Learning constitutive relations from indirect observations using deep neural networks
From MaRDI portal
Publication:781968
DOI10.1016/J.JCP.2020.109491zbMATH Open1437.65192arXiv1905.12530OpenAlexW3022973855MaRDI QIDQ781968FDOQ781968
Kailai Xu, Eric Darve, Daniel Z. Huang, C. Farhat
Publication date: 21 July 2020
Published in: Journal of Computational Physics (Search for Journal in Brave)
Abstract: We present a new approach for predictive modeling and its uncertainty quantification for mechanical systems, where coarse-grained models such as constitutive relations are derived directly from observation data. We explore the use of a neural network to represent the unknown constitutive relations, compare the neural networks with piecewise linear functions, radial basis functions, and radial basis function networks, and show that the neural network outperforms the others in certain cases. We analyze the approximation error of the neural networks using a scaling argument. The training and predicting processes in our framework combine the finite element method, automatic differentiation, and neural networks (or other function approximators). Our framework also allows uncertainty quantification in the form of confidence intervals. Numerical examples on a multiscale fiber-reinforced plate problem and a nonlinear rubbery membrane problem from solid mechanics demonstrate the effectiveness of our framework.
Full work available at URL: https://arxiv.org/abs/1905.12530
Recommendations
- Learning constitutive relations using symmetric positive definite neural networks
- Learning nonlocal constitutive models with neural networks
- Constitutive artificial neural networks: a fast and general approach to predictive data-driven constitutive modeling by deep learning
- Learning viscoelasticity models from indirect data using deep neural networks
- A deep learning framework for constitutive modeling based on temporal convolutional network
- Recurrent neural networks (RNNs) learn the constitutive law of viscoelasticity
- scientific article; zbMATH DE number 1213967
- Deep autoencoders for physics-constrained data-driven nonlinear materials modeling
- Constructive deep ReLU neural network approximation
- Reliable extrapolation of deep neural operators informed by physics or sparse observations
Artificial neural networks and deep learning (68T07) Laplace operator, Helmholtz equation (reduced wave equation), Poisson equation (35J05) Finite element, Rayleigh-Ritz and Galerkin methods for boundary value problems involving PDEs (65N30) Numerical differentiation (65D25)
Cites Work
- Bayesian calibration of computer models. (With discussion)
- A Limited Memory Algorithm for Bound Constrained Optimization
- Line search algorithms with guaranteed sufficient decrease
- Cross-covariance functions for multivariate geostatistics
- Determination of the size of the representative volume element for random composites: Statistical and numerical approach.
- Chebyshev and Fourier spectral methods.
- Analysis of Composite Materials—A Survey
- Barycentric Lagrange Interpolation
- Spectral Methods
- The Wiener--Askey Polynomial Chaos for Stochastic Differential Equations
- \(FE^2\) multiscale approach for modelling the elastoviscoplastic behaviour of long fibre SiC/Ti composite materials
- An algorithm for selecting a good value for the parameter \(c\) in radial basis function interpolation
- Title not available (Why is that?)
- Modeling uncertainty in flow simulations via generalized polynomial chaos.
- Title not available (Why is that?)
- Two algorithms for constructing a Delaunay triangulation
- Error estimates and condition numbers for radial basis function interpolation
- Laplacian smoothing and Delaunay triangulations
- Title not available (Why is that?)
- Toward realization of computational homogenization in practice
- Identification of elastic parameters by displacement field measurement
- Finite element computation of large rubber membrane deformations
- Advanced Lectures on Machine Learning
- Active Subspace Methods in Theory and Practice: Applications to Kriging Surfaces
- Title not available (Why is that?)
- Computational homogenization of nonlinear elastic materials using neural networks
- Deep learning-based numerical methods for high-dimensional parabolic partial differential equations and backward stochastic differential equations
- Data-driven computational mechanics
- Genetic evolution of nonlinear material constitutive models
- Machine learning strategies for systems with invariance properties
- Finite element analysis of V-ribbed belts using neural network based hyperelastic material model
- Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
- Reynolds averaged turbulence modelling using deep neural networks with embedded invariance
- Deep UQ: learning deep neural network surrogate models for high dimensional uncertainty quantification
- A framework for data-driven analysis of materials under uncertainty: countering the curse of dimensionality
- Data driven computing with noisy material data sets
- A multiscale multi-permeability poroplasticity model linked by recursive homogenizations and deep learning
Cited In (42)
- Learning constitutive models from microstructural simulations via a non-intrusive reduced basis method
- Efficient derivative-free Bayesian inference for large-scale inverse problems
- Constitutive artificial neural networks: a fast and general approach to predictive data-driven constitutive modeling by deep learning
- Hybrid FEM-NN models: combining artificial neural networks with the finite element method
- Learning nonlocal constitutive models with neural networks
- Uncertainty quantification in scientific machine learning: methods, metrics, and comparisons
- Learning viscoelasticity models from indirect data using deep neural networks
- On sparse regression, \(L_p\)-regularization, and automated model discovery
- A mechanics‐informed artificial neural network approach in data‐driven constitutive modeling
- DPK: Deep Neural Network Approximation of the First Piola-Kirchhoff Stress
- Improved Analysis of PINNs: Alleviate the CoD for Compositional Solutions
- A deep learning energy method for hyperelasticity and viscoelasticity
- Learning homogenization for elliptic operators
- Image inversion and uncertainty quantification for constitutive laws of pattern formation
- An equivariant neural operator for developing nonlocal tensorial constitutive models
- Learning generative neural networks with physics knowledge
- Deep learning in computational mechanics: a review
- Geometric deep learning for computational mechanics. I: Anisotropic hyperelasticity
- Sobolev training of thermodynamic-informed neural networks for interpretable elasto-plasticity models with level set hardening
- Solving inverse problems in stochastic models using deep neural networks and adversarial training
- Iterated Kalman methodology for inverse problems
- Hybrid analysis and modeling, eclecticism, and multifidelity computing toward digital twin revolution
- Bayesian calibration for large‐scale fluid structure interaction problems under embedded/immersed boundary framework
- Probabilistic partition of unity networks for high‐dimensional regression problems
- Molecular dynamics inferred transfer learning models for finite‐strain hyperelasticity of monoclinic crystals: Sobolev training and validations against physical constraints
- Multiscale modeling of inelastic materials with thermodynamics-based artificial neural networks (TANN)
- A mechanics-informed deep learning framework for data-driven nonlinear viscoelasticity
- Spectral operator learning for parametric PDEs without data reliance
- Frame-independent vector-cloud neural network for nonlocal constitutive modeling on arbitrary grids
- Physics constrained learning for data-driven inverse modeling from sparse observations
- Data-driven elasto-(visco)-plasticity involving hidden state variables
- Automated discovery of generalized standard material models with EUCLID
- Transfer learning based physics-informed neural networks for solving inverse problems in engineering structures under different loading scenarios
- An indirect training approach for implicit constitutive modelling using recurrent neural networks and the virtual fields method
- NN-mCRE: a modified constitutive relation error framework for unsupervised learning of nonlinear state laws with physics-augmented neural networks
- Physics-integrated neural differentiable (PiNDiff) model for composites manufacturing
- Bayesian-EUCLID: discovering hyperelastic material laws with uncertainties
- Unsupervised discovery of interpretable hyperelastic constitutive laws
- Material Modeling via Thermodynamics-Based Artificial Neural Networks
- A computationally tractable framework for nonlinear dynamic multiscale modeling of membrane woven fabrics
- Addressing discontinuous root-finding for subsequent differentiability in machine learning, inverse problems, and control
- Physical informed neural network for thermo-hydral analysis of fire-loaded concrete
Uses Software
This page was built for publication: Learning constitutive relations from indirect observations using deep neural networks
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q781968)
