Neural Network Algorithms for Awonisotropic Diffusion Equations

Ouyang Haocong; Fang Jinwei

doi:10.12052/gdutxb.240155

English Version Journal Login

您当前的位置：

首页 >

文章列表页 >

Neural Network Algorithms for Awonisotropic Diffusion Equations

更新时间：2026-05-27

- Neural Network Algorithms for Awonisotropic Diffusion Equations
- Vol. 42, Issue 6, Pages: 70-77, 125(2025)
- 作者机构：
  
  广东工业大学数学与统计学院,广东,广州,510520
- 作者简介：
- 基金信息：
- DOI：10.12052/gdutxb.240155
  CLC： TP183
- Published：2025
- 稿件说明：
移动端阅览
欧阳浩聪, 房金伟. 各向异性扩散方程的神经网络算法[J]. 广东工业大学学报, 2025, 42(6): 70–77.

Ouyang Haocong, Fang Jinwei. Neural Network Algorithms for Awonisotropic Diffusion Equations[J]. 2025, 42(6): 70-77, 125.
欧阳浩聪, 房金伟. 各向异性扩散方程的神经网络算法[J]. 广东工业大学学报, 2025, 42(6): 70–77. DOI： 10.12052/gdutxb.240155

Ouyang Haocong, Fang Jinwei. Neural Network Algorithms for Awonisotropic Diffusion Equations[J]. 2025, 42(6): 70-77, 125. DOI： 10.12052/gdutxb.240155.

摘要

基于物理信息神经网络(Physics-Informed Neural Network

PINN) 求解各向异性扩散方程。首先提出了垂直平行采样法，对材料坐标系的不同方向进行数据点采集。实验结果表明，相比传统的随机采样方法，该方法能够准确捕捉问题中各向异性变化特性，提高神经网络预测精度。其次提出了加权积分离散法，与普通的PINN方法不同，该方法将计算数据点误差改为计算积分型误差函数，通过嵌入权函数并使用高斯积分数值方法进行离散。实验结果表明，使用该方法预测的结果准确性具有较大的提升，并且收敛稳定性更强和收敛速度更快。

Abstract

In this research

the anisotropic diffusion equation is solved based on physics-informed neural network (PINN) . Firstly

a vertical parallel sampling method is proposed to collect data points in different directions of the material coordinate system. The experimental results show that the method can accurately capture the anisotropic variation characteristics in the problem and improve the prediction accuracy of the neural network compared with the traditional random sampling method. Secondly

the weighted integral discretization method is proposed

which is different from the ordinary PINN method in that it changes the calculation of the data point error to the calculation of the integral-type error function

which is discretized by embedding the weight function and using the Gaussian integral numerical method. The experimental results show that the accuracy of the predicted results using this method has a significant improvement

and the convergence stability is stronger and the convergence speed is faster.

关键词

Keywords

references

CAO F J, YAO Y Z. Conservative positivity-preserving algorithm for Kershaw scheme of anisotropic diffusion problems[J]. Chinese Journal of Computational Physics, 2017, 34(3): 283-293.

CHANDA S, BALAJI C, VENKATESHAN S P, et al . Simultaneous estimation of principal thermal conductivities of an anisotropic composite medium: an inverse analysis[J ] . Journal of Heat Transfer, 2013, 135(2): 8.

MEIGEL F J, DARWENT T, BASTIN L, et al . Dispersive transport dynamics in porous media emerge from local correlations[J ] . Nature Communications, 2022, 13(1): 9.

ZHANG W, Al KOBAISI M. On the monotonicity and positivity of physics-informed neural networks for highly anisotropic diffusion equations[J]. Energies, 2022, 15(18): 18.

TAN J, LIU J. Solving anisotropic subdiffusion problems in annuli and shells[J]. Journal of Computational and Applied Mathematics, 2022, 401: 13.

PROTEKTOR D O, KOLODYAZHNY V M, LISIN D O, et al . A meshless method of solving three-dimensional nonstationary heat conduction problems in anisotropic materials[J ] . Cybernetics and Systems Analysis, 2021, 57(3): 470-480.

WU J, GAO Z. Interpolation-based second-order monotone finite volume schemes for anisotropic diffusion equations on general grids[J]. Journal of Computational Physics, 2014, 275: 569-588.

QI Y F, WU L, CAO H K, et al . 3D anisotropic forward modeling for direct-current resistivity method based on multiscale finite-element algorithm[J ] . Chinese Journal of Geop hysics, 2024, 67(9): 3605-3617.

ONG T H. Schémas volumes finis pour des opérateurs de diffusion anisotropes hétérogenes sur des maillages non-conformes[D]. Paris: Université de l'Est Parisien , 2012.

LE POTIER C, THANH H. A cell-centered scheme for heterogeneous anisotropic diffusion problems on general meshes[J]. International Journal on Finite Volumes, 2012, 8: 1-40.

BASSI F, REBAY S. A high-order accurate discontinuous finite element method for the numerical solution of the compressible Navier-Stokes equations[J]. Journal of Computational Physics, 1997, 131(2): 267-279.

HYMAN J, MOREL J, SHASHKOV M, et al . Mimetic finite difference methods for diffusion equations[J ] . Computational Geosciences, 2002, 6: 333-352.

YANG T, WANG Y. A new tailored finite point method for strongly anisotropic diffusion equation on misaligned grids[J]. Applied Mathematics and Computation, 2019, 355: 85-95.

VAN ES B, KOREN B, DE BLANK H J. Finite-difference schemes for anisotropic diffusion[J]. Journal of Computational Physics, 2014, 272: 526-549.

GREEN D, HU X, LORE J, et al . An efficient high-order numerical solver for diffusion equations with strong anisotropy[J ] . Computer Physics Communications, 2022, 276: 13.

HELD M, WIESENBERGER M, STEGMEIR A. Three discontinuous Galerkin schemes for the anisotropic heat conduction equation on non-aligned grids[J]. Computer Physics Communications, 2016, 199: 29-39.

GAO Z, WU J. A small stencil and extremum-preserving scheme for anisotropic diffusion problems on arbitrary 2D and 3D meshes[J]. Journal of Computational Physics, 2013, 250: 308-331.

AAVATSMARK I, BARKVE T, BOE O, et al . Discretization on non-orthogonal, quadrilateral grids for inhomogeneous, anisotropic media[J ] . Journal of Computational Physics, 1996, 127(1): 2-14.

FLORES G F G, SANCHEZ P B. New quality measures for quadrilaterals and new discrete functionals for grid generation[J]. Mathematical and Computational Applications, 2023, 28(5): 16.

SHENG Z, YUAN G. The finite volume scheme preserving extremum principle for diffusion equations on polygonal meshes[J]. Journal of Computational Physics, 2011, 230(7): 2588-2604.

ZHAO T, ZHOU Y, CHENG Y Q, et al . Solving forward and inverse problems of the heat conduction equation using physics-informed neural networks[J ] . Acta Aerodynamica Sinica, 2021, 39(5): 19-26.

RAISSI M, PERDIKARIS P, KARNIADAKIS G E. Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations[J]. Journal of Computational Physics, 2019, 378: 686-707.

NIAKI S A, HAGHIGHAT E, CAMPBELL T, et al . Physics-informed neural network for modelling the thermochemical curing process of composite-tool systems during manufacture[J ] . Computer Methods in Applied Mechanics and Engineering, 2021, 384: 23.

RAO C, SUN H, LIU Y. Physics-informed deep learning for incompressible laminar flows[J]. Theoretical and Applied Mechanics Letters, 2020, 10(3): 207-212.

HAGHIGHAT E, RAISSI M, MOURE A, et al . A physics-informed deep learning framework for inversion and surrogate modeling in solid mechanics[J ] . Computer Methods in Applied Mechanics and Engineering, 2021, 379: 22.

NICOLI K, ANDERS C J, FUNCKE L, et al . Physics-informed Bayesian optimization of variational quantum circuits[J ] . Advances in Neural Information Processing Systems, 2024, 36: 18341-18376.

SON S, LEE H, JEONG D, et al . A novel physics-informed neural network for modeling electromagnetism of a permanent magnet synchronous motor[J ] . Advanced Engineering Informatics, 2023, 57: 18.

MOSELEY B, MARKHAM A, NISSEN-MEYER T. Finite basis physics-informed neural networks (FBPINNs): a scalable domain decomposition approach for solving differential equations[J]. Advances in Computational Mathematics, 2023, 49(4): 39.

SONG J, ZHONG M, KARNIADAKIS G E, et al . Two-stage initial-value iterative physics-informed neural networks for simulating solitary waves of nonlinear wave equations[J ] . Journal of Computational Physics, 2024, 505: 22.

SHUKLA K, DI LEONI P C, BLACKSHIRE J, et al . Physics-informed neural network for ultrasound nondestructive quantification of surface breaking cracks[J ] . Journal of Nondestructive Evaluation, 2020, 39(3): 20.

JAGTAP A D, KHARAZMI E, KARNIADAKIS G E. Conservative physics-informed neural networks on discrete domains for conservation laws: applications to forward and inverse problems[J]. Computer Methods in Applied Mechanics and Engineering, 2020, 365: 27.

LU L, MENG X, MAO Z, et al . DeepXDE: a deep learning library for solving differential equations[J ] . SIAM Review, 2021, 63(1): 208-228.

FANG Y, WU G Z, WANG Y Y, et al . Data-driven femtosecond optical soliton excitations and parameters discovery of the high-order NLSE using the PINN[J ] . Nonlinear Dynamics, 2021, 105(1): 603-616.

ZHANG L H, ZHOU B, LUO Y, et al . An adaptive collocation point algorithm for physics-informed neural networks[J ] . Chinese Journal of Theoretical and Applied Mechanics, 2024, 56(10): 3069-3083.

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

The Cross-twin Physics-informed Neural Network for Wave Equations with Conserved Quantities

Research on Foreign Object Detection in Power Transmission Lines Based on the AAGV-YOLOX Model

A Study of Power Quality Regulation Strategies of Energy Storage Converters

Bimodal Iterative Cross-attention Fusion Ensemble Framework

Research and Implementation of a Pulsed Laser Ranging Method Based on Digital Mixing

Related Author

Li Jianhao

Related Institution

No data

AI问答

⁰