Abstract:A global prediction of macroscale deformation and microstructure evolution under complex mechanical and thermal fields is crucial for customizing the expected shape and performance of a component. To overcome the shortcomings of multiscale models on predictions at individual scale, a cross-scale model for global prediction from representative volume element (RVE) to component is created here, in which the cellular automata crystal Plasticity finite element method (CACPFEM) and artificial neural network (ANN) were combined cleverly. The CACPFEM model, which fully couples the heterogeneous deformation and the microstructure evolution, e.g., dynamic recrystallization (DRX), is to account for the responses of RVE. Quantities of constant and varying loading paths were applied to RVE in order to reflect the dependences of responses on strain rate, temperature, microstructure and deformation mode. All of the responses (including mechanical and microstructural ones) form a huge database, based on which back propagation (BP) ANN models with Marquardt-Levenberg (M-L) algorithms were established through training, validating, testing and loops of optimizations. The outputs of the ANN models are set as the microstructural evolution (including DRX volume fraction and average grain size) and the dynamically varying macroscale parameters of J2-J3 constitutive model dependent on loading paths and microstructure, which are then applied to the FEM model to predict the responses of a component. Thus, a bridge was built to connect the responses of RVE and component. In turn, the deformation history of a local region at the component can also be applied to RVE to further study the microscale deformation mechanism and microstructural evolution. With the cross-scale model, the results reflecting the characteristics of anisotropy, tension-compression asymmetry, dependences on strain rate, temperature, microstructure and deformation mode were obtained. It benefits from the physically-based CACPFEM, the J2-J3 constitutive model with dependences on deformation condition and microstructure evolution, the well optimized ANN model, and their innovative combination. The optimization strategy guarantees the cross-scale prediction accuracy. The applications of the cross-scale model to the uni-axial compression of a revolving billet and the extension to the yield surface prediction and the forging process of a new shape billet show the cross-scale prediction capability of the model.

A micromechanics-based artificial neural networks model for rapid prediction of mechanical response in short fiber reinforced rubber composites

Machine Learning and Materials Informatics Approaches for Predicting Transverse Mechanical Properties of Unidirectional CFRP Composites with Microvoids

Prediction of Macroscopic Compressive Mechanical Properties for 2.5D Woven Composites Based on Artificial Neural Network

A micromechanics-based recurrent neural networks model for path-dependent cyclic deformation of short fiber composites

A mechanics-informed machine learning approach for modeling the elastoplastic behavior of fiber-reinforced composites

Artificial Neural Networks in Prediction of Mechanical Behavior of High Performance Plastic Composites

Predicting Mechanical Properties of High-Performance Fiber-Reinforced Cementitious Composites by Integrating Micromechanics and Machine Learning

An innovative computational framework for the analysis of complex mechanical behaviors of short fiber reinforced polymer composites

Artificial Neural Network-Based Homogenization Model for Predicting Multiscale Thermo-Mechanical Properties of Woven Composites

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

Micromechanical Predictions on Elastic Moduli of a Short Fiber Composite with Arbitrary Geometric Combination

Study of effective elastic properties of heterogeneous materials with an artificial neural network model

Artificial Neural Network (Ann)-Based Residual Strength Prediction of Carbon Fibre Reinforced Composites (cfrcs) after Impact

Tensile Performance Prediction of CFRPs with Voids Using Multiscale Analysis and Neural Networks

CFRP-RC fracture strength prediction: A novel approach using deep learning

Numerical Prediction of Mechanical Properties of Rubber Composites Reinforced by Aramid Fiber under Large Deformation

Cross-scale Prediction from RVE to Component

Strength prediction and progressive damage analysis of carbon fiber reinforced polymer-laminate with circular holes by an efficient Artificial Neural Network

LS-DYNA Machine Learning-based Multiscale Method for Nonlinear Modeling of Short Fiber-Reinforced Composites

Mechanical Properties Prediction of Composite Laminate with FEA and Machine Learning Coupled Method

Highly Efficient Simulation of Composites by Determining Failure Initiation and Fracture Angle with Artificial Neural Networks