Learning constitutive models from microstructural simulations via a non-intrusive reduced basis method
From MaRDI portal
Publication:2237428
DOI10.1016/J.CMA.2021.113924zbMATH Open1506.74234arXiv2104.04451OpenAlexW3170539680WikidataQ118178326 ScholiaQ118178326MaRDI QIDQ2237428FDOQ2237428
Authors: Theron Guo, Karen Veroy, Ondřej Rokoš 
Publication date: 27 October 2021
Published in: Computer Methods in Applied Mechanics and Engineering (Search for Journal in Brave)
Abstract: In order to optimally design materials, it is crucial to understand the structure-property relations in the material by analyzing the effect of microstructure parameters on the macroscopic properties. In computational homogenization, the microstructure is thus explicitly modeled inside the macrostructure, leading to a coupled two-scale formulation. Unfortunately, the high computational costs of such multiscale simulations often render the solution of design, optimization, or inverse problems infeasible. To address this issue, we propose in this work a non-intrusive reduced basis method to construct inexpensive surrogates for parametrized microscale problems; the method is specifically well-suited for multiscale simulations since the coupled simulation is decoupled into two independent problems: (1) solving the microscopic problem for different (loading or material) parameters and learning a surrogate model from the data; and (2) solving the macroscopic problem with the learned material model. The proposed method has three key features. First, the microscopic stress field can be fully recovered. Second, the method is able to accurately predict the stress field for a wide range of material parameters; furthermore, the derivatives of the effective stress with respect to the material parameters are available and can be readily utilized in solving optimization problems. Finally, it is more data efficient, i.e. requiring less training data, as compared to directly performing a regression on the effective stress. For the microstructures in the two test problems considered, the mean approximation error of the effective stress is as low as 0.1% despite using a relatively small training dataset. Embedded into the macroscopic problem, the reduced order model leads to an online speed up of approximately three orders of magnitude while maintaining a high accuracy as compared to the FE solver.
Full work available at URL: https://arxiv.org/abs/2104.04451
Recommendations
- Learning constitutive models from microstructural simulations via a non-intrusive reduced basis method: extension to geometrical parameterizations
- Two-stage data-driven homogenization for nonlinear solids using a reduced order model
- Towards surrogate modeling of material microstructures through the processing variables
- A nonlinear data-driven reduced order model for computational homogenization with physics/pattern-guided sampling
- Micromechanics-based surrogate models for the response of composites: a critical comparison between a classical mesoscale constitutive model, hyper-reduction and neural networks
Gaussian process regressionmultiscale analysisproper orthogonal decompositioncomputational homogenizationnon-intrusive reduced basis method
Micromechanics of solids (74M25) Mathematical modeling or simulation for problems pertaining to mechanics of deformable solids (74-10)
Cites Work
- Reduced basis methods for partial differential equations. An introduction
- An `empirical interpolation' method: Application to efficient reduced-basis discretization of partial differential equations
- Nonlinear model reduction via discrete empirical interpolation
- Nonuniform transformation field analysis
- A numerical method for computing the overall response of nonlinear composites with complex microstructure
- Reduced basis hybrid computational homogenization based on a mixed incremental formulation
- The proper generalized decomposition for advanced numerical simulations. A primer
- Efficient non linear model reduction via a least-squares Petrov-Galerkin projection and compressive tensor approximations
- Transformation field analysis of inelastic composite materials
- An approach to micro-macro modeling of heterogeneous materials
- The LATIN multiscale computational method and the proper generalized decomposition
- Multi-scale computational homogenization: trends and challenges
- Hyper-reduction of mechanical models involving internal variables
- The reduced model multiscale method (R3M) for the nonlinear homogenization of hyperelastic media at finite strains
- Computational micro-to-macro transitions of discretized microstructures undergoing small strains
- Computational homogenization for multiscale crack modeling. Implementational and computational aspects
- Computational micro-to-macro transitions for discretized micro-structures of heterogeneous materials at finite strains based on the minimization of averaged incremental energy.
- Certified reduced basis methods for parametrized partial differential equations
- On the computation of the macroscopic tangent for multiscale volumetric homogenization problems
- A comparison of projection-based model reduction concepts in the context of nonlinear biomechanics
- Advanced Lectures on Machine Learning
- POD-DEIM model order reduction for strain-softening viscoplasticity
- Study of efficient homogenization algorithms for nonlinear problems
- Computational homogenization of nonlinear elastic materials using neural networks
- Title not available (Why is that?)
- High-performance model reduction techniques in computational multiscale homogenization
- Self-consistent clustering analysis: an efficient multi-scale scheme for inelastic heterogeneous materials
- Data-driven computational mechanics
- Recent advances on topology optimization of multiscale nonlinear structures
- Non-intrusive reduced order modeling of nonlinear problems using neural networks
- The deep Ritz method: a deep learning-based numerical algorithm for solving variational problems
- Model-free data-driven inelasticity
- Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
- Reduced order modeling for nonlinear structural analysis using Gaussian process regression
- Constitutive artificial neural networks: a fast and general approach to predictive data-driven constitutive modeling by deep learning
- Learning constitutive relations using symmetric positive definite neural networks
- A framework for data-driven analysis of materials under uncertainty: countering the curse of dimensionality
- Data driven computing with noisy material data sets
- Learning constitutive relations from indirect observations using deep neural networks
- A non-intrusive reduced basis method for elastoplasticity problems in geotechnics
- A numerical study of different projection-based model reduction techniques applied to computational homogenisation
- Projection-based model reduction: formulations for physics-based machine learning
- An energy approach to the solution of partial differential equations in computational mechanics via machine learning: concepts, implementation and applications
- Self-consistent clustering analysis for multiscale modeling at finite strains
- Model order reduction of nonlinear homogenization problems using a Hashin-Shtrikman type finite element method
- Micromorphic computational homogenization for mechanical metamaterials with patterning fluctuation fields
- A non-intrusive multifidelity method for the reduced order modeling of nonlinear problems
- Smart constitutive laws: inelastic homogenization through machine learning
- A physics-informed deep learning framework for inversion and surrogate modeling in solid mechanics
- Data-driven reduced order model with temporal convolutional neural network
- Accelerating multiscale finite element simulations of history-dependent materials using a recurrent neural network
- A machine learning based plasticity model using proper orthogonal decomposition
- A recurrent neural network-accelerated multi-scale model for elasto-plastic heterogeneous materials subjected to random cyclic and non-proportional loading paths
Cited In (8)
- Physically recurrent neural networks for path-dependent heterogeneous materials: embedding constitutive models in a data-driven surrogate
- Micromechanics-based surrogate models for the response of composites: a critical comparison between a classical mesoscale constitutive model, hyper-reduction and neural networks
- Multiscale modeling of linear elastic heterogeneous structures via localized model order reduction
- Learning Invariant Representation of Multiscale Hyperelastic Constitutive Law from Sparse Experimental Data
- Stabilization of generalized empirical interpolation method (GEIM) in presence of noise: a novel approach based on Tikhonov regularization
- A reduced order model for geometrically parameterized two-scale simulations of elasto-plastic microstructures under large deformations
- Two-stage data-driven homogenization for nonlinear solids using a reduced order model
- Learning constitutive models from microstructural simulations via a non-intrusive reduced basis method: extension to geometrical parameterizations
Uses Software
This page was built for publication: Learning constitutive models from microstructural simulations via a non-intrusive reduced basis method
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2237428)
