Neural network-based limiter with transfer learning
From MaRDI portal
Complexity and performance of numerical algorithms (65Y20) Artificial neural networks and deep learning (68T07) Packaged methods for numerical algorithms (65Y15) Multistep, Runge-Kutta and extrapolation methods for ordinary differential equations (65L06) Finite element, Rayleigh-Ritz and Galerkin methods for initial value and initial-boundary value problems involving PDEs (65M60) PDEs on graphs and networks (ramified or polygonal spaces) (35R02) Numerical methods for partial differential equations, initial value and time-dependent initial-boundary value problems (65M99)
Abstract: A neural network is trained using simulation data from a Runge Kutta discontinuous Galerkin (RKDG) method and a modal high order limiter. With this methodology, we design one and two-dimensional black-box shock detection functions. Furthermore, we describe a strategy to adapt the shock detection function to different numerical schemes without the need of a full training cycle and large dataset. We evaluate the performance of the neural network on a RKDG scheme for validation. To evaluate the domain adaptation properties of this neural network limiter, our methodology is verified on a residual distribution scheme (RDS), both in one and two-dimensional problems, and on Cartesian and unstructured meshes. Lastly, we report on the quality of the numerical solutions when using a neural based shock detection method, in comparison to more traditional limiters, as well as on the computational impact of using this method in existing codes.
Recommendations
- An artificial neural network as a troubled-cell indicator
- A neural network based shock detection and localization approach for discontinuous Galerkin methods
- Applications of limiters, neural networks and polynomial annihilation in higher-order FD/FV schemes
- Using deep neural networks for detecting spurious oscillations in discontinuous Galerkin solutions of convection-dominated convection-diffusion equations
- A priori neural networks versus a posteriori MOOD loop: a high accurate 1D FV scheme testing bed
Cites work
- A neural network based shock detection and localization approach for discontinuous Galerkin methods
- An artificial neural network as a troubled-cell indicator
- Application of conservative residual distribution schemes to the solution of the shallow water equations on unstructured meshes
- Continuous interior penalty finite element method for the time-dependent Navier-Stokes equations: space discretization and convergence
- Deep learning
- Deep, Narrow Sigmoid Belief Networks Are Universal Approximators
- Explicit Runge-Kutta residual distribution schemes for time dependent problems: second order case
- High-order residual distribution scheme for the time-dependent Euler equations of fluid dynamics
- Limiters for high-order discontinuous Galerkin methods
- Multi-dimensional Optimal Order Detection (MOOD) — a Very High-Order Finite Volume Scheme for Conservation Laws on Unstructured Meshes
- On a Class of High Resolution Total-Variation-Stable Finite-Difference Schemes
- On high order strong stability preserving Runge-Kutta and multi step time discretizations
- Optimal approximation of piecewise smooth functions using deep ReLU neural networks
- Optimization methods for large-scale machine learning
- Parallel, adaptive finite element methods for conservation laws
- Practical mathematical optimization. An introduction to basic optimization theory and classical and new gradient-based algorithms.
- Residual distribution schemes: current status and future trends
- Solution of two-dimensional Riemann problems for gas dynamics without Riemann problem solvers
- TVB Runge-Kutta Local Projection Discontinuous Galerkin Finite Element Method for Conservation Laws II: General Framework
- The Runge-Kutta discontinuous Galerkin method for conservation laws. I: Multidimensional systems
- Weak solutions of nonlinear hyperbolic equations and their numerical computation
- \textit{A posteriori} correction of high-order discontinuous Galerkin scheme through subcell finite volume formulation and flux reconstruction
- ``\textit{A posteriori} limited high order and robust schemes for transient simulations of fluid flows in gas dynamics
Cited in
(7)- Machine learning-based WENO5 scheme
- Third-order hybrid finite volume WENO method with a multilayer perceptron troubled-cell indicator for hyperbolic conservation laws
- Applications of limiters, neural networks and polynomial annihilation in higher-order FD/FV schemes
- Using deep neural networks for detecting spurious oscillations in discontinuous Galerkin solutions of convection-dominated convection-diffusion equations
- Learning WENO for entropy stable schemes to solve conservation laws
- Preface to the focused issue on high-order numerical methods for evolutionary PDEs
- Convolution Neural Network Shock Detector for Numerical Solution of Conservation Laws
This page was built for publication: Neural network-based limiter with transfer learning
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6098315)
