PhyGeoNet: Physics-informed geometry-adaptive convolutional neural networks for solving parameterized steady-state PDEs on irregular domain

•Enable CNN-based physics-informed deep learning for PDEs on irregular domain.•The proposed network can be trained without any labeled data.•Boundary conditions are strictly encoded in a hard manner.•Investigated complex parametric PDEs, e.g., Naiver-Stokes with varying geometries.•Shows improvement...

Full description

Saved in:
Bibliographic Details
Published in:Journal of computational physics Vol. 428; p. 110079
Main Authors: Gao, Han, Sun, Luning, Wang, Jian-Xun
Format: Journal Article
Language:English
Published: Cambridge Elsevier Inc 01.03.2021
Elsevier Science Ltd
Subjects:
Artificial neural networks
Boundary conditions
Computational fluid dynamics
Computational physics
Coordinate transformations
Domains
Label-free
Machine learning
Navier-Stokes
Neural networks
Parameterization
Partial differential equations
Physics
Physics-constrained deep learning
Physics-informed neural networks
Poisson equation
Steady state
Surrogate modeling
Thermodynamics
Physics-constrained deep learning
Navier-Stokes
Partial differential equations
Label-free
Surrogate modeling
Physics-informed neural networks
ISSN:0021-9991, 1090-2716
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•Enable CNN-based physics-informed deep learning for PDEs on irregular domain.•The proposed network can be trained without any labeled data.•Boundary conditions are strictly encoded in a hard manner.•Investigated complex parametric PDEs, e.g., Naiver-Stokes with varying geometries.•Shows improvements of efficiency and accuracy over FC-NN formulations. Recently, the advent of deep learning has spurred interest in the development of physics-informed neural networks (PINN) for efficiently solving partial differential equations (PDEs), particularly in a parametric setting. Among all different classes of deep neural networks, the convolutional neural network (CNN) has attracted increasing attention in the scientific machine learning community, since the parameter-sharing feature in CNN enables efficient learning for problems with large-scale spatiotemporal fields. However, one of the biggest challenges is that CNN only can handle regular geometries with image-like format (i.e., rectangular domains with uniform grids). In this paper, we propose a novel physics-constrained CNN learning architecture, aiming to learn solutions of parametric PDEs on irregular domains without any labeled data. In order to leverage powerful classic CNN backbones, elliptic coordinate mapping is introduced to enable coordinate transforms between the irregular physical domain and regular reference domain. The proposed method has been assessed by solving a number of steady-state PDEs on irregular domains, including heat equations, Navier-Stokes equations, and Poisson equations with parameterized boundary conditions, varying geometries, and spatially-varying source fields. Moreover, the proposed method has also been compared against the state-of-the-art PINN with fully-connected neural network (FC-NN) formulation. The numerical results demonstrate the effectiveness of the proposed approach and exhibit notable superiority over the FC-NN based PINN in terms of efficiency and accuracy.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0021-9991
1090-2716
DOI:10.1016/j.jcp.2020.110079