Abstract:The structure of the non-linear constitutive models is a key to control non-linear behaviours of materials. Because the non-linear mechanical mechanism is not clearly understood in most cases, it is very difficult to assume the structure of the model in advance. The recognition of the structure of the model from experimental results can help understanding of the mechanism. This recognition is a dynamic search problem being highly multimodal, multi-variable with high order, and needing a large parameter space. How to obtain a global optimum solution is a key to this problem. In this paper, a hybrid evolutionary algorithm is proposed for coupling recognition of the structure of the non-linear constitutive material model and its coefficients in global space using global response information, e.g. load vs deflection data, obtained from the structural test. Genetic programming is used to recognize the structure of the non-linear stress-strain relationship without any assumption in advance and the genetic algorithm is then used to recognize its coefficients. The non-linear stress-strain relationship thus found can not only satisfy the dynamic change in its structure but also its variables and coefficients. Non-linear finite element analysis is used to transfer the load-deflection information to the stress-strain data. The potential of the proposed method is demonstrated by applying it to the macro-mechanical modelling of the non-linear behaviour of composite materials. A non-linear material model for the unidirectional ply is recognized by using experimental data of a lamina plate [(+/-45)(6)](s). The obtained non-linear constitutive model gave good predictions in coincidence with the non-linear behaviours of the [(+/-30)(6)](s), [(0/+/-45)(3)](s) and [(0/+/-45)(4)](s) plates. The results indicate that the coupling non-linear constitutive model of the structure and its coefficients can identify the model which the traditional constitutive model theory is unable to recognize. Copyright (C) 2004 John Wiley Sons, Ltd.

Autoprogressive Training of Neural Network Constitutive Models

Automated model discovery of finite strain elastoplasticity from uniaxial experiments

Learning solutions of thermodynamics-based nonlinear constitutive material models using physics-informed neural networks

A generic physics-informed neural network-based constitutive model for soft biological tissues

A neural network enhanced system for learning nonlinear constitutive law and failure initiation criterion of composites using indirectly measurable data

Genetic evolution of nonlinear material constitutive models

Learning solution of nonlinear constitutive material models using physics-informed neural networks: COMM-PINN

A neural network-based material cell for elastoplasticity and its performance in FE analyses of boundary value problems

A new family of Constitutive Artificial Neural Networks towards automated model discovery

Learning constitutive relations using symmetric positive definite neural networks

Dynamic Analysis of Nonlinear Civil Engineering Structures using Artificial Neural Network with Adaptive Training

Coupling Recognition of the Structure and Parameters of Non‐linear Constitutive Material Models Using Hybrid Evolutionary Algorithms

A Physics Preserving Neural Network Based Approach for Constitutive Modeling of Isotropic Fibrous Materials

Learning the physics-consistent material behavior from experimentally measurable data via PDE-constrained optimization

Viscoelasticty with physics-augmented neural networks: model formulation and training methods without prescribed internal variables

Assessing Feed-Forward Backpropagation Artificial Neural Networks for Strain-Rate-Sensitive Mechanical Modeling

Nonlinear electro-elastic finite element analysis with neural network constitutive models

Deep Long Short-Term Memory Neural Network for Accelerated Elastoplastic Analysis of Heterogeneous Materials: an Integrated Data-Driven Surrogate Approach

Structural stress prediction based on deformations using artificial neural networks trained with artificial noise

Artificial neural networks and intelligent finite elements in non-linear structural mechanics

A deep learning framework for constitutive modeling based on temporal convolutional network