FE^ANN: an efficient data-driven multiscale approach based on physics-constrained neural networks and automated data mining
From MaRDI portal
Publication:6101611
Abstract: Herein, we present a new data-driven multiscale framework called FE which is based on two main keystones: the usage of physics-constrained artificial neural networks (ANNs) as macroscopic surrogate models and an autonomous data mining process. Our approach allows the efficient simulation of materials with complex underlying microstructures which reveal an overall anisotropic and nonlinear behavior on the macroscale. Thereby, we restrict ourselves to finite strain hyperelasticity problems for now. By using a set of problem specific invariants as the input of the ANN and the Helmholtz free energy density as the output, several physical principles, e.g., objectivity, material symmetry, compatibility with the balance of angular momentum and thermodynamic consistency are fulfilled a priori. The necessary data for the training of the ANN-based surrogate model, i.e., macroscopic deformations and corresponding stresses, are collected via computational homogenization of representative volume elements (RVEs). Thereby, the core feature of the approach is given by a completely autonomous mining of the required data set within an overall loop. In each iteration of the loop, new data are generated by gathering the macroscopic deformation states from the macroscopic finite element (FE) simulation and a subsequently sorting by using the anisotropy class of the considered material. Finally, all unknown deformations are prescribed in the RVE simulation to get the corresponding stresses and thus to extend the data set. The proposed framework consequently allows to reduce the number of time-consuming microscale simulations to a minimum. It is exemplarily applied to several descriptive examples, where a fiber reinforced composite with a highly nonlinear Ogden-type behavior of the individual components is considered.
Recommendations
- A multiscale, data-driven approach to identifying thermo-mechanically coupled laws -- bottom-up with artificial neural networks
- Local approximate Gaussian process regression for data-driven constitutive models: development and comparison with neural networks
- Multiscale modeling of inelastic materials with thermodynamics-based artificial neural networks (TANN)
- A deep material network for multiscale topology learning and accelerated nonlinear modeling of heterogeneous materials
- Accelerating multiscale finite element simulations of history-dependent materials using a recurrent neural network
Cites work
- \(FE^2\) multiscale approach for modelling the elastoviscoplastic behaviour of long fibre SiC/Ti composite materials
- A data-driven computational homogenization method based on neural networks for the nonlinear anisotropic electrical response of graphene/polymer nanocomposites
- A deep material network for multiscale topology learning and accelerated nonlinear modeling of heterogeneous materials
- A manifold learning approach to data-driven computational elasticity and inelasticity
- A method of two-scale analysis with micro-macro decoupling scheme: application to hyperelastic composite materials
- A recurrent neural network-accelerated multi-scale model for elasto-plastic heterogeneous materials subjected to random cyclic and non-proportional loading paths
- Accelerating multiscale finite element simulations of history-dependent materials using a recurrent neural network
- Algorithmic two-scale transition for magneto-electro-mechanically coupled problems, FE \({}^2\)-scheme: localization and homogenization
- An efficient monolithic solution scheme for FE^2 problems
- An FE-DMN method for the multiscale analysis of thermomechanical composites
- Anisotropic hyperelastic constitutive models for finite deformations combining material theory and data-driven approaches with application to cubic lattice metamaterials
- Automated constitutive modeling of isotropic hyperelasticity based on artificial neural networks
- Automated solution of differential equations by the finite element method. The FEniCS book
- Computational homogenization of heterogeneous materials with finite elements
- Computational homogenization of nonlinear elastic materials using neural networks
- Computational micro-to-macro transitions of discretized microstructures undergoing small strains
- Constitutive artificial neural networks: a fast and general approach to predictive data-driven constitutive modeling by deep learning
- Constrained neural network training and its application to hyperelastic material modeling
- Data driven computing with noisy material data sets
- Data-driven computational mechanics
- Data-driven fracture mechanics
- Data-driven tissue mechanics with polyconvex neural ordinary differential equations
- Exploring the 3D architectures of deep material network in data-driven multiscale mechanics
- Finite electro-elasticity with physics-augmented neural networks
- Geometric deep learning for computational mechanics. I: Anisotropic hyperelasticity
- scientific article; zbMATH DE number 1390907 (Why is no real title available?)
- Local approximate Gaussian process regression for data-driven constitutive models: development and comparison with neural networks
- Machine learning based multiscale calibration of mesoscopic constitutive models for composite materials: application to brain white matter
- Model-data-driven constitutive responses: application to a multiscale computational framework
- Model-free data-driven inelasticity
- Model-free data-driven simulation of inelastic materials using structured data sets, tangent space information and transition rules
- 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)
- Nonlinear solid mechanics. A continuum approach for engineering
- Numerical implementation of a neural network based material model in finite element analysis
- On physics-informed data-driven isotropic and anisotropic constitutive models through probabilistic machine learning and space-filling sampling
- On the micromechanics of deep material networks
- Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
- Random heterogeneous materials. Microstructure and macroscopic properties
- Sobolev training of thermodynamic-informed neural networks for interpretable elasto-plasticity models with level set hardening
- The thermodynamics of elastic materials with heat conduction and viscosity
- Thermodynamically consistent machine-learned internal state variable approach for data-driven modeling of path-dependent materials
Cited in
(36)- Local approximate Gaussian process regression for data-driven constitutive models: development and comparison with neural networks
- A multiscale, data-driven approach to identifying thermo-mechanically coupled laws -- bottom-up with artificial neural networks
- Multiscale modeling of inelastic materials with thermodynamics-based artificial neural networks (TANN)
- Accelerating multiscale finite element simulations of history-dependent materials using a recurrent neural network
- An FE-DMN method for the multiscale analysis of short fiber reinforced plastic components
- Accelerated offline setup of homogenized microscopic model for multi‐scale analyses using neural network with knowledge transfer
- Advanced discretization techniques for hyperelastic physics-augmented neural networks
- A comparative study on different neural network architectures to model inelasticity
- Surrogate modeling for the homogenization of elastoplastic composites based on RBF interpolation
- Neural network-based multiscale modeling of finite strain magneto-elasticity with relaxed convexity criteria
- Efficient multiscale modeling of heterogeneous materials using deep neural networks
- Viscoelastic constitutive artificial neural networks (vCANNs) -- a framework for data-driven anisotropic nonlinear finite viscoelasticity
- Nonlinear electro-elastic finite element analysis with neural network constitutive models
- Physics-constrained symbolic model discovery for polyconvex incompressible hyperelastic materials
- Deep learning in computational mechanics: a review
- Learning solutions of thermodynamics-based nonlinear constitutive material models using physics-informed neural networks
- Model-free chemomechanical interfaces: history-dependent damage under transient mass diffusion
- Equivariant graph convolutional neural networks for the representation of homogenized anisotropic microstructural mechanical response
- Reconstruction of unstable atmospheric surface layer streamwise turbulence based on multi-layer perceptron neural network architecture
- Viscoelasticty with physics-augmented neural networks: model formulation and training methods without prescribed internal variables
- Direct data-driven algorithms for multiscale mechanics
- Non-intrusive parametric hyper-reduction for nonlinear structural finite element formulations
- Rediscovering hyperelasticity by deep symbolic regression
- Mixing data-driven and physics-based constitutive models using uncertainty-driven phase fields
- Multiscale modeling of viscoelastic shell structures with artificial neural networks
- An efficient Poisson solver and a data-driven surrogate model for magnetic stray field calculations
- A comparison of classical phenomenological, hybrid neural network, and high performance ROM \(\text{FE}^2\) models for open-cell foams: efficiency, accuracy, and flexibility
- HYDRA: symbolic feature engineering of overparameterized Eulerian hyperelasticity models for fast inference time
- Neural networks meet anisotropic hyperelasticity: a framework based on generalized structure tensors and isotropic tensor functions
- Radial basis function-based surrogate computational homogenization for elastoplastic composites at finite strain
- A non-intrusive, online reduced order method for non-linear micro-heterogeneous materials
- Higher-order multi-scale deep Ritz method (HOMS-DRM) and its convergence analysis for solving thermal transfer problems of composite materials
- Physics-augmented neural networks for constitutive modeling of hyperelastic geometrically exact beams
- EquiNO: a physics-informed neural operator for multiscale simulations
- A data-driven multiscale scheme for anisotropic finite strain magneto-elasticity
- Neural-initialized Newton: accelerating nonlinear finite elements via operator learning
This page was built for publication: \(\mathrm{FE^{ANN}}\): an efficient data-driven multiscale approach based on physics-constrained neural networks and automated data mining
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6101611)
