Multi-Scale Deep Neural Network (MscaleDNN) Methods for Oscillatory Stokes Flows in Complex Domains

Multi-Scale Deep Neural Network (MscaleDNN) Methods for Oscillatory Stokes Flows in Complex Domains

Year:    2020

Author:    Bo Wang, Wenzhong Zhang, Wei Cai

Communications in Computational Physics, Vol. 28 (2020), Iss. 5 : pp. 2139–2157

Abstract

In this paper, we study a multi-scale deep neural network (MscaleDNN) as a meshless numerical method for computing oscillatory Stokes flows in complex domains. The MscaleDNN employs a multi-scale structure in the design of its DNN using radial scalings to convert the approximation of high frequency components of the highly oscillatory Stokes solution to one of lower frequencies. The MscaleDNN solution to the Stokes problem is obtained by minimizing a loss function in terms of $L^2$ norm of the residual of the Stokes equation. Three forms of loss functions are investigated based on vorticity-velocity-pressure, velocity-stress-pressure, and velocity-gradient of velocity-pressure formulations of the Stokes equation. We first conduct a systematic study of the MscaleDNN methods with various loss functions on the Kovasznay flow in comparison with normal fully connected DNNs. Then, Stokes flows with highly oscillatory solutions in a 2-D domain with six randomly placed holes are simulated by the MscaleDNN. The results show that MscaleDNN has faster convergence and consistent error decays in the simulation of Kovasznay flow for all three tested loss functions. More importantly, the MscaleDNN is capable of learning highly oscillatory solutions when the normal DNNs fail to converge.

You do not have full access to this article.

Already a Subscriber? Sign in as an individual or via your institution

Journal Article Details

Publisher Name:    Global Science Press

Language:    English

DOI:    https://doi.org/10.4208/cicp.OA-2020-0192

Communications in Computational Physics, Vol. 28 (2020), Iss. 5 : pp. 2139–2157

Published online:    2020-01

AMS Subject Headings:    Global Science Press

Copyright:    COPYRIGHT: © Global Science Press

Pages:    19

Keywords:    Deep neural network Stokes equation multi-scale meshless methods.

Author Details

Bo Wang

Wenzhong Zhang

Wei Cai

  1. Kronecker Neural Networks Overcome Spectral Bias for PINN-Based Wavefield Computation

    Waheed, Umair Bin

    IEEE Geoscience and Remote Sensing Letters, Vol. 19 (2022), Iss. P.1

    https://doi.org/10.1109/LGRS.2022.3209901 [Citations: 14]
  2. Fourier warm start for physics-informed neural networks

    Jin, Ge | Wong, Jian Cheng | Gupta, Abhishek | Li, Shipeng | Ong, Yew-Soon

    Engineering Applications of Artificial Intelligence, Vol. 132 (2024), Iss. P.107887

    https://doi.org/10.1016/j.engappai.2024.107887 [Citations: 3]
  3. Three ways to solve partial differential equations with neural networks — A review

    Blechschmidt, Jan | Ernst, Oliver G.

    GAMM-Mitteilungen, Vol. 44 (2021), Iss. 2

    https://doi.org/10.1002/gamm.202100006 [Citations: 119]
  4. Phase space approach to solving higher order differential equations with artificial neural networks

    Tori, Floriano | Ginis, Vincent

    Physical Review Research, Vol. 4 (2022), Iss. 4

    https://doi.org/10.1103/PhysRevResearch.4.043090 [Citations: 1]
  5. A Multi-scale Deep Neural Network Technique for Surrogate Modeling of Microwave Components

    Li, Zheng | Li, Xiao-Chun | Wu, Ze-Ming

    2024 Photonics & Electromagnetics Research Symposium (PIERS), (2024), P.1

    https://doi.org/10.1109/PIERS62282.2024.10617970 [Citations: 0]
  6. Physics-informed neural network combined with characteristic-based split for solving forward and inverse problems involving Navier–Stokes equations

    Hu, Shuang | Liu, Meiqin | Zhang, Senlin | Dong, Shanling | Zheng, Ronghao

    Neurocomputing, Vol. 573 (2024), Iss. P.127240

    https://doi.org/10.1016/j.neucom.2024.127240 [Citations: 2]
  7. Subspace Decomposition Based Dnn Algorithm for Elliptic-Type Multi-Scale Pdes

    Li, Xi-An | Xu, Zhi-Qin John | Zhang, Lei

    SSRN Electronic Journal, Vol. (2022), Iss.

    https://doi.org/10.2139/ssrn.4020731 [Citations: 3]
  8. Kronecker neural networks for the win: Overcoming spectral bias for PINN-based wavefield computation

    Waheed, Umair bin

    Second International Meeting for Applied Geoscience & Energy, (2022), P.1644

    https://doi.org/10.1190/image2022-3747298.1 [Citations: 0]
  9. Physics-informed machine learning

    Karniadakis, George Em | Kevrekidis, Ioannis G. | Lu, Lu | Perdikaris, Paris | Wang, Sifan | Yang, Liu

    Nature Reviews Physics, Vol. 3 (2021), Iss. 6 P.422

    https://doi.org/10.1038/s42254-021-00314-5 [Citations: 2689]
  10. Deep Learning for Solving and Estimating Dynamic Macro-finance Models

    Fan, Benjamin | Qiao, Edward | Jiao, Anran | Gu, Zhouzhou | Li, Wenhao | Lu, Lu

    Computational Economics, Vol. (2024), Iss.

    https://doi.org/10.1007/s10614-024-10693-3 [Citations: 0]
  11. MPIPN: a multi physics-informed PointNet for solving parametric acoustic-structure systems

    Wang, Chu | Wu, Jinhong | Wang, Yanzhi | Zha, Zhijian | Zhou, Qi

    Engineering with Computers, Vol. (2024), Iss.

    https://doi.org/10.1007/s00366-024-01998-w [Citations: 0]
  12. Physics‐Informed Neural Networks (PINNs) for Wave Propagation and Full Waveform Inversions

    Rasht‐Behesht, Majid | Huber, Christian | Shukla, Khemraj | Karniadakis, George Em

    Journal of Geophysical Research: Solid Earth, Vol. 127 (2022), Iss. 5

    https://doi.org/10.1029/2021JB023120 [Citations: 144]
  13. A multiscale differential‐algebraic neural network‐based method for learning dynamical systems

    Huang, Yin | Ding, Jieyu

    International Journal of Mechanical System Dynamics, Vol. 4 (2024), Iss. 1 P.77

    https://doi.org/10.1002/msd2.12102 [Citations: 1]
  14. A Physics-Informed Neural Network for Quantifying the Microstructural Properties of Polycrystalline Nickel Using Ultrasound Data: A promising approach for solving inverse problems

    Shukla, Khemraj | Jagtap, Ameya D. | Blackshire, James L. | Sparkman, Daniel | Em Karniadakis, George

    IEEE Signal Processing Magazine, Vol. 39 (2022), Iss. 1 P.68

    https://doi.org/10.1109/MSP.2021.3118904 [Citations: 68]
  15. A deep domain decomposition method based on Fourier features

    Li, Sen | Xia, Yingzhi | Liu, Yu | Liao, Qifeng

    Journal of Computational and Applied Mathematics, Vol. 423 (2023), Iss. P.114963

    https://doi.org/10.1016/j.cam.2022.114963 [Citations: 14]
  16. Isogeometric neural networks: A new deep learning approach for solving parameterized partial differential equations

    Gasick, Joshua | Qian, Xiaoping

    Computer Methods in Applied Mechanics and Engineering, Vol. 405 (2023), Iss. P.115839

    https://doi.org/10.1016/j.cma.2022.115839 [Citations: 7]
  17. Physics-informed neural network combined with characteristic-based split for solving Navier–Stokes equations

    Hu, Shuang | Liu, Meiqin | Zhang, Senlin | Dong, Shanling | Zheng, Ronghao

    Engineering Applications of Artificial Intelligence, Vol. 128 (2024), Iss. P.107453

    https://doi.org/10.1016/j.engappai.2023.107453 [Citations: 8]
  18. Subspace decomposition based DNN algorithm for elliptic type multi-scale PDEs

    Li, Xi-An | Xu, Zhi-Qin John | Zhang, Lei

    Journal of Computational Physics, Vol. 488 (2023), Iss. P.112242

    https://doi.org/10.1016/j.jcp.2023.112242 [Citations: 6]
  19. General-Kindred physics-informed neural network to the solutions of singularly perturbed differential equations

    Wang, Sen | Zhao, Peizhi | Ma, Qinglong | Song, Tao

    Physics of Fluids, Vol. 36 (2024), Iss. 11

    https://doi.org/10.1063/5.0235410 [Citations: 0]
  20. Solving a class of multi-scale elliptic PDEs by Fourier-based mixed physics informed neural networks

    Li, Xi'an | Wu, Jinran | Tai, Xin | Xu, Jianhua | Wang, You-Gan

    Journal of Computational Physics, Vol. 508 (2024), Iss. P.113012

    https://doi.org/10.1016/j.jcp.2024.113012 [Citations: 1]
  21. Modified failproof physics-informed neural network framework for fast and accurate optical fiber transmission link modeling

    Uduagbomen, Joshua | Leeson, Mark S. | Liu, Zheng | Lakshminarayana, Subhash | Xu, Tianhua

    Applied Optics, Vol. 63 (2024), Iss. 14 P.3794

    https://doi.org/10.1364/AO.524426 [Citations: 0]
  22. Solving the one dimensional vertical suspended sediment mixing equation with arbitrary eddy diffusivity profiles using temporal normalized physics-informed neural networks

    Zhang, Shaotong | Deng, Jiaxin | Li, Xi'an | Zhao, Zixi | Wu, Jinran | Li, Weide | Wang, You-Gan | Jeng, Dong-Sheng

    Physics of Fluids, Vol. 36 (2024), Iss. 1

    https://doi.org/10.1063/5.0179223 [Citations: 1]
  23. Flow field reconstruction from sparse sensor measurements with physics-informed neural networks

    Hosseini, Mohammad Yasin | Shiri, Yousef

    Physics of Fluids, Vol. 36 (2024), Iss. 7

    https://doi.org/10.1063/5.0211680 [Citations: 2]
  24. A practical approach to flow field reconstruction with sparse or incomplete data through physics informed neural network

    Xu, Shengfeng | Sun, Zhenxu | Huang, Renfang | Guo, Dilong | Yang, Guowei | Ju, Shengjun

    Acta Mechanica Sinica, Vol. 39 (2023), Iss. 3

    https://doi.org/10.1007/s10409-022-22302-x [Citations: 31]
  25. A Linear Frequency Principle Model to Understand the Absence of Overfitting in Neural Networks

    Zhang, Yaoyu | Luo, Tao | Ma, Zheng | Xu, Zhi-Qin John

    Chinese Physics Letters, Vol. 38 (2021), Iss. 3 P.038701

    https://doi.org/10.1088/0256-307X/38/3/038701 [Citations: 10]
  26. Overview Frequency Principle/Spectral Bias in Deep Learning

    Xu, Zhi-Qin John | Zhang, Yaoyu | Luo, Tao

    Communications on Applied Mathematics and Computation, Vol. (2024), Iss.

    https://doi.org/10.1007/s42967-024-00398-7 [Citations: 2]
  27. Computational Science – ICCS 2023

    Hierarchical Learning to Solve PDEs Using Physics-Informed Neural Networks

    Han, Jihun | Lee, Yoonsang

    2023

    https://doi.org/10.1007/978-3-031-36024-4_42 [Citations: 1]
  28. Physics-informed deep learning of rate-and-state fault friction

    Rucker, Cody | Erickson, Brittany A.

    Computer Methods in Applied Mechanics and Engineering, Vol. 430 (2024), Iss. P.117211

    https://doi.org/10.1016/j.cma.2024.117211 [Citations: 2]