skip to main content

Explore Research Products in the NSF-PAR

This content will become publicly available on August 1, 2024

Title: Denoising diffusion algorithm for inverse design of microstructures with fine-tuned nonlinear material properties
We introduce a denoising diffusion algorithm to discover microstructures with nonlinear fine-tuned properties. Denoising diffusion probabilistic models are generative models that use diffusion-based dynamics to gradually denoise images and generate realistic synthetic samples. By learning the reverse of a Markov diffusion process, we design an artificial intelligence to efficiently manipulate the topology of microstructures to generate a massive number of prototypes that exhibit constitutive responses sufficiently close to designated nonlinear constitutive behaviors. To identify the subset of microcstructures with sufficiently precise fine-tuned properties, a convolutional neural network surrogate is trained to replace high-fidelity finite element simulations to filter out prototypes outside the admissible range. Results of this study indicate that the denoising diffusion process is capable of creating microstructures of fine-tuned nonlinear material properties within the latent space of the training data. More importantly, this denoising diffusion algorithm can be easily extended to incorporate additional topological and geometric modifications by introducing high-dimensional structures embedded in the latent space. Numerical experiments are conducted on the open-source mechanical MNIST data set (Lejeune, 2020). Consequently, this algorithm is not only capable of performing inverse design of nonlinear effective media, but also learns the nonlinear structure–property map to quantitatively understand the multiscale interplay among the geometry, topology, and their effective macroscopic properties.  more » « less
Award ID(s):
1846875
NSF-PAR ID:
10487106
Author(s) / Creator(s):
;
Publisher / Repository:
Computer Methods in Applied Mechanics and Engineering
Date Published:
Journal Name:
Computer Methods in Applied Mechanics and Engineering
Volume:
413
Issue:
C
ISSN:
0045-7825
Page Range / eLocation ID:
116126
Subject(s) / Keyword(s):
["Denoising diffusion","Inverse design","Latent space","Fine-tuning material properties"]
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; ( , Journal of Mechanical Design)
    null (Ed.)
    Abstract The data-driven approach is emerging as a promising method for the topological design of multiscale structures with greater efficiency. However, existing data-driven methods mostly focus on a single class of microstructures without considering multiple classes to accommodate spatially varying desired properties. The key challenge is the lack of an inherent ordering or “distance” measure between different classes of microstructures in meeting a range of properties. To overcome this hurdle, we extend the newly developed latent-variable Gaussian process (LVGP) models to create multi-response LVGP (MR-LVGP) models for the microstructure libraries of metamaterials, taking both qualitative microstructure concepts and quantitative microstructure design variables as mixed-variable inputs. The MR-LVGP model embeds the mixed variables into a continuous design space based on their collective effects on the responses, providing substantial insights into the interplay between different geometrical classes and material parameters of microstructures. With this model, we can easily obtain a continuous and differentiable transition between different microstructure concepts that can render gradient information for multiscale topology optimization. We demonstrate its benefits through multiscale topology optimization with aperiodic microstructures. Design examples reveal that considering multiclass microstructures can lead to improved performance due to the consistent load-transfer paths for micro- and macro-structures. 
    more » « less
  2. ; ; ; ( , Information Visualization)
    Transformer-based language models such as BERT and its variants have found widespread use in natural language processing (NLP). A common way of using these models is to fine-tune them to improve their performance on a specific task. However, it is currently unclear how the fine-tuning process affects the underlying structure of the word embeddings from these models. We present TopoBERT, a visual analytics system for interactively exploring the fine-tuning process of various transformer-based models – across multiple fine-tuning batch updates, subsequent layers of the model, and different NLP tasks – from a topological perspective. The system uses the mapper algorithm from topological data analysis (TDA) to generate a graph that approximates the shape of a model’s embedding space for an input dataset. TopoBERT enables its users (e.g. experts in NLP and linguistics) to (1) interactively explore the fine-tuning process across different model-task pairs, (2) visualize the shape of embedding spaces at multiple scales and layers, and (3) connect linguistic and contextual information about the input dataset with the topology of the embedding space. Using TopoBERT, we provide various use cases to exemplify its applications in exploring fine-tuned word embeddings. We further demonstrate the utility of TopoBERT, which enables users to generate insights about the fine-tuning process and provides support for empirical validation of these insights. 
    more » « less
  3. ; ; ( , Polymer Composites)

    The stochastic nature of fiber distribution over transversal cross‐section in unidirectional fiber‐reinforced composites has a significant effect on the microscopical distribution of stresses and strains. Towards simulation‐based analysis and design of fiber‐reinforced composite, this paper concentrates on the issue of random fiber distributions and a new approach is developed to effectively generate microstructures of composite material with high fiber volume fraction. The method, named random distribution particle swarm optimization (PSO) algorithm, is proposed based on the particle movement regulation and position updating principle of PSO. The method creates microstructures with different fiber volume fractions conveniently by distributing each single fiber with inter‐fiber distance arrangement. The performance of algorithm is analyzed statistically and the results demonstrate that it is capable of catching the typical distribution of these materials for both low and high fiber volume fractions. Finite‐element analysis is conducted to predict effective properties of the generated microstructures and the simulation results exhibit reasonable agreement with experimental results. The developed algorithm is easily realized and highly efficient without introducing heuristic steps to achieve high volume fraction. The approach provides a valuable alternative to generate random numerical models used in micromechanical analysis of unidirectional fiber‐reinforced composites. POLYM. COMPOS., 40:1643–1653, 2019. © 2018 Society of Plastics Engineers

     
    more » « less
  4. ; ; ; ( , ASME 2020 Conference on Smart Materials, Adaptive Structures and Intelligent Systems)
    null (Ed.)
    Abstract

    Composites can be tailored to specific applications by adjusting process variables. These variables include those related to composition, such as volume fraction of the constituents and those associated with processing methods, methods that can affect composite topology. In the case of particle matrix composites, orientation of the inclusions affects the resulting composite properties, particularly so in instances where the particles can be oriented and arranged into structures. In this work, we study the effects of coupled electric and magnetic field processing with externally applied fields on those structures, and consequently on the resulting material properties that arise. The ability to vary these processing conditions with the goal of generating microstructures that yield target material properties adds an additional level of control to the design of composite material properties. Moreover, while analytical models allow for the prediction of resulting composite properties from constituents and composite topology, these models do not build upward from process variables to make these predictions.

    This work couples simulation of the formation of microscale architectures, which result from coupled electric and magnetic field processing of particulate filled polymer matrix composites, with finite element analysis of those structures to provide a direct and explicit linkages between process, structure, and properties. This work demonstrates the utility of these method as a tool for determining composite properties from constituent and processing parameters. Initial particle dynamics simulation incorporating electromagnetic responses between particles and between the particles and the applied fields, including dielectrophoresis, are used to stochastically generate representative volume elements for a given set of process variables. Next, these RVEs are analyzed as periodic structures using FEA yielding bulk material properties. The results are shown to converge for simulation size and discretization, validating the RVE as an appropriate representation of the composite volume. Calculated material properties are compared to traditional effective medium theory models. Simulations allow for mapping of composite properties with respect to not only composition, but also fundamentally from processing simulations that yield varying particle configurations, a step not present in traditional or more modern effective medium theories such as the Halpin Tsai or double-inclusion theories.

     
    more » « less
  5. ; ; ; ; ; ( , Journal of Mechanical Design)
    null (Ed.)
    Abstract Scientific and engineering problems often require the use of artificial intelligence to aid understanding and the search for promising designs. While Gaussian processes (GP) stand out as easy-to-use and interpretable learners, they have difficulties in accommodating big data sets, categorical inputs, and multiple responses, which has become a common challenge for a growing number of data-driven design applications. In this paper, we propose a GP model that utilizes latent variables and functions obtained through variational inference to address the aforementioned challenges simultaneously. The method is built upon the latent-variable Gaussian process (LVGP) model where categorical factors are mapped into a continuous latent space to enable GP modeling of mixed-variable data sets. By extending variational inference to LVGP models, the large training data set is replaced by a small set of inducing points to address the scalability issue. Output response vectors are represented by a linear combination of independent latent functions, forming a flexible kernel structure to handle multiple responses that might have distinct behaviors. Comparative studies demonstrate that the proposed method scales well for large data sets with over 104 data points, while outperforming state-of-the-art machine learning methods without requiring much hyperparameter tuning. In addition, an interpretable latent space is obtained to draw insights into the effect of categorical factors, such as those associated with “building blocks” of architectures and element choices in metamaterial and materials design. Our approach is demonstrated for machine learning of ternary oxide materials and topology optimization of a multiscale compliant mechanism with aperiodic microstructures and multiple materials. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email