Abstract:Metamaterials can be designed to contain functional gradients with negative Poisson's ratio that have counterintuitive behavior compared to monolithic materials. These negative Poisson's ratio materials, referred to as auxetics, are important to engineering sciences because of their unique mechanical expansion. Previous studies have explored compliant actuators using analytical and numerically derived mechanics of materials principles. However, the control of compliant gradient mechanisms frequently uses complex analytical equations combined with traditional control algorithms making them difficult to design. In order to confront the design processes and computational load, this study uses machine learning to predict errors in compliant auxetic designs based on a mathematically optimal deformation. Finite element analysis and experimental specimens validate the theoretical mechanical behavior of a specific auxetic configuration as well as demonstrate the manufacturing capabilities of additive manufacturing of graded auxetic materials. We then use pseudorandomized images and their respective computational deformation results to train a regressive model and predict the deviation from the optimal behavior. The model predicts the deviation from the desired behavior with a mean average percent error below 5% for the validation set. This study subsequently proposes a scalable workflow design process connecting the unique performance of auxetics to machine learning design predictions.This article is protected by copyright. All rights reserved.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to accelerate the design process of auxetic metamaterials with negative Poisson's ratio (NPR) and reduce their computational burden. Specifically, researchers use deep - learning techniques to predict the behavioral deviations of these metamaterials during deformation, thereby simplifying and optimizing the design process. ### Problem Background Traditional metamaterial design relies on complex analytical equations and traditional control algorithms, which makes the design process very complex and time - consuming. Especially for metamaterials with gradient functions, the complexity of their internal structures and mechanical behaviors further increases the design difficulty. To overcome these challenges, this paper proposes a deep - learning - based method to accelerate the design of metamaterials and improve design efficiency. ### Main Objectives 1. **Predict Deformation Errors**: Predict the behavioral deviations of metamaterials under specific deformations through a machine - learning model to ensure the accuracy of the design. 2. **Optimize the Design Process**: Use deep learning to reduce the consumption of design time and computational resources, making the design process more efficient. 3. **Verify Theoretical Models**: Through finite - element analysis (FEA) and experimental verification, ensure the consistency between theoretical models and actual results. ### Solutions - **Data Generation**: Train a regression model using pseudo - random images and their corresponding calculated deformation results. - **Model Training**: Use a convolutional neural network (CNN) for training to predict the deviation between the deformation behavior and the ideal behavior. - **Experimental Verification**: Manufacture experimental samples by 3D printing and conduct experimental tests to verify the accuracy of the model predictions. ### Key Formulas - **Mean Absolute Error (MAE)**: \[ \text{MAE}=\frac{1}{n}\sum_{i = 1}^{n}|x_i-\hat{x}_i| \] where \(x_i\) is the actual deformation value and \(\hat{x}_i\) is the predicted deformation value. - **Mean Squared Error (MSE)**: \[ \text{MSE}=\frac{1}{n}\sum_{i = 1}^{n}(x_i-\hat{x}_i)^2 \] - **Gaussian Filter**: \[ G(x,y)=\frac{1}{2\pi\sigma^2}e^{-\frac{x^2 + y^2}{2\sigma^2}} \] where \(\sigma = 10\). Through the above methods, this research not only improves the efficiency of metamaterial design but also provides new ideas and technical means for future research.

Accelerating Auxetic Metamaterial Design with Deep Learning

Quasi-static and dynamic responses of gradient hexachiral auxetics: Experimental and numerical analysis

A Deep Learning Approach for Reverse Design of Gradient Mechanical Metamaterials

Obtaining auxetic and isotropic metamaterials in counterintuitive design spaces: an automated optimization approach and experimental characterization

Modeling and Design of Zero‐Stiffness Elastomer Springs Using Machine Learning

An autonomous design algorithm to experimentally realize three-dimensionally isotropic auxetic network structures without compromising density

3D printed auxetic metamaterials with tunable mechanical properties and morphological fitting abilities

Spatial tuning of the positive and negative Poisson's ratio of metamaterials through gradient design

Deep Learning in Mechanical Metamaterials: From Prediction and Generation to Inverse Design

A critical review on the application of machine learning in supporting auxetic metamaterial design

Auxetic Metamaterials for Biomedical Devices: Current Situation, Main Challenges, and Research Trends

Optimization of Graded Metamaterials for Control of Energy Transmission Using a Genetic Algorithm

FFT-based surrogate modeling of auxetic metamaterials with real-time prediction of effective elastic properties and swift inverse design

Machine learning guided design of mechanically efficient metamaterials with auxeticity

Design Study for Multifunctional 3D Re‐entrant Auxetics

3D Tiled Auxetic Metamaterial: A New Family of Mechanical Metamaterial with High Resilience and Mechanical Hysteresis

Continuous fiber reinforced meta-composites with tailorable Poisson's ratio and effective elastic modulus: Design and experiment

Machine learning analysis/optimization of auxetic performance of a polymeric meta-hybrid structure of re-entrant and meta-trichiral

Auxetic mechanical metamaterials: from soft to stiff

Design of 2D re-entrant auxetic lattice structures with extreme elastic mechanical properties

Design of multiphase auxetic metamaterials by a parametric color level set method