Discretizationnet: a machine-learning based solver for Navier-Stokes equations using finite volume discretization
From MaRDI portal
Publication:2021855
Abstract: Over the last few decades, existing Partial Differential Equation (PDE) solvers have demonstrated a tremendous success in solving complex, non-linear PDEs. Although accurate, these PDE solvers are computationally costly. With the advances in Machine Learning (ML) technologies, there has been a significant increase in the research of using ML to solve PDEs. The goal of this work is to develop an ML-based PDE solver, that couples important characteristics of existing PDE solvers with ML technologies. The two solver characteristics that have been adopted in this work are: 1) the use of discretization-based schemes to approximate spatio-temporal partial derivatives and 2) the use of iterative algorithms to solve linearized PDEs in their discrete form. In the presence of highly non-linear, coupled PDE solutions, these strategies can be very important in achieving good accuracy, better stability and faster convergence. Our ML-solver, DiscretizationNet, employs a generative CNN-based encoder-decoder model with PDE variables as both input and output features. During training, the discretization schemes are implemented inside the computational graph to enable faster GPU computation of PDE residuals, which are used to update network weights that result into converged solutions. A novel iterative capability is implemented during the network training to improve the stability and convergence of the ML-solver. The ML-Solver is demonstrated to solve the steady, incompressible Navier-Stokes equations in 3-D for several cases such as, lid-driven cavity, flow past a cylinder and conjugate heat transfer.
Recommendations
- A neural network multigrid solver for the Navier-Stokes equations
- Deep learning of the spanwise-averaged Navier-Stokes equations
- The deep learning Galerkin method for the general Stokes equations
- Deep learning solver for solving advection-diffusion equation in comparison to finite difference methods
- Accelerating high order discontinuous Galerkin solvers using neural networks: 3D compressible Navier-Stokes equations
- NSFnets (Navier-Stokes flow nets): physics-informed neural networks for the incompressible Navier-Stokes equations
- Physics-Driven Learning of the Steady Navier-Stokes Equations using Deep Convolutional Neural Networks
- Active training of physics-informed neural networks to aggregate and interpolate parametric solutions to the Navier-Stokes equations
- Machine learning design of volume of fluid schemes for compressible flows
- A finite element based deep learning solver for parametric PDEs
Cites work
- A Cartesian cut cell method for incompressible viscous flow
- A composite neural network that learns from multi-fidelity data: application to function approximation and inverse PDE problems
- A sharp-interface immersed boundary method with improved mass conservation and reduced spurious pressure oscillations
- Conservative physics-informed neural networks on discrete domains for conservation laws: applications to forward and inverse problems
- Data-driven discovery of PDEs in complex datasets
- Deep learning of vortex-induced vibrations
- Fronts propagating with curvature-dependent speed: Algorithms based on Hamilton-Jacobi formulations
- Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations
- Hidden physics models: machine learning of nonlinear partial differential equations
- Machine learning in cardiovascular flows modeling: predicting arterial blood pressure from non-invasive 4D flow MRI data using physics-informed neural networks
- Neural algorithm for solving differential equations
- Numerical study of the turbulent flow past an airfoil with trailing edge separation
- Physics-constrained deep learning for high-dimensional surrogate modeling and uncertainty quantification without labeled data
- Physics-informed neural networks for high-speed flows
- Physics-informed semantic inpainting: application to geostatistical modeling
- PPINN: parareal physics-informed neural network for time-dependent PDEs
- Surrogate modeling for fluid flows based on physics-constrained deep learning without simulation data
- Weak adversarial networks for high-dimensional partial differential equations
Cited in
(41)- PhyCRNet: physics-informed convolutional-recurrent network for solving spatiotemporal PDEs
- Physics-informed graph neural Galerkin networks: a unified framework for solving PDE-governed forward and inverse problems
- CAN-PINN: a fast physics-informed neural network based on coupled-automatic-numerical differentiation method
- Learning finite element convergence with the multi-fidelity graph neural network
- An exploratory study on machine learning to couple numerical solutions of partial differential equations
- Structure preservation for the deep neural network multigrid solver
- A physics-informed convolutional neural network for the simulation and prediction of two-phase Darcy flows in heterogeneous porous media
- QBoost for regression problems: solving partial differential equations
- VPVnet: a velocity-pressure-vorticity neural network method for the Stokes' equations under reduced regularity
- MOD-Net: a machine learning approach via model-operator-data network for solving PDEs
- Learning to Discretize: Solving 1D Scalar Conservation Laws via Deep Reinforcement Learning
- DiscretizationNet
- Machine Learning Surrogate Modeling for Meshless Methods: Leveraging Universal Approximation
- A nonlocal energy‐informed neural network for isotropic elastic solids with cracks under thermomechanical loads
- JAX-fluids: a fully-differentiable high-order computational fluid dynamics solver for compressible two-phase flows
- Accelerating explicit time-stepping with spatially variable time steps through machine learning
- Solving seepage equation using physics-informed residual network without labeled data
- A conservative hybrid deep learning method for Maxwell-Ampère-Nernst-Planck equations
- Physical informed memory networks for solving PDEs: implementation and applications
- Efficient simulation of two-dimensional time-fractional Navier-Stokes equations using RBF-FD approach
- Solving the discretised multiphase flow equations with interface capturing on structured grids using machine learning libraries
- Neural-integrated meshfree (NIM) method: a differentiable programming-based hybrid solver for computational mechanics
- Enhancing physics informed neural networks for solving Navier-Stokes equations
- Machine learning predictive model for dynamic response of rising bubbles impacting on a horizontal wall
- Cubic and quartic hyperbolic B-splines comparison for coupled Navier Stokes equation via differential quadrature method -- a statistical aspect
- Advanced physics-informed neural networks for numerical approximation of the coupled Schrödinger-KdV equation
- Transfer learning enhanced nonlocal energy-informed neural network for quasi-static fracture in rock-like materials
- Deep learning in computational mechanics: a review
- f-PICNN: a physics-informed convolutional neural network for partial differential equations with space-time domain
- A multifidelity machine learning based semi-Lagrangian finite volume scheme for linear transport equations and the nonlinear Vlasov-Poisson system
- Differentiability in unrolled training of neural physics simulators on transient dynamics
- MBNO: Mamba-based neural operators for solving partial differential equations
- Accelerating space-time methods using physics-informed neural networks
- Physics informed neural network framework for unsteady discretized reduced order system
- MRF-PINN: a multi-receptive-field convolutional physics-informed neural network for solving partial differential equations
- Preconditioned FEM-based neural networks for solving incompressible fluid flows and related inverse problems
- Learning to solve PDEs with finite volume-informed neural networks in a data-free approach
- SK-PINN: accelerated physics-informed deep learning by smoothing kernel gradients
- SFVnet: finite-volume informed U-Net for compressible flow prediction with sparse data under ill-conditions
- A Systematic Analysis of Automatic Differentiation versus Discretization-based Constraints for Physics-Informed PDE Solvers
- Overcoming the loss conditioning bottleneck in optimization-based PDE solvers: a well-conditioned loss function
Describes a project that uses
Uses Software
This page was built for publication: Discretizationnet: a machine-learning based solver for Navier-Stokes equations using finite volume discretization
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2021855)
