|
|
用于非局部Allen-Cahn方程的能量耗散型EDNN方法
|
Abstract:
非局部Allen-Cahn的数值模拟在实际应用中得到了广泛应用。然而，开发一种既高效又符合物理定律的数值方法仍是一个重大挑战。近年来，利用神经网络求解偏微分方程显示出了巨大的潜力。受这些研究的启发，我们提出在Evolutionary Deep Neural Network (EDNN)中引入一个辅助变量，以保持偏微分方程的基本物理特性。该方法确保离散数值格式具有无条件能量耗散特性，从而将问题框定为一个最小化任务。我们对非局部Allen-Cahn方程进行了数值模拟，验证了我们修正过的EDNN的准确性和效率。
The numerical simulation of nonlocal Allen-Cahn equations has been widely applied in practical applications. However, developing an efficient numerical method that adheres to physical laws remains a significant challenge. Recently, the use of neural networks to solve partial differential equations has demonstrated great potential. Inspired by these studies, we propose incorporating an auxiliary variable into the Evolutionary Deep Neural Network (EDNN) framework to preserve the fundamental physical properties of partial differential equations. This approach ensures that the discrete numerical scheme possesses an unconditionally energy dissipation property, thereby framing the problem as a minimization task. To validate the accuracy and efficiency of our modified EDNN, we conducted numerical simulations of the nonlocal Allen-Cahn equation.
| [1] | Wang, N., Chang, H. and Zhang, D. (2021) Deep‐Learning‐Based Inverse Modeling Approaches: A Subsurface Flow Example. Journal of Geophysical Research: Solid Earth, 126, JB020549. https://doi.org/10.1029/2020jb020549 |
| [2] | Xu, H., Chang, H. and Zhang, D. (2021) DL-PDE: Deep-Learning Based Data-Driven Discovery of Partial Differential Equations from Discrete and Noisy Data. Communications in Computational Physics, 29, 698-728. https://doi.org/10.4208/cicp.oa-2020-0142 |
| [3] | Raissi, M., Perdikaris, P. and Karniadakis, G.E. (2017) Physics Informed Deep Learning (Part I): Data-Driven Solutions of Nonlinear Partial Differential Equations. arXiv: 1711.10561. |
| [4] | Goswami, S., Anitescu, C., Chakraborty, S. and Rabczuk, T. (2020) Transfer Learning Enhanced Physics Informed Neural Network for Phase-Field Modeling of Fracture. Theoretical and Applied Fracture Mechanics, 106, Article 102447. https://doi.org/10.1016/j.tafmec.2019.102447 |
| [5] | Qiu, R., Huang, R., Xiao, Y., Wang, J., Zhang, Z., Yue, J., et al. (2022) Physics-Informed Neural Networks for Phase-Field Method in Two-Phase Flow. Physics of Fluids, 34, Article 052109. https://doi.org/10.1063/5.0091063 |
| [6] | Kiyani, E., Silber, S., Kooshkbaghi, M. and Karttunen, M. (2022) Machine-Learning-Based Data-Driven Discovery of Nonlinear Phase-Field Dynamics. Physical Review E, 106, Article 065303. https://doi.org/10.1103/physreve.106.065303 |
| [7] | Du, Y. and Zaki, T.A. (2021) Evolutional Deep Neural Network. Physical Review E, 104, Article 045303. https://doi.org/10.1103/physreve.104.045303 |
| [8] | Allen, S.M. and Cahn, J.W. (1979) A Microscopic Theory for Antiphase Boundary Motion and Its Application to Antiphase Domain Coarsening. Acta Metallurgica, 27, 1085-1095. https://doi.org/10.1016/0001-6160(79)90196-2 |
| [9] | Cheng, Q., Liu, C. and Shen, J. (2021) Generalized SAV Approaches for Gradient Systems. Journal of Computational and Applied Mathematics, 394, Article 113532. https://doi.org/10.1016/j.cam.2021.113532 |
| [10] | Meng, X., Cheng, A. and Liu, Z. (2023) The Stabilized Exponential-SAV Approach Preserving Maximum Bound Principle for Nonlocal Allen-Cahn Equation. arXiv:2307.13934. |
