Abstract:Multiscale modeling is an effective approach for investigating multiphysics systems with largely disparate size features, where models with different resolutions or heterogeneous descriptions are coupled together for predicting the system's response. The solver with lower fidelity (coarse) is responsible for simulating domains with homogeneous features, whereas the expensive high-fidelity (fine) model describes microscopic features with refined discretization, often making the overall cost prohibitively high, especially for time-dependent problems. In this work, we explore the idea of multiscale modeling with machine learning and employ DeepONet, a neural operator, as an efficient surrogate of the expensive solver. DeepONet is trained offline using data acquired from the fine solver for learning the underlying and possibly unknown fine-scale dynamics. It is then coupled with standard PDE solvers for predicting the multiscale systems with new boundary/initial conditions in the coupling stage. The proposed framework significantly reduces the computational cost of multiscale simulations since the DeepONet inference cost is negligible, facilitating readily the incorporation of a plurality of interface conditions and coupling schemes. We present various benchmarks to assess the accuracy and efficiency, including static and time-dependent problems. We also demonstrate the feasibility of coupling of a continuum model (finite element methods, FEM) with a neural operator, serving as a surrogate of a particle system (Smoothed Particle Hydrodynamics, SPH), for predicting mechanical responses of anisotropic and hyperelastic materials. What makes this approach unique is that a well-trained over-parametrized DeepONet can generalize well and make predictions at a negligible cost.

Neural Physical Simulation with Multi-Resolution Hash Grid Encoding

Multi-scale hash encoding based neural geometry representation

On Fast Simulation of Dynamical System with Neural Vector Enhanced Numerical Solver

NeuralStagger: Accelerating Physics-Constrained Neural PDE Solver with Spatial-Temporal Decomposition

A Multi-resolution Adaptive Encoding Method for 3-D Reconstruction of DPO-SDF

A GPU-based computational framework that bridges neuron simulation and artificial intelligence

A High Resolution Physics-informed Neural Networks for High-Dimensional Convection-Diffusion-reaction Equations.

Convolutional Hierarchical Deep Learning Neural Networks-Tensor Decomposition (C-HiDeNN-TD): a scalable surrogate modeling approach for large-scale physical systems

Differentiable Multi-Fidelity Fusion: Efficient Learning of Physics Simulations with Neural Architecture Search and Transfer Learning

Learning to Simulate Complex Physical Systems: A Case Study

A novel paradigm for solving PDEs: multi scale neural computing

UGrid: An Efficient-And-Rigorous Neural Multigrid Solver for Linear PDEs

Multi-resolution partial differential equations preserved learning framework for spatiotemporal dynamics

Efficient physics-informed neural networks using hash encoding

Heuristic Tree-Partition-Based Parallel Method for Biophysically Detailed Neuron Simulation

Interfacing finite elements with deep neural operators for fast multiscale modeling of mechanics problems

Towards General Neural Surrogate Solvers with Specialized Neural Accelerators

Multiscale Neural Operator: Learning Fast and Grid-independent PDE Solvers

Coupled Multiphysics Solver for Irregular Regions Based on Graph Neural Network

A fully differentiable GNN-based PDE Solver: With Applications to Poisson and Navier-Stokes Equations