☆ 4.6 Article

On the use of graph neural networks and shape-function-based gradient computation in the deep energy method

INTERNATIONAL JOURNAL FOR NUMERICAL METHODS IN ENGINEERING (2023)

相关参考文献

注意：仅列出部分参考文献，下载原文获取全部文献信息。

Article Engineering, Multidisciplinary

Physics-informed graph neural Galerkin networks: A unified framework for solving PDE-governed forward and inverse problems

Han Gao et al.

Summary: Despite the promise of PINNs in solving forward and inverse problems, there are technical challenges that hinder complex and realistic applications. To address these challenges, a discrete PINN framework based on GCN and PDE is proposed to solve PDEs in both forward and inverse settings, offering improved scalability, boundary enforcement, and handling of irregular geometries.

COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING (2022)

添加到收藏夹

Article Computer Science, Interdisciplinary Applications

The mixed Deep Energy Method for resolving concentration features in finite strain hyperelasticity

Jan N. Fuhg et al.

Summary: This paper introduces the research status of deep neural networks as universal approximators of PDEs and compares the advantages and disadvantages of two different methods. An improved deep energy method is proposed to solve the fine features in solid mechanics and a numerical integration scheme is introduced to enhance the flexibility of the method.

JOURNAL OF COMPUTATIONAL PHYSICS (2022)

添加到收藏夹

Article Mechanics

Flow completion network: Inferring the fluid dynamics from incomplete flow information using graph neural networks

Xiaodong He et al.

Summary: This paper introduces a novel neural network, FCN, which can infer the fluid dynamics of a flow field and the force acting on a body from incomplete data. The FCN is capable of handling both structured and unstructured data and utilizes flow field information and gradient information simultaneously. Evaluation results show that the proposed FCN outperforms traditional CNN and DNN models in terms of predicting force coefficients.

PHYSICS OF FLUIDS (2022)

添加到收藏夹

Article Engineering, Multidisciplinary

A mixed formulation for physics-informed neural networks as a potential solver for engineering problems in heterogeneous domains: Comparison with finite element method

Shahed Rezaei et al.

Summary: Physics Informed Neural Networks (PINNs) can find solutions to boundary value problems by minimizing a loss function that incorporates governing equations, initial and boundary conditions. This study proposes an improved method that uses the spatial gradient of the primary variable as an output and applies the strong form of the equation as a physical constraint. By comparing with finite element methods, it is shown that this approach has advantages, and the potential of combining PINN with physical FE simulations for designing composite materials is discussed.

COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING (2022)

添加到收藏夹

Article Mechanics

A deep learning energy method for hyperelasticity and viscoelasticity

Diab W. Abueidda et al.

Summary: In this study, the potential energy formulation and deep learning are merged to introduce the deep energy method, which shows potential for solving deformation problems in hyperelastic and viscoelastic materials.

EUROPEAN JOURNAL OF MECHANICS A-SOLIDS (2022)

添加到收藏夹

Article Mechanics

Deep autoencoder based energy method for the bending, vibration, and buckling analysis of Kirchhoff plates with transfer learning

Xiaoying Zhuang et al.

Summary: This paper introduces a Deep Autoencoder based Energy Method (DAEM) for the bending, vibration, and buckling analysis of Kirchhoff plates. The DAEM utilizes higher-order continuity, integrates a deep autoencoder and the minimum total potential principle, and serves as an unsupervised feature learning method. It efficiently identifies patterns, minimizes total potential energy, extracts fundamental frequencies and critical buckling loads, alleviates gradient problems, and improves computational efficiency and generality through transfer learning.

EUROPEAN JOURNAL OF MECHANICS A-SOLIDS (2021)

添加到收藏夹

Article Engineering, Multidisciplinary

Parametric deep energy approach for elasticity accounting for strain gradient effects

Vien Minh Nguyen-Thanh et al.

Summary: The Parametric Deep Energy Method (P-DEM) presented in this work solves elasticity problems considering strain gradient effects using physics-informed neural networks (PINNs). It does not require classical discretization and simplifies implementation by defining potential energy. Normalized inputs in a parametric/reference space prevent vanishing gradients and enable faster convergence. The method utilizes NURBS basis functions for forward-backward mapping and Gauss quadrature for approximating total potential energy in the loss function.

COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING (2021)

添加到收藏夹

Article Mechanics

Graph neural networks for laminar flow prediction around random two-dimensional shapes

J. Chen et al.

Summary: The research introduces a GCNN structure as a surrogate model for predicting laminar flows around 2D obstacles, it can be directly applied on triangular meshes and easily coupled with CFD solvers. The model is trained on a dataset containing CFD-computed laminar flows around 2000 random 2D shapes, and accuracy levels are compared with standard U-net architectures.

PHYSICS OF FLUIDS (2021)

添加到收藏夹

Article Mechanics

Teaching the incompressible Navier-Stokes equations to fast neural surrogate models in three dimensions

Nils Wandel et al.

Summary: This work introduces significant extensions to a deep learning framework for fluid simulations, enabling fast and accurate 3D simulations while conditioning the neural fluid model on additional information about viscosity and density. The method allows for simulating laminar and turbulent flows without prior fluid simulation data, showing improvements in accuracy, speed, and generalization capabilities compared to current 3D NN-based fluid models.

PHYSICS OF FLUIDS (2021)

添加到收藏夹

Article Engineering, Multidisciplinary

A physics-informed deep learning framework for inversion and surrogate modeling in solid mechanics

Ehsan Haghighat et al.

Summary: This study presents the application of Physics Informed Neural Networks (PINN) in solid mechanics, improving accuracy and convergence with a multi-network model and Isogeometric Analysis. The study demonstrates the importance of honoring physics in improving robustness and highlights the potential application of PINN in sensitivity analysis and surrogate modeling.

COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING (2021)

添加到收藏夹

Article Engineering, Multidisciplinary

Meshless physics-informed deep learning method for three-dimensional solid mechanics

Diab W. Abueidda et al.

Summary: Deep learning and the collocation method are merged to solve partial differential equations describing structures' deformation, offering a meshfree approach that avoids spatial discretization and data generation bottlenecks.

INTERNATIONAL JOURNAL FOR NUMERICAL METHODS IN ENGINEERING (2021)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Graph convolutional networks applied to unstructured flow field data

Francis Ogoke et al.

Summary: The method utilizes a graph convolutional neural network to infer fields on unstructured meshes, demonstrating the accurate prediction of global properties from spatially irregular measurements, such as predicting drag force around airfoils from scattered velocity measurements.

MACHINE LEARNING-SCIENCE AND TECHNOLOGY (2021)

添加到收藏夹

Article Mechanics

A deep energy method for finite deformation hyperelasticity

Vien Minh Nguyen-Thanh et al.

EUROPEAN JOURNAL OF MECHANICS A-SOLIDS (2020)

添加到收藏夹

Article Computer Science, Interdisciplinary Applications

Modeling the dynamics of PDE systems with physics-constrained deep auto-regressive networks

Nicholas Geneva et al.

JOURNAL OF COMPUTATIONAL PHYSICS (2020)

添加到收藏夹

Article Engineering, Multidisciplinary

FEA-Net: A physics-guided data-driven model for efficient mechanical response prediction

Houpu Yao et al.

COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING (2020)

添加到收藏夹

Article Engineering, Multidisciplinary

An energy approach to the solution of partial differential equations in computational mechanics via machine learning: Concepts, implementation and applications

E. Samaniego et al.

COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING (2020)

添加到收藏夹

Article Engineering, Civil

Physics-guided convolutional neural network (PhyCNN) for data-driven seismic response modeling

Ruiyang Zhang et al.

ENGINEERING STRUCTURES (2020)

添加到收藏夹

Article Engineering, Multidisciplinary

Geometric deep learning for computational mechanics Part I: anisotropic hyperelasticity

Nikolaos N. Vlassis et al.

COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING (2020)

添加到收藏夹

Article Materials Science, Multidisciplinary

Learning to fail: Predicting fracture evolution in brittle material models using recurrent graph convolutional neural networks

Max Schwarzer et al.

COMPUTATIONAL MATERIALS SCIENCE (2019)

添加到收藏夹

Article Computer Science, Interdisciplinary Applications

Physics-constrained deep learning for high-dimensional surrogate modeling and uncertainty quantification without labeled data

Yinhao Zhu et al.

JOURNAL OF COMPUTATIONAL PHYSICS (2019)

添加到收藏夹

Article Mathematics

The Deep Ritz Method: A Deep Learning-Based Numerical Algorithm for Solving Variational Problems

E. Weinan et al.

COMMUNICATIONS IN MATHEMATICS AND STATISTICS (2018)

添加到收藏夹

© Peeref 2019-2024. All rights reserved.