 1 训练数据

Fig. 1. Training data

下载: 全尺寸图片幻灯片

图 2 基于大气压介质阻挡放电特性构造的人工神经网络图

Fig. 2. Diagram of artificial neural network constructed based on the characteristics of atmospheric pressure dielectric barrier discharge

下载: 全尺寸图片幻灯片

图 3 通过机器学习预测He等离子体在频率为10 kHz下的电流密度, 并与流体模拟的结果进行比较　(a) V₀ = 2150 V下的电流密度; (b) V₀ = 2450 V下的电流密度

Fig. 3. Prediction of current density of He plasma at ambient pressure (f = 10 kHz) via machine learning with comparison of the results by fluid simulation: (a) Current density at ambient pressure (V₀ = 2150 V); (b) current density at ambient pressure (V₀ = 2450 V)

下载: 全尺寸图片幻灯片

图 4 通过机器学习预测He等离子体在f = 10 kHz下的电子密度、离子密度和电场分布, 并与流体模拟的结果进行比较　(a) 电子密度与离子密度; (b) 电场强度

Fig. 4. Prediction of electron density, ion density and electric field of He plasma at ambient pressure (f = 10 kHz) via machine learning with comparison of the results by fluid simulation: (a) Electron density and ion density; (b) electric field

下载: 全尺寸图片幻灯片

图 5 电流密度峰值随电压幅值的变化

Fig. 5. Peak current density as a function of amplitude of applied voltage

下载: 全尺寸图片幻灯片

图 6 电流密度峰值时刻的最大电子密度与最大电场强度随电压幅值的变化

Fig. 6. Maximum electron density and electric field at the moment when the current density gets to the peak value as a function of amplitude of applied voltage

下载: 全尺寸图片幻灯片

图 7 通过机器学习预测He等离子体在V₀ = 2000 V下的电流密度, 并与流体模拟的结果进行比较　(a) f = 14.05 kHz下的电流密度; (b) f = 29.05 kHz下的电流密度

Fig. 7. Prediction of current density of He plasma at ambient pressure (V₀ = 2000 V) via machine learning with comparison of the results by fluid simulation: (a) Current density at ambient pressure (f = 14.05 kHz); (b) current density at ambient pressure (f = 29.05 kHz)

下载: 全尺寸图片幻灯片

图 8 通过机器学习预测He等离子体在V₀ = 2000 V下的电子密度、离子密度和电场分布, 并与流体模型模拟的结果进行比较　(a)电子密度与离子密度; (b) 电场强度

Fig. 8. Prediction of electron density, ion density and electric field of He plasma at ambient pressure (V₀ = 2000 V) via machine learning with comparison of the results by fluid simulation: (a) Electron density and ion density; (b) electric field

下载: 全尺寸图片幻灯片

图 9 不同频率下的电流密度峰值

Fig. 9. Peak current density at different frequencies

下载: 全尺寸图片幻灯片

图 10 不同频率下最大电流时刻的电子密度峰值和电场强度峰值

Fig. 10. Peak electron density and peak electric field intensity at the moment of maximum current at different frequencies

下载: 全尺寸图片幻灯片

图 11 通过机器学习预测He等离子体在多输入属性变化下的电流密度, 并与流体模拟的结果进行比较　(a) f = 14.05 kHz下的电流密度; (b) f = 29.05 kHz下的电流密度

Fig. 11. Prediction of current density of He plasma under the change of multi input attributes machine learning with comparison of the results by fluid simulation: (a) Current density at ambient pressure (f = 14.05 kHz); (b) current density at ambient pressure (f = 29.05 kHz)

下载: 全尺寸图片幻灯片

图 12 通过机器学习预测He等离子体在多输入属性变化下的电子密度和离子密度, 并与流体模拟的结果进行比较　(a) f = 14.05 kHz的电子密度; (b) f = 29.05 kHz的离子密度

Fig. 12. Prediction of electron density and ion density of He plasma under the change of multi input attributes via machine learning with comparison of the results by fluid simulation: (a) Electron density at ambient pressure (f = 14.05 kHz); (b) ion density at ambient pressure (f = 29.05 kHz)

下载: 全尺寸图片幻灯片

图 13 通过机器学习预测He等离子体在多输入参数变化下的电场强度分布, 并与流体模拟的结果进行比较　(a) f = 14.05 kHz的场强分布; (b) f = 29.05 kHz的场强分布

Fig. 13. Prediction of electric field of He plasma under the change of multi input attributes via machine learning with comparison of the results by fluid simulation: (a) Electric field at ambient pressure (f = 14.05 kHz); (b) electric field at ambient pressure (f = 29.05 kHz)

下载: 全尺寸图片幻灯片

[1]	Von Woedtke T, Metelmann H R, Weltmann K D 2014 Contrib. Plasma Phys. 54 104 Google Scholar
[2]	Agarwal P, Girshick S L 2014 Plasma Chem. Plasma Process. 34 489 Google Scholar
[3]	Chen Q, Li J, Li Y 2015 J. Phys. D Appl. Phys. 48 424005 Google Scholar
[4]	Von Keudell A, Schulz-Von Der Gathen V 2017 Plasma Sources Sci. Technol. 26 113001 Google Scholar
[5]	Bruggeman P J, Iza F, Brandenburg R 2017 Plasma Sources Sci. Technol. 26 123002 Google Scholar
[6]	Zhang Y T, Chi Y Y, He J 2014 Plasma Process. Polym. 11 639 Google Scholar
[7]	Wang X C, Bai J X, Zhang T H, Sun Y, Zhang Y T 2022 Vacuum 203 111200 Google Scholar
[8]	张泰恒, 王绪成, 张远涛 2021 物理学报 70 215201 Google Scholar Zhang T H, Wang X C, Zhang Y T 2021 Acta Phys. Sin. 70 215201 Google Scholar
[9]	Iqbal M M, Turner M M 2015 Plasma Process. Polym. 12 1104 Google Scholar
[10]	Zhang Y T, Wang D Z, Kong M G 2005 J. Appl. Phys. 98 113308 Google Scholar
[11]	Alves L, Bogaerts A, Guerra V, Turner M 2018 Plasma Sources Sci. Technol. 27 023002 Google Scholar
[12]	Wang G, Kuang Y, Zhang Y T 2019 Plasma Sci. Technol. 22 015404 Google Scholar
[13]	Brodtkorb A R, Hagen T R, Sætra M L 2013 J. Parallel Distrib. Comput. 73 4 Google Scholar
[14]	Pandey M, Fernandez M, Gentile F, Isayev O, Tropsha A, Stern A C, Cherkasov A 2022 Nat. Mach. Intell. 4 211 Google Scholar
[15]	Eklund A, Dufort P, Forsberg D, LaConte S M 2013 Med. Image Anal. 17 1073 Google Scholar
[16]	Carleo G, Cirac I, Cranmer K, Daudet L, Schuld M, Tishby N, Vogt-Maranto L, Zdeborová L 2019 Rev. Mod. Phys. 91 045002 Google Scholar
[17]	Piccione A, Berkery J, Sabbagh S, Andreopoulos Y 2020 Nucl. Fusion 60 046033 Google Scholar
[18]	Fu Y, Eldon D, Erickson K, Kleijwegt K, Lupin-Jimenez L, Boyer M D, Eidietis N, Barbour N, Izacard O, Kolemen E 2020 Phys. Plasma 27 022501 Google Scholar
[19]	Mesbah A, Graves D B 2019 J. Phys. D Appl. Phys. 52 30LT02 Google Scholar
[20]	Jordan M I, Mitchell T M 2015 Science 349 255 Google Scholar
[21]	Wang H, Lei Z, Zhang X, Zhou B, Peng J 2019 Energy Convers. Manage. 198 111799 Google Scholar
[22]	Bonzanini A D, Shao K, Stancampiano A, Graves D B, Mesbah A 2021 IEEE Trans. Radiat. Plasma Med. Sci. 6 16 Google Scholar
[23]	Hong Y, Hou B, Jiang H, Zhang J 2020 Wiley Interdiscip. Rev. Comput. Mol. Sci. 10 e1450 Google Scholar
[24]	Abiodun O I, Jantan A, Omolara A E, Dada K V, Mohamed N A, Arshad H 2018 Heliyon 4 e00938 Google Scholar
[25]	Dongare A, Kharde R, Kachare A D 2012 IJEIT 2 189
[26]	Kukreja H, Bharath N, Siddesh C, Kuldeep S 2016 Int. J. Adv. Res. Innov. Ideas Educ. 1 27
[27]	Han J, Jentzen A, Weinan E 2018 Proc. Natl. Acad. Sci. 115 8505 Google Scholar
[28]	Zhong L, Gu Q, Wu B 2020 Comput. Phys. Commun. 257 107496 Google Scholar
[29]	张远涛, 王德真, 王艳辉 2005 物理学报 54 4808 Google Scholar Zhang Y T, Wang D Z, Wang Y H 2005 Acta Phys. Sin. 54 4808 Google Scholar
[30]	王艳辉, 王德真 2003 物理学报 52 1694 Google Scholar Wang Y H, Wang D Z 2003 Acta Phys. Sin. 52 1694 Google Scholar
[31]	Massines F, Rabehi A, Decomps P, Gadri R B, Ségur P, Mayoux C 1998 J. Appl. Phys. 83 2950 Google Scholar
[32]	Zhang Y T, Wang D Z, Kong M G 2006 J. Appl. Phys. 100 063304 Google Scholar
[33]	He J, Zhang Y T 2012 Plasma Process. Polym. 9 919 Google Scholar
[34]	Wang Y, Zhang Y, Wang D Z, Kong M G 2007 Appl. Phys. Lett. 90 071501 Google Scholar
[35]	Yuan X, Raja L L 2003 IEEE Trans. Plasma Sci. 31 495 Google Scholar
[36]	Song S, Guo Y, Choe W, Zhang J, Zhang J, Shi J 2012 Phys. Plasma 19 123508 Google Scholar
[37]	Zhang Y T, Wang Y H 2018 Phys. Plasma 25 023509 Google Scholar
[38]	Simeni M S, Zheng Y, Barnat E V, Bruggeman P J 2021 Plasma Sources Sci. Technol. 30 055004 Google Scholar
[39]	王艳辉 2006 博士学位论文 (大连: 大连理工大学) Wang Y H 2006 Ph. D. Dissertation (Dalian: Dalian University of Technology) (in Chinese)
[40]	Vanraes P, Nikiforov A, Bogaerts A, Leys C 2018 Sci. Rep. 8 1 Google Scholar
[41]	Massines F, Segur P, Gherardi N, Khamphan C, Ricard A 2003 Surf. Coat. Tech. 174 8 Google Scholar
[42]	徐学基, 诸定昌 1996 气体放电物理 (上海: 复旦大学出版社) 第277页 Xu X J, Zhu D C 1996 Discharge Physics of Gas (Shanghai: Fudan University Press) p277 (in Chinese)
[43]	刘勇, 何湘宁, 马飞 2005 高电压技术 31 55 Google Scholar Liu Y, He X N, Ma F 2005 High Volt. Eng. 31 55 Google Scholar
[44]	Sadeghi B 2000 J. Mater. Process. Technol. 103 411 Google Scholar
[45]	Gawehn E, Hiss J A, Brown J B, Schneider G 2018 Expert Opin. Drug Discovery 13 579 Google Scholar
[46]	He J, Hu J T, Liu D W, Zhang Y T 2013 Plasma Sources Sci. Technol. 22 035008 Google Scholar
[47]	Golubovskii Y B, Maiorov V, Behnke J, Behnke J 2002 J. Phys. D Appl. Phys. 36 39 Google Scholar

计量

文章访问数: 2398
PDF下载量: 62
被引次数: 0

姓名
邮箱
手机号码
标题
留言内容
验证码

搜索

留言板

结合机器学习的大气压介质阻挡放电数值模拟研究