-
容性耦合等离子体（CCP）的流体模拟对于理解放电物理机制非常重要，但其高昂的计算成本制约了大范围参数化探索。为突破该限制，本研究开发了一种深度学习代理模型，旨在以近瞬时推理速度复现一维CCP流体模型的输出结果。该模型精确预测了容性耦合氩等离子体流体模拟中关键等离子体参数的空间分布，包括电子密度、电子温度及电场分布，并将所需计算时间从数小时压缩至毫秒量级。除加速优势外，代理模型学习过程还揭示了根植于等离子体物理的非对称推理能力。代理模型可从复杂的低压物理域外推至更简单的高压物理域，反之则不可行，表明低压状态具有更完整的物理信息。进一步，本研究建立一个模型推理的置信边界，确保预测结果的物理可靠性。最终，本研究为创建高保真、超快速的流体模拟等离子体替代提供了方案。Fluid simulations of capacitively coupled plasmas (CCP) are crucial for understanding their discharge physics, yet the high computational cost poses a major bottleneck. To overcome this limitation, we have developed a deep learning-based surrogate model designed to replicate the output of a onedimensional CCP fluid model with near-instantaneous inference speed. Through a systematic evaluation of three architectures—Feedforward Neural Network (FNN), Attention-enhanced Long Short-Term Memory network (ALSTM), and Convolutional-Transformer hybrid network (CTransformer)—we found that the sequence-structured ALSTM model achieves the optimal balance between speed and accuracy, with an overall prediction error of only 1.73% for electron density, electric field, and electron temperature in argon discharge. This study not only achieves simulation acceleration but also reveals that the model can accu- rately extrapolate from low-pressure conditions dominated by complex non-local effects to high-pressure conditions governed by simple local effects, whereas the reverse extrapolation fails. This phenomenon suggests that training under low-pressure conditions enables the model to capture more comprehensive data features. From the perspective of model weights, both low-pressure and high-pressure models assign key weights to the sheath region, but the low-pressure model exhibits higher weight peaks in the sheath, indicating a stronger ability to capture the key physical process of sheath dynamics. In contrast, the high-pressure model, due to its lower weights in the sheath region, may fail to adequately resolve the complex dynamics of the sheath when predicting new operating conditions, thereby limiting its ability for high-fidelity extrapolation. To ensure the reliability of this data-driven model in practical applications, we established a trust boundary of 5% normalized mean absolute spatial error for model performance through systematic extrapolation experiments. When the model’s extrapolation error falls below this threshold, the spatial distribution curves of predicted parameters such as electron density and electron temperature closely match the true physical distributions. However, once the error exceeds this critical point, systematic deviations such as morphological distortion and amplitude discrepancies begin to appear in the predicted spatial distributions, significantly diverging from the true physical laws. In the future, we will develop neural network models capable of processing high-dimensional spatial data and incorporating multi-dimensional input features such as various discharge gases, ultimately realizing a dedicated AI model for the field of capacitively coupled plasmas.
-
[1] Kim H C, Iza F and Yang S S, Radmilović-Radjenović M and Lee J K 2005 J. Phys. D: Appl. Phys. 38 R283
[2] Wang X C and Zhang Y T 2023 J. Appl. Phys. 133143301
[3] Zhang Y T, Gao S H and Zhu Y Y 2023 J. Appl. Phys. 133053303
[4] Sethi S P, Das D P and Behera S K 2023 IEEE Trans. Plasma Sci. 511434
[5] Kim B, Im S and Yoo G 2021 Electronics 1049
[6] Xiao T Q, Wu Z, Christofides P D, Armaou A and Ni D 2021 Ind. Eng. Chem. Res. 61638
[7] Liau L C K, Huang C J, Chen C C, Huang C S, Chen T, Lin S C and Kuo L C 2002 Sol. Energy Mater. Sol. Cells 71169
[8] Pan J, Liu Y, Zhang S, Hu X C, Liu Y D and Shao T 2023 Energy Convers. Manage. 277116620.
[9] Wan C G, Yu Z, Wang F, Liu X J and Li J G 2021 Nucl. Fusion 61066015
[10] Yang Y, Yang S, Li C and Yu Z H 2021 IEEE Access 967232
[11] Borghei M and Ghassemi M 2022 IEEE Trans. Dielectr. Electr. Insul. 29319
[12] Seo J, Kim S, Jalalvand A, Conlin R, Rothstein A, Abbate J, Erickson K, Wai J, Shousha R and Kolemen E 2024 Nature 626746
[13] Dave B, Patel S, Shivani R, Purohit S and Chaudhury B 2022 Contrib. Plasma Phys. 63 e202200051
[14] Marvin M and Seymour P 1969 MIT Press 6318
[15] Riedmiller M 1994 Computer Standards & Interfaces 16265
[16] Krizhevsky A, Sutskever I and Hinton G E 2012 Advances in neural information processing systems 25
[17] Elman J L 1990 Cognitive science 14179
[18] Hpchreiter S and Schmidhuber J 1997 Neural computation 91735
[19] Vaswani A, Shazeer N, Parmar N, Uszkoreit J, Jones L, Gomez A N, Kaiser Ł and Polosukhin 2017 Advances in neural information processing systems 30
[20] Liu Z M, Wang Y X, Vaidya S, Ruehle F, Halverson J, Soljačić M, Hou Y T and Tegmark M 20242404.19756v5[cs.LG]
[21] Von R L, Maryer S, Beckh K, Georgiev B, Giesselbach S, Heese R, Kirsch B, Pfrommer J, Pick A, Ramamurthy R, Walczak M, Garcke J, Bauckhage C and Schuecker J. 2021 IEEE Trans. Knowl. Data Eng. 35614
[22] Willard J, Jia X, Xu S M, Steinbach M and Kumar V 2020 arXiv:2003.04919v6[physics.comp-ph]
[23] Basir S 2022 arXiv:2209.09988v3[cs.LG]
[24] Luo X, Yuan S Y, Tang H W, Xu D, Ran Q H, Cen Y H and Ling D F 2024 Hydrol. Processes 38 e151431
[25] Li J, Wu X and Li Z 2025 Appl. Ocean Res. 161104661
[26] Guo Y, Li L, Xiang Z, Gui J, Shi S, Lei Z and Xu X 2025 Atomic Energy Science and Technology 591085
[27] Li W K and Zhang Y T 2025 J. Appl. Phys. 137203304
[28] Noh H, Lee J and Yoon E 2025 J. Comput. Phys. 523113665
[29] Shi H Y, Wang S and Wang P Y 2024 J. Environ. Chem. Eng. 12112998
[30] Marcus G 2018 arXiv:1801.00631v1[cs.AI]
[31] Battaglia P W, Hamrick J B, Bapst V, Sanchez-Gonzalez A, Zambaldi V, Malinowski M, Tacchetti A, Raposo D, Santoro A, Faulkner R, Gulcehre C, Song F, Ballard A, Gilmer J, Dahl G, Vaswani A, Allen K, Nash C, Langston V, Dyer C, Heess N, Wierstra D, Kohli P, Botvinick M, Vinyals O, Li Y, Pascanu R. 2018 arXiv:1806.01261v3[cs.LG]
[32] Xu K, Zhang M, Li J, Du S S, Kawarabayashi K I and Jegelka S 2020 arXiv:2009.11848v5[cs.LG]
[33] Wu Y, Zhu Z, Liu F, Chrysos G and Cevher V 2022 Advances in neural information processing systems 3526980
[34] Hestness J, Narang S, Ardalani N, Diamos G, Jun H, Kianinejad H, Patwary M M A, Yang Y and Zhou Y Q 2017 arXiv:1712.00409v1[cs.LG]
[35] Manzhos S and Ihara M 2023 J. Phys. Chem. A 1277823
[36] Jain S and Wallace B C 2019 arXiv:1902.10186v3[cs.CL]
[37] Donko Z, Derzsi A, Vass M, Horvth B, Wilczek S, Horvath B and Hartmann P 2021 Plasma Sources Sci. Technol 30095017
[38] Jia W Z, Zhang Q Z, Wang X F, Song Y H and Wang Y N 2019 J. Phys. D: Appl. Phys 52015206
计量
- 文章访问数: 21
- PDF下载量: 2
- 被引次数: 0
下载:
