Learning constitutive models from microstructural simulations via a non-intrusive reduced basis method
From MaRDI portal
(Redirected from Publication:2237428)
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.
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
Cites work
- scientific article; zbMATH DE number 1213967 (Why is no real title available?)
- A comparison of projection-based model reduction concepts in the context of nonlinear biomechanics
- A framework for data-driven analysis of materials under uncertainty: countering the curse of dimensionality
- A machine learning based plasticity model using proper orthogonal decomposition
- A non-intrusive multifidelity method for the reduced order modeling of nonlinear problems
- A non-intrusive reduced basis method for elastoplasticity problems in geotechnics
- A numerical method for computing the overall response of nonlinear composites with complex microstructure
- A numerical study of different projection-based model reduction techniques applied to computational homogenisation
- A physics-informed deep learning framework for inversion and surrogate modeling in solid mechanics
- 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
- Advanced Lectures on Machine Learning
- An `empirical interpolation' method: Application to efficient reduced-basis discretization of partial differential equations
- An approach to micro-macro modeling of heterogeneous materials
- An energy approach to the solution of partial differential equations in computational mechanics via machine learning: concepts, implementation and applications
- Certified reduced basis methods for parametrized partial differential equations
- Computational homogenization for multiscale crack modeling. Implementational and computational aspects
- Computational homogenization of nonlinear elastic materials using neural networks
- Computational micro-to-macro transitions for discretized micro-structures of heterogeneous materials at finite strains based on the minimization of averaged incremental energy.
- 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
- Data driven computing with noisy material data sets
- Data-driven computational mechanics
- Data-driven reduced order model with temporal convolutional neural network
- Efficient non linear model reduction via a least-squares Petrov-Galerkin projection and compressive tensor approximations
- High-performance model reduction techniques in computational multiscale homogenization
- Hyper-reduction of mechanical models involving internal variables
- Learning constitutive relations from indirect observations using deep neural networks
- Learning constitutive relations using symmetric positive definite neural networks
- Micromorphic computational homogenization for mechanical metamaterials with patterning fluctuation fields
- Model order reduction of nonlinear homogenization problems using a Hashin-Shtrikman type finite element method
- Model-free data-driven inelasticity
- Multi-scale computational homogenization: trends and challenges
- Non-intrusive reduced order modeling of nonlinear problems using neural networks
- Nonlinear model reduction via discrete empirical interpolation
- Nonuniform transformation field analysis
- On the computation of the macroscopic tangent for multiscale volumetric homogenization problems
- POD-DEIM model order reduction for strain-softening viscoplasticity
- Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
- Projection-based model reduction: formulations for physics-based machine learning
- Recent advances on topology optimization of multiscale nonlinear structures
- Reduced basis hybrid computational homogenization based on a mixed incremental formulation
- Reduced basis methods for partial differential equations. An introduction
- Reduced order modeling for nonlinear structural analysis using Gaussian process regression
- Self-consistent clustering analysis for multiscale modeling at finite strains
- Self-consistent clustering analysis: an efficient multi-scale scheme for inelastic heterogeneous materials
- Smart constitutive laws: inelastic homogenization through machine learning
- Study of efficient homogenization algorithms for nonlinear problems
- The LATIN multiscale computational method and the proper generalized decomposition
- The deep Ritz method: a deep learning-based numerical algorithm for solving variational problems
- The proper generalized decomposition for advanced numerical simulations. A primer
- The reduced model multiscale method (R3M) for the nonlinear homogenization of hyperelastic media at finite strains
- Transformation field analysis of inelastic composite materials
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
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)
