Accelerating Particle and Fluid Simulations with Differentiable Graph Networks for Solving Forward and Inverse Problems

Krishna Kumar,Yongjin Choi

2023-09-23

Abstract:We leverage physics-embedded differentiable graph network simulators (GNS) to accelerate particulate and fluid simulations to solve forward and inverse problems. GNS represents the domain as a graph with particles as nodes and learned interactions as edges. Compared to modeling global dynamics, GNS enables learning local interaction laws through edge messages, improving its generalization to new environments. GNS achieves over 165x speedup for granular flow prediction compared to parallel CPU numerical simulations. We propose a novel hybrid GNS/Material Point Method (MPM) to accelerate forward simulations by minimizing error on a pure surrogate model by interleaving MPM in GNS rollouts to satisfy conservation laws and minimize errors achieving 24x speedup compared to pure numerical simulations. The differentiable GNS enables solving inverse problems through automatic differentiation, identifying material parameters that result in target runout distances. We demonstrate the ability of GNS to solve inverse problems by iteratively updating the friction angle (a material property) by computing the gradient of a loss function based on the final and target runouts, thereby identifying the friction angle that best matches the observed runout. The physics-embedded and differentiable simulators open an exciting new paradigm for AI-accelerated design, control, and optimization.

Geophysics,Machine Learning,Computational Physics

What problem does this paper attempt to address?

The paper aims to address forward and inverse problems in particle and fluid simulations and accelerate these simulation processes. Specifically: 1. **Accelerating Simulations**: By using a physics-embedded differentiable graph network simulator (GNS) to accelerate particle and fluid simulations, GNS achieves over 165 times speedup in particle flow prediction compared to traditional CPU-based numerical simulation methods. 2. **Forward Simulation Acceleration**: A new hybrid GNS/Material Point Method (MPM) is proposed, which intersperses MPM within GNS to satisfy conservation laws and minimize errors, achieving forward simulation speeds 24 times faster than pure numerical simulations. 3. **Inverse Problem Solving**: Utilizing the automatic differentiation capability of GNS to solve inverse problems, such as identifying material parameters (e.g., friction angle) that lead to a target running distance through gradient optimization. This method overcomes the issue of traditional simulators being unable to compute gradients in inverse mode. In summary, the main goal of this paper is to significantly improve the speed of particle and fluid simulations by introducing a physics-embedded differentiable graph network simulator and effectively solve inverse problems, thereby opening up a new paradigm for AI-accelerated design, control, and optimization.

Accelerating Particle and Fluid Simulations with Differentiable Graph Networks for Solving Forward and Inverse Problems

Inverse analysis of granular flows using differentiable graph neural network simulator

GNS: A generalizable Graph Neural Network-based simulator for particulate and fluid modeling

Integrating Graph Neural Network-Based Surrogate Modeling with Inverse Design for Granular Flows

Three-dimensional granular flow simulation using graph neural network-based learned simulator

Learning Incompressible Fluid Dynamics from Scratch -- Towards Fast, Differentiable Fluid Models that Generalize

Graph Neural Network-based surrogate model for granular flows

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

Physics-constrained neural networks as multi-material Riemann solvers for compressible two-gas simulations

Solving Inverse Problems in Steady State Navier-Stokes Equations using Deep Neural Networks

Physics-informed MeshGraphNets (PI-MGNs): Neural finite element solvers for non-stationary and nonlinear simulations on arbitrary meshes

Neural SPH: Improved Neural Modeling of Lagrangian Fluid Dynamics

Accelerating high order discontinuous Galerkin solvers using neural networks: 3D compressible Navier-Stokes equations

Accurately Solving Physical Systems with Graph Learning

Accelerating Simulation of Two-Phase Flows with Neural PDE Surrogates

Multi-scale Graph Neural Network for Physics-Informed Fluid Simulation

MPMNet: A Data-Driven MPM Framework for Dynamic Fluid-Solid Interaction

Graph Neural Networks (gnns) Based Accelerated Numerical Simulation

Rethinking materials simulations: Blending direct numerical simulations with neural operators

A Gradient-based Deep Neural Network Model for Simulating Multiphase Flow in Porous Media