Technology Category
Machine Learning: Deep Generative Models & AutoencodersSearch and Optimization: Metareasoning and MetaheuristicsCognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning
Deep learning-aided inverse design of porous metamaterials
🤖 AI Summary
To address the high computational cost and trial-and-error reliance in inverse design of porous metamaterials, this paper proposes a property-variational autoencoder (pVAE) framework. The pVAE couples a variational autoencoder with a regression network to establish a bidirectional mapping between microstructural configurations and hydraulic properties—specifically porosity and permeability—within an interpretable latent space. Hydraulic property prediction is achieved via joint CNN–lattice Boltzmann method (LBM) modeling, trained on both synthetic data and real micro-CT images. The framework enables structural interpolation, target-property-driven inverse generation, and rapid performance evaluation. It significantly reduces computational overhead while preserving physical consistency, thereby enabling intelligent, customized design of porous microstructures. This work establishes a novel data-driven paradigm for the development of advanced porous materials.
Technology Category
Application Category
📝 Abstract
The ultimate aim of the study is to explore the inverse design of porous metamaterials using a deep learning-based generative framework. Specifically, we develop a property-variational autoencoder (pVAE), a variational autoencoder (VAE) augmented with a regressor, to generate structured metamaterials with tailored hydraulic properties, such as porosity and permeability. While this work uses the lattice Boltzmann method (LBM) to generate intrinsic permeability tensor data for limited porous microstructures, a convolutional neural network (CNN) is trained using a bottom-up approach to predict effective hydraulic properties. This significantly reduces the computational cost compared to direct LBM simulations. The pVAE framework is trained on two datasets: a synthetic dataset of artificial porous microstructures and CT-scan images of volume elements from real open-cell foams. The encoder-decoder architecture of the VAE captures key microstructural features, mapping them into a compact and interpretable latent space for efficient structure-property exploration. The study provides a detailed analysis and interpretation of the latent space, demonstrating its role in structure-property mapping, interpolation, and inverse design. This approach facilitates the generation of new metamaterials with desired properties. The datasets and codes used in this study will be made open-access to support further research.
Problem
Research questions and friction points this paper is trying to address.
Inverse design of porous metamaterials using deep learning
Predict hydraulic properties with reduced computational cost
Generate metamaterials with tailored porosity and permeability
Innovation
Methods, ideas, or system contributions that make the work stand out.
Deep learning-based generative framework for metamaterials
Property-variational autoencoder with regressor augmentation
Convolutional neural network predicts hydraulic properties
🔎 Similar Papers
No similar papers found.
