Mesh-based simulations are central to modeling complex physical systems in many disciplines across science and engineering. Mesh representations support powerful numerical integration methods and their resolution can be adapted to strike favorable trade-offs between accuracy and efficiency. However, high-dimensional scientific simulations are very expensive to run, and solvers and parameters must often be tuned individually to each system studied. Here we introduce MeshGraphNets, a framework for learning mesh-based simulations using graph neural networks. Our model can be trained to pass messages on a mesh graph and to adapt the mesh discretization during forward simulation. Our results show it can accurately predict the dynamics of a wide range of physical systems, including aerodynamics, structural mechanics, and cloth. The model's adaptivity supports learning resolution-independent dynamics and can scale to more complex state spaces at test time. Our method is also highly efficient, running 1-2 orders of magnitude faster than the simulation on which it is trained. Our approach broadens the range of problems on which neural network simulators can operate and promises to improve the efficiency of complex, scientific modeling tasks.
Mesh-based Super-Resolution of Fluid Flows with Multiscale Graph Neural Networks
Transportation and Power Systems Division, Argonne National Laboratory(阿贡国家实验室运输与电力系统部门);Computational Science Division, Argonne National Laboratory(阿贡国家实验室计算科学部门);Argonne Leadership Computing Facility, Argonne National Laboratory(阿贡领导计算设施);Mathematics and Computer Science Division, Argonne National Laboratory(阿贡国家实验室数学与计算机科学部门);College of Information Sciences and Technology, Pennsylvania State University(宾夕法尼亚州立大学信息科学与技术学院)
G-PARC: Graph-Physics Aware Recurrent Convolutional Neural Networks for Spatiotemporal Dynamics on Unstructured Meshes
School of Data Science, University of Virginia(数据科学学院,弗吉尼亚大学);University of Virginia(弗吉尼亚大学);Department of Engineering Systems and Environment, University of Virginia(工程系统与环境系,弗吉尼亚大学);Department of Mechanical and Aerospace Engineering, University of Virginia(机械与航空航天工程系,弗吉尼亚大学);University of Iowa(爱荷华大学)
Super-resolution of turbulent reacting flows on complex meshes using graph neural networks
FLAME Laboratory, Department of Computational and Data Sciences, Indian Institute of Science, Bengaluru, 560012, India(印度科学研究院流体动力学与能源实验室,班加罗尔);CAPS Laboratory, Department of Mechanical and Process Engineering, ETH Zürich, 8092 Zürich, Switzerland(苏黎世联邦理工学院机械与过程工程实验室,苏黎世);Jülich Supercomputing Centre, Forschungszentrum Jülich GmbH, Jülich, 52425, Germany(尤利希超算中心,尤利希)
G-Adaptivity: optimised graph-based mesh relocation for finite element methods
Department of Applied Mathematics;Theoretical Physics, University of Cambridge, UK;Department of Mathematical Sciences, University of Bath, UK;Department of Computer Science, University of Bath, UK;School of Computing;Mathematical Sciences, University of Leicester, UK;Laboratoire Jacques-Louis Lions, Sorbonne Universit\'e, France;Department of Computer Science;Technology, University of Cambridge, UK
DeepCrysTet: A Deep Learning Approach Using Tetrahedral Mesh for Predicting Properties of Crystalline Materials
SAKURA internet Inc.(樱花互联网公司);National Institute for Materials Science (NIMS)(国立材料科学研究所);The University of Tokyo(东京大学);RIKEN(理化学研究所);Kyoto University(京都大学)
A Survey of AI Methods for Geometry Preparation and Mesh Generation in Engineering Simulation
Sandia National Laboratories(桑迪亚国家实验室);Old Dominion University(旧 Dominion 大学);Los Alamos National Laboratory(洛斯阿拉莫斯国家实验室);New York University / Stony Brook University(纽约大学 / 斯通布鲁克大学);CEA(法国原子能委员会);Dalian University of Technology(大连理工大学);Carnegie Mellon University(卡内基梅隆大学)
MMGP: a Mesh Morphing Gaussian Process-based machine learning method for regression of physical problems under non-parameterized geometrical variability
An extensible Benchmarking Graph-Mesh dataset for studying Steady-State Incompressible Navier-Stokes Equations
Sorbonne Université(索邦大学);CNRS(法国国家科学研究中心);École Normale Supérieure Paris Saclay(巴黎萨克雷高等师范学院);Extrality(Extrality公司);Criteo AI Lab(Criteo人工智能实验室)
\emph{Geometric Deep Learning}(GDL)的最新进展显示了其提供强大的数据驱动模型的潜力。这为探索从Graph-Mesh数据学习由\emph{Partial Differential Equations}(PDE)管理的物理系统的新方法提供了动力。然而,尽管做出了努力并取得了最新成就,但仍有几个研究方向尚未探索,进展仍然远远不能满足现实世界现象的物理要求。主要障碍之一是缺乏基准数据集和共同的物理评估协议。在本文中,我们提出了一个2-D图形网格数据集,以研究在高雷诺数制度的翼型气流(从$10^6$和超越)。我们还介绍了在翼型上的应力的度量,以评估GDL模型的重要物理量。此外,我们提供广泛的GDL基线。
MeshGraphNet-Transformer: Scalable Mesh-based Learned Simulation for Solid Mechanics
Keysight-UZ Chair of the National Strategy on Artificial Intelligence(Keysight-UZ人工智能国家战略主席);Aragon Institute of Engineering Research (I3A)(阿兰格工程研究所);Universidad de Zaragoza(萨拉戈萨大学)
Mask-Morph Graph U-Net: A Generalisable Mesh-Based Surrogate for Crashworthiness Field Prediction under Large Geometric Variation
Dyson School of Design Engineering, Imperial College London(帝国理工学院伦敦设计工程学院);TUM School of Engineering and Design, Technical University of Munich(慕尼黑技术大学工程与设计学院);Faculty of Mechanical Engineering, OTH Regensburg(雷根斯堡机械工程学院);NVIDIA(NVIDIA公司)
Meshless Shape Optimization using Neural Networks and Partial Differential Equations on Graphs
Institute of Mathematics, University of Würzburg(维尔茨堡大学数学研究所);Center for Artificial Intelligence and Data Science (CAIDAS), University of Würzburg(维尔茨堡大学人工智能与数据科学中心(CAIDAS))
Mesh Variational Autoencoders with Edge Contraction Pooling
Institute of Computing Technology, CAS(中国科学院计算技术研究所);University of Chinese Academy of Sciences(中国科学院大学);Cardiff University(卡迪夫大学);City University of Hong Kong(香港城市大学)
MeshDiffusion: Score-based Generative 3D Mesh Modeling
Mila(米拉人工智能研究所);Université de Montréal(蒙特利尔大学);Max Planck Institute for Intelligent Systems(马克斯·普朗克智能系统研究所);ETH Zürich(苏黎世联邦理工学院);McGill University(麦吉尔大学);University of Cambridge(剑桥大学)