Abstract:In this study, we developed an inverse analysis framework that proposes a microstructure for dual-phase (DP) steel that exhibits high strength and ductility. The inverse analysis method proposed in this study involves repeated random searches on a model that combines a generative adversarial network (GAN), which generates microstructures, and a convolutional neural network (CNN), which predicts the maximum stress and working limit strain from DP steel microstructures. GAN was trained using images of DP steel microstructures generated by the phase-field method. CNN was trained using images of DP steel microstructures, the maximum stress and the working limit strain calculated by the dislocation-crystal plasticity finite element method. The constructed framework made an efficient search for microstructures possible because of a low-dimensional search space by a latent variable of GAN. The multiple deformation modes were considered in this framework, which allowed the required microstructures to be explored under complex deformation modes. A microstructure with a fine grain size was proposed by using the developed framework.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the trade - off between high strength and high ductility in dual - phase steel (DP steel). Specifically, the research aims to develop an inverse analysis framework that can propose the microstructure of dual - phase steel with high strength and high ductility. By combining the generative adversarial network (GAN) and the convolutional neural network (CNN), this framework can efficiently explore the microstructure that meets the required mechanical properties in a low - dimensional search space and takes into account the complex deformation requirements under multiple deformation modes. ### Background and Objectives of the Paper Dual - phase steel is composed of a soft phase (ferrite) and a hard phase (martensite) and is widely used because of its remarkable mechanical properties. However, the main problem faced by dual - phase steel is the trade - off between strength and ductility. In order to improve the performance of the material, it is necessary to find a microstructure that has both high strength and high ductility. Experimental studies have shown that the mechanical properties of dual - phase steel are affected by the spatial distribution of ferrite and martensite. For example, when the martensite phase forms a chain - like network structure around the ferrite phase, the dual - phase steel has high strength but low ductility; when the martensite phase exists in isolation, the opposite is true. ### Research Methods 1. **Inverse Analysis Framework**: - **Generative Adversarial Network (GAN)**: It is used to generate microstructure images of dual - phase steel. GAN learns to generate new microstructures by training on the microstructure images of dual - phase steel generated by the phase - field method. - **Conative Neural Network (CNN)**: It is used to predict the maximum stress (\(\sigma_{\text{max}}\)) and the working limit strain (\(\epsilon_{\text{lim}}\)) from the microstructure images. CNN is trained by training on the microstructure images generated by the phase - field method and the maximum stress and the working limit strain calculated by the dislocation - crystal plasticity finite element method (FEM). 2. **Consideration of Multiple Deformation Modes**: - This framework takes into account four deformation modes: stretching in the x - direction, stretching in the y - direction, shearing in the x - direction, and shearing in the y - direction. This makes it possible to explore the required microstructure under complex deformation modes. ### Main Steps 1. **Random Selection of Latent Variables**: Randomly select a series of latent variables \(z\) from the latent variable space, and GAN outputs the microstructure images of dual - phase steel according to the selected \(z\). 2. **Prediction of Mechanical Properties**: CNN predicts the maximum stress \(\sigma_{\text{max}}\) and the working limit strain \(\epsilon_{\text{lim}}\) under each deformation mode from the images. 3. **Calculation of Normalized Indicators**: For each deformation mode, calculate the product of the normalized maximum stress and the working limit strain \(\bar{\sigma}_{\text{max}} \cdot \bar{\epsilon}_{\text{lim}}\). 4. **Saving the Optimal Solution**: If \(\bar{\sigma}_{\text{max}} \cdot \bar{\epsilon}_{\text{lim}}\) at the current step is the highest, save the corresponding dual - phase steel microstructure and deformation mode as a temporary optimal solution. 5. **Determination of the Final Solution**: After multiple iterations, determine the dual - phase steel microstructure and deformation mode with the highest \(\bar{\sigma}_{\text{max}} \cdot \bar{\epsilon}_{\text{lim}}\) as the final optimal solution. ### Results and Discussion Through this inverse analysis framework, researchers can efficiently explore the microstructure of dual - phase steel with high strength and high ductility. Experimental results show that this framework can effectively predict mechanical properties and performs well under multiple deformation modes. The finally proposed microstructure has a small grain size, which helps to improve the comprehensive performance of dual - phase steel. ### Conclusions This research has successfully developed an inverse analysis framework that can efficiently explore and propose the microstructure of dual - phase steel with high strength and high ductility. By combining GAN and CNN, this framework not only improves the search efficiency but also takes into account the complex requirements under multiple deformation modes, providing new ideas and methods for the material development of dual - phase steel.

Investigation on optimal microstructure of dual-phase steel with high strength and ductility by machine learning

Maximization of strength–ductility balance of dual-phase steels using generative adversarial networks and Bayesian optimization

Comparative study of the experimentally observed and GAN-generated 3D microstructures in dual-phase steels

Employing Deep Learning in Non‐parametric Inverse Visualization of Elastic–plastic Mechanisms in Dual‐phase Steels

Analysis and design of dual-phase steel microstructure for enhanced ductile fracture resistance

Deep learning-based method for microstructure-property linkage of dual-phase steel

Achieving High Strength and High Ductility of Dual-Phase Steel Via Alternating Lamellar Microstructure

Dual-phase steels microstructure and properties consideration based on artificial intelligence techniques

High resolution in situ mapping of microstrain and microstructure evolution reveals damage resistance criteria in dual phase steels

A deep learning method for extensible microstructural quantification of DP steel enhanced by physical metallurgy-guided data augmentation

Hybrid Data-Driven Deep Learning Framework for Material Mechanical Properties Prediction with the Focus on Dual-Phase Steel Microstructures

Accelerating the Discovery of New DP Steel Using Machine Learning-Based Multiscale Materials Simulations

Virtual Laboratory Enabled Constitutive Modelling of Dual Phase Steels

Building a Quantitative Composition-Microstructure-property Relationship of Dual-Phase Steels Via Multimodal Data Mining

Two- and Three-Dimensional Modeling and Simulations of Grain Growth Behavior in Dual-Phase Steel Using Monte Carlo and Machine Learning

Study on Microstructural Evolution of DP Steel Considering the Interface Layer Based on Multi Mechanism Strain Gradient Theory

Failure Mode And Ductility Of Dual Phase Steel With Edge Crack

Strain-Hardening Behaviors of Dual Phase Steels with Microstructure Features

Mathematical Morphology Applied to the Study of Dual Phase Steel Formation

An Overview of Dual-Phase Steels: Advances in Microstructure-Oriented Processing and Micromechanically Guided Design

Achieving excellent strength and ductility synergy by δ phase strengthening in low-density high-manganese steel