skip to main content


Title: Generative Adversarial Networks and Mixture Density Networks-Based Inverse Modeling for Microstructural Materials Design
Abstract

There are two broad modeling paradigms in scientific applications: forward and inverse. While forward modeling estimates the observations based on known causes, inverse modeling attempts to infer the causes given the observations. Inverse problems are usually more critical as well as difficult in scientific applications as they seek to explore the causes that cannot be directly observed. Inverse problems are used extensively in various scientific fields, such as geophysics, health care and materials science. Exploring the relationships from properties to microstructures is one of the inverse problems in material science. It is challenging to solve the microstructure discovery inverse problem, because it usually needs to learn a one-to-many nonlinear mapping. Given a target property, there are multiple different microstructures that exhibit the target property, and their discovery also requires significant computing time. Further, microstructure discovery becomes even more difficult because the dimension of properties (input) is much lower than that of microstructures (output). In this work, we propose a framework consisting of generative adversarial networks and mixture density networks for inverse modeling of structure–property linkages in materials, i.e., microstructure discovery for a given property. The results demonstrate that compared to baseline methods, the proposed framework can overcome the above-mentioned challenges and discover multiple promising solutions in an efficient manner.

 
more » « less
Award ID(s):
2053929
NSF-PAR ID:
10379295
Author(s) / Creator(s):
; ; ; ; ; ;
Publisher / Repository:
Springer Science + Business Media
Date Published:
Journal Name:
Integrating Materials and Manufacturing Innovation
Volume:
11
Issue:
4
ISSN:
2193-9764
Page Range / eLocation ID:
p. 637-647
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Multi-principal element alloys (MPEAs) with remarkable performances possess great potential as structural, functional, and smart materials. However, their efficient performance-orientated design in a wide range of compositions and types is an extremely challenging issue, because of properties strongly dependent upon the composition and composition-dominated microstructure. Here, we propose a multistage-design approach integrating machine learning, physical laws and a mathematical model for developing the desired-property MPEAs in a very time-efficient way. Compared to the existing physical model- or machine-learning-assisted material development, the forward-and-inverse problems, including identifying the target property and unearthing the optimal composition, can be tackled with better efficiency and higher accuracy using our proposed avenue, which defeats the one-step component-performance design strategy by multistage-design coupling constraints. Furthermore, we developed a new multi-phase MPEA at the minimal time and cost, whose high strength-ductility synergy exceeded those of its system and subsystem reported so far by searching for the optimal combination of phase fraction and composition. The present work suggests that the property-guided composition and microstructure are precisely tailored through the newly built approach with significant reductions of the development period and cost, which is readily extendable to other multi-principal element materials. 
    more » « less
  2. Abstract Materials design aims to identify the material features that provide optimal properties for various engineering applications, such as aerospace, automotive, and naval. One of the important but challenging problems for materials design is to discover multiple polycrystalline microstructures with optimal properties. This paper proposes an end-to-end artificial intelligence (AI)-driven microstructure optimization framework for elastic properties of materials. In this work, the microstructure is represented by the Orientation Distribution Function (ODF) that determines the volume densities of crystallographic orientations. The framework was evaluated on two crystal systems, cubic and hexagonal, for Titanium (Ti) in Joint Automated Repository for Various Integrated Simulations (JARVIS) database and is expected to be widely applicable for materials with multiple crystal systems. The proposed framework can discover multiple polycrystalline microstructures without compromising the optimal property values and saving significant computational time. 
    more » « less
  3. Data-driven methods have attracted increasingly more attention in materials research since the advent of the material genome initiative. The combination of materials science with computer science, statistics, and data-driven methods aims to expediate materials research and applications and can utilize both new and archived research data. In this paper, we present a data driven and deep learning approach that builds a portion of the structure–property relationship for polymer nanocomposites. Analysis of archived experimental data motivates development of a computational model which allows demonstration of the approach and gives flexibility to sufficiently explore a wide range of structures. Taking advantage of microstructure reconstruction methods and finite element simulations, we first explore qualitative relationships between microstructure descriptors and mechanical properties, resulting in new findings regarding the interplay of interphase, volume fraction and dispersion. Then we present a novel deep learning approach that combines convolutional neural networks with multi-task learning for building quantitative correlations between microstructures and property values. The performance of the model is compared with other state-of-the-art strategies including two-point statistics and structure descriptor-based approaches. Lastly, the interpretation of the deep learning model is investigated to show that the model is able to capture physical understandings while learning. 
    more » « less
  4. Bridging the gaps among various categories of stochastic microstructures remains a challenge in the design representation of microstructural materials. Each microstructure category requires certain unique mathematical and statistical methods to define the design space (design representation). The design representation methods are usually incompatible between two different categories of stochastic microstructures. The common practice of pre-selecting the microstructure category and the associated design representation method before conducting rigorous computational design restricts the design freedom and hinders the discovery of innovative microstructure designs. To overcome this issue, this paper proposes and compares two novel methods, the deep generative modeling-based method and the curvature functional-based method, to understand their pros and cons in designing mixed-category stochastic microstructures for desired properties. For the deep generative modeling-based method, the Variational Autoencoder is employed to generate an unstructured latent space as the design space. For the curvature functional-based method, the microstructure geometry is represented by curvature functionals, of which the functional parameters are employed as the microstructure design variables. Regressors of the microstructure design variables-property relationship are trained for microstructure design optimization. A comparative study is conducted to understand the relative merits of these two methods in terms of computational cost, continuous transition, design scalability, design diversity, dimensionality of the design space, interpretability of the statistical equivalency, and design performance.

     
    more » « less
  5. Joan Bruna, Jan S (Ed.)
    In recent years, the field of machine learning has made phenomenal progress in the pursuit of simulating real-world data generation processes. One notable example of such success is the variational autoencoder (VAE). In this work, with a small shift in perspective, we leverage and adapt VAEs for a different purpose: uncertainty quantification in scientific inverse problems. We introduce UQ-VAE: a flexible, adaptive, hybrid data/model-constrained framework for training neural networks capable of rapid modelling of the posterior distribution representing the unknown parameter of interest. Specifically, from divergence-based variational inference, our framework is derived such that most of the information usually present in scientific inverse problems is fully utilized in the training procedure. Additionally, this framework includes an adjustable hyperparameter that allows selection of the notion of distance between the posterior model and the target distribution. This introduces more flexibility in controlling how optimization directs the learning of the posterior model. Further, this framework possesses an inherent adaptive optimization property that emerges through the learning of the posterior uncertainty. Numerical results for an elliptic PDE-constrained Bayesian inverse problem are provided to verify the proposed framework. 
    more » « less