Deep-HyROMnet: a deep learning-based operator approximation for hyper-reduction of nonlinear parametrized PDEs
From MaRDI portal
Publication:2103427
DOI10.1007/S10915-022-02001-8zbMATH Open1503.65225arXiv2202.02658OpenAlexW4304615543MaRDI QIDQ2103427FDOQ2103427
Authors: Ludovica Cicci, Stefania Fresca, Andrea Manzoni 
Publication date: 13 December 2022
Published in: Journal of Scientific Computing (Search for Journal in Brave)
Abstract: To speed-up the solution to parametrized differential problems, reduced order models (ROMs) have been developed over the years, including projection-based ROMs such as the reduced-basis (RB) method, deep learning-based ROMs, as well as surrogate models obtained via a machine learning approach. Thanks to its physics-based structure, ensured by the use of a Galerkin projection of the full order model (FOM) onto a linear low-dimensional subspace, RB methods yield approximations that fulfill the physical problem at hand. However, to make the assembling of a ROM independent of the FOM dimension, intrusive and expensive hyper-reduction stages are usually required, such as the discrete empirical interpolation method (DEIM), thus making this strategy less feasible for problems characterized by (high-order polynomial or nonpolynomial) nonlinearities. To overcome this bottleneck, we propose a novel strategy for learning nonlinear ROM operators using deep neural networks (DNNs). The resulting hyper-reduced order model enhanced by deep neural networks, to which we refer to as Deep-HyROMnet, is then a physics-based model, still relying on the RB method approach, however employing a DNN architecture to approximate reduced residual vectors and Jacobian matrices once a Galerkin projection has been performed. Numerical results dealing with fast simulations in nonlinear structural mechanics show that Deep-HyROMnets are orders of magnitude faster than POD-Galerkin-DEIM ROMs, keeping the same level of accuracy.
Full work available at URL: https://arxiv.org/abs/2202.02658
Recommendations
- A comprehensive deep learning-based approach to reduced order modeling of nonlinear time-dependent parametrized PDEs
- Non-intrusive reduced order modeling of nonlinear problems using neural networks
- POD-DL-ROM: enhancing deep learning-based reduced order models for nonlinear parametrized PDEs by proper orthogonal decomposition
- Non-intrusive model reduction of large-scale, nonlinear dynamical systems using deep learning
- Nonintrusive Reduced-Order Models for Parametric Partial Differential Equations via Data-Driven Operator Inference
deep learningproper orthogonal decompositionreduced order modelingnonlinear time-dependent PDEshyper-reduction
Numerical optimization and variational techniques (65K10) Artificial neural networks and deep learning (68T07) Computational learning theory (68Q32) Numerical interpolation (65D05) Finite element, Rayleigh-Ritz and Galerkin methods for initial value and initial-boundary value problems involving PDEs (65M60) Numerical methods for partial differential equations, initial value and time-dependent initial-boundary value problems (65M99)
Cites Work
- Reduced basis methods for partial differential equations. An introduction
- Finding structure with randomness: probabilistic algorithms for constructing approximate matrix decompositions
- Deep learning
- An `empirical interpolation' method: Application to efficient reduced-basis discretization of partial differential equations
- Nonlinear model reduction via discrete empirical interpolation
- Title not available (Why is that?)
- Missing Point Estimation in Models Described by Proper Orthogonal Decomposition
- Numerical models for differential problems
- A survey of projection-based model reduction methods for parametric dynamical systems
- Efficient non linear model reduction via a least-squares Petrov-Galerkin projection and compressive tensor approximations
- Efficient reduced-basis treatment of nonaffine and nonlinear partial differential equations
- A Class of Methods for Solving Nonlinear Simultaneous Equations
- Reduced basis approximation for nonlinear parametrized evolution equations based on empirical operator interpolation
- Thermodynamically consistent orthotropic activation model capturing ventricular systolic wall thickening in cardiac electromechanics
- Certified reduced basis methods for parametrized partial differential equations
- Model reduction and neural networks for parametric PDEs
- Nonlinear model order reduction based on local reduced-order bases
- Active stress vs. active strain in mechanobiology: constitutive issues
- POD-DEIM model order reduction for strain-softening viscoplasticity
- POD-based model reduction with empirical interpolation applied to nonlinear elasticity
- Structure-preserving, stability, and accuracy properties of the energy-conserving sampling and weighting method for the hyper reduction of nonlinear finite element dynamic models
- Non-intrusive model reduction of large-scale, nonlinear dynamical systems using deep learning
- Model reduction of dynamical systems on nonlinear manifolds using deep convolutional autoencoders
- Data-driven operator inference for nonintrusive projection-based model reduction
- Data-driven reduced order modeling for time-dependent problems
- Non-intrusive reduced order modeling of nonlinear problems using neural networks
- Model reduction of parametrized systems. Selected contributions based on the presentations at the MoRePaS conference, SISSA, Trieste, Italy, October 13--16, 2015
- Reduced order modeling for nonlinear structural analysis using Gaussian process regression
- Efficient model reduction of parametrized systems by matrix discrete empirical interpolation
- Numerical approximation of parametrized problems in cardiac electrophysiology by a local reduced basis method
- A comprehensive deep learning-based approach to reduced order modeling of nonlinear time-dependent parametrized PDEs
- Localized Discrete Empirical Interpolation Method
- Projection-based model reduction: formulations for physics-based machine learning
- POD-DL-ROM: enhancing deep learning-based reduced order models for nonlinear parametrized PDEs by proper orthogonal decomposition
- A matrix DEIM technique for model reduction of nonlinear parametrized problems in cardiac mechanics
- Operator inference for non-intrusive model reduction of systems with non-polynomial nonlinear terms
- The deal.II library, version 9.2
- Computational bottlenecks for PROMs: precomputation and hyperreduction
- A deep learning approach to Reduced Order Modelling of parameter dependent partial differential equations
- Projection-based reduced order models for parameterized nonlinear time-dependent problems arising in cardiac mechanics
Cited In (13)
- A comprehensive deep learning-based approach to reduced order modeling of nonlinear time-dependent parametrized PDEs
- Deep learning methods for partial differential equations and related parameter identification problems
- POD-DL-ROM: enhancing deep learning-based reduced order models for nonlinear parametrized PDEs by proper orthogonal decomposition
- lifex-cfd: an open-source computational fluid dynamics solver for cardiovascular applications
- Model reduction of coupled systems based on non-intrusive approximations of the boundary response maps
- A nonlinear-manifold reduced-order model and operator learning for partial differential equations with sharp solution gradients
- Error estimates for POD-DL-ROMs: a deep learning framework for reduced order modeling of nonlinear parametrized PDEs enhanced by proper orthogonal decomposition
- Dosnet as a non-black-box PDE solver: when deep learning meets operator splitting
- Deep learning‐based reduced order models for the real‐time simulation of the nonlinear dynamics of microstructures
- HiDeNN-TD: reduced-order hierarchical deep learning neural networks
- Uncertainty quantification for nonlinear solid mechanics using reduced order models with Gaussian process regression
- Non-intrusive parametric hyper-reduction for nonlinear structural finite element formulations
- A graph convolutional autoencoder approach to model order reduction for parametrized PDEs
Uses Software
This page was built for publication: Deep-HyROMnet: a deep learning-based operator approximation for hyper-reduction of nonlinear parametrized PDEs
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2103427)
