Abstract:Liquid metals (LM) are embedded in an elastomer matrix to obtain soft composites with unique thermal, dielectric, and mechanical properties. They have applications in soft robotics, biomedical engineering, and wearable electronics. By linking the structure to the properties of these materials, it is possible to perform material design rationally. Liquid-metal embedded elastomers (LMEEs) have been designed for targeted electro-thermo-mechanical properties by semi-supervised learning of structure-property (SP) links in a variational autoencoder network (VAE). The design parameters are the microstructural descriptors that are physically meaningful and have affine relationships with the synthetization of the studied particulate composite. The machine learning (ML) model is trained on a generated dataset of microstructural descriptors with their multifunctional property quantities as their labels. Sobol sequence is used for in-silico Design of Experiment (DoE) by sampling the design space to generate a comprehensive dataset of 3D microstructure realizations via a packing algorithm. The mechanical responses of the generated microstructures are simulated using a previously developed Finite Element (FE) model, considering the surface tension induced by LM inclusions, while the linear thermal and dielectric constants are homogenized with the help of our in-house Fast Fourier Transform (FFT) package. Following the training by minimization of an appropriate loss function, the VAE encoder acts as the surrogate of numerical solvers of the multifunctional homogenizations, and its decoder is used for the material design. Our results indicate the satisfactory performance of the surrogate model and the inverse calculator with respect to high-fidelity numerical simulations validated with LMEE experimental results.

What problem does this paper attempt to address?

The problems that this paper attempts to solve are as follows: Through machine - learning methods, especially variational auto - encoders (VAE), establish a connection between the microstructure of liquid - metal - embedded elastomers (LMEEs) and their multi - physical properties (mechanical, thermal, and dielectric properties), thereby achieving the rational design of materials. Specifically, the author hopes: 1. **Establish the connection between microstructure and properties**: Generate a data set containing a large number of microstructure descriptors and their corresponding multi - physical properties, and train a machine - learning model to reveal how the microstructure of these materials affects their macroscopic properties. 2. **Accelerate the material design process**: Use the trained VAE model as a surrogate model to quickly predict the properties of new materials, and perform reverse design through the decoder to find the microstructure with specific target properties, thereby replacing the traditional computationally - intensive numerical simulations and experimental trial - and - error methods. 3. **Improve the design accuracy**: Ensure that the proposed machine - learning framework can accurately predict the multi - physical properties of LMEEs, and can generate multiple possible microstructures in reverse design, rather than just reproducing the input structure. ### Key issues - **Selection of microstructure descriptors**: Select appropriate physical descriptors (such as volume fraction, aspect ratio, etc.) to characterize the microstructure of LMEEs. - **Data generation and processing**: Use the Sobol sequence to generate a uniformly - distributed training data set, and calculate the corresponding multi - physical properties through finite - element analysis (FEA) and fast Fourier transform (FFT). - **Model training and evaluation**: Train the VAE model and evaluate its performance as a surrogate model and a reverse - design tool through multiple metrics. ### Formula representation The formulas involved in the paper include: - Neo - Hookean strain - energy expression: \[ U = C_{10}(\overline{I}_1 - 3)+\frac{(J_{el} - 1)^2}{D_1} \] where \(C_{10}=\frac{\mu_0}{2}\), \(D_1 = \frac{2}{K_0}\). - Constitutive equation of thermal conductivity: \[ \bar{q}=K\bar{e} \] where \(K\) is the thermal - conductivity tensor, \(\bar{e}\) is the temperature - gradient vector, and \(\bar{q}\) is the heat - flow vector. - Constitutive equation of dielectric constant: \[ \bar{D}=\epsilon\bar{E} \] where \(\epsilon\) is the absolute dielectric constant, \(\bar{D}\) is the electric - displacement field, and \(\bar{E}\) is the applied electric field. - Loss function: \[ Loss = L_{encoder}+L_{decoder} \] where \[ L_{decoder}=L_1=\frac{1}{N}\sum_{i = 1}^{N}(x_i-\hat{x}_i)^2 \] \[ L_{encoder}=L_2=\frac{1}{N}\sum_{i = 1}^{N}(y_i-\hat{y}_i)^2 \] Through the solution of these problems, this research aims to provide an efficient and accurate method for the design of multifunctional composite materials.

Linking Properties to Microstructure in Liquid Metal Embedded Elastomers via Machine Learning

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

Lightweight, Thermally Conductive Liquid Metal Elastomer Composite with Independently Controllable Thermal Conductivity and Density

A supervised machine learning approach for accelerating the design of particulate composites: Application to thermal conductivity

Deep generative modeling for mechanistic-based learning and design of metamaterial systems

Physics-informed machine learning for composition-process-property alloy design: shape memory alloy demonstration

(Invited) Combining Machine Learning, DFT, EFM, and Modeling to Design Nanodielectric Behavior

Prediction of Equivalent Elastic Modulus for Metal-Coated Lattice Based on Machine Learning

Computational morphogenesis for liquid crystal elastomer metamaterial

Unveiling Processing--Property Relationships in Laser Powder Bed Fusion: The Synergy of Machine Learning and High-throughput Experiments

A multifunctional shape-morphing elastomer with liquid metal inclusions

Regression analysis for predicting the elasticity of liquid crystal elastomers

Machine‐Learning Microstructure for Inverse Material Design

Smooth Like Butter: Evaluating Multi-Lattice Transitions in Property-Augmented Latent Spaces

Universal Design Methodology for Printable Microstructural Materials via a New Deep Generative Learning Model: Application to a Piezocomposite

Liquid metal-based elastomer heat conduction enhancement enabled by stretching

Accelerating microstructure modelling via machine learning: a new method combining Autoencoder and ConvLSTM

Creating a Microstructure Latent Space with Rich Material Information for Multiphase Alloy Design

Customizable anisotropic microlattices for additive manufacturing: Machine learning accelerated design, mechanical properties and structural-property relationships

Segregated and Non‐Settling Liquid Metal Elastomer via Jamming of Elastomeric Particles

Integrating machine learning and multi-objective optimization to investigate the magnetic and mechanical properties of FeSiCr soft magnetic composite processed by selective laser melting