A data-driven framework for structure-property correlation in ordered and disordered cellular metamaterials

Area of Science:

• Materials Science
• Mechanical Engineering
• Computational Mechanics

Background:

• Understanding the relationship between microstructure and macroscopic properties is crucial for designing advanced cellular metamaterials.
• Current methods often lack the ability to deeply connect specific morphological features to material performance.

Purpose of the Study:

• To develop a unified framework for predicting macroscopic properties of cellular metamaterials.
• To reveal the connection between key morphological characteristics and material properties using machine learning.
• To identify critical microstructural features influencing material behavior.

Main Methods:

• Integration of machine learning models with interpretability algorithms.
• Analysis of strut orientation and its impact on effective stiffness.
• Examination of shear moduli and mean cell compactness.
• Refinement of Maxwell's criteria for frame structure rigidity.

Main Results:

• The framework successfully predicts macroscopic properties and links them to morphological features.
• Strut orientation is identified as critical for stiffness in specific microstructures, leading to counterintuitive material behavior.
• Mean cell compactness emerges as a key feature for predicting shear moduli.
• A refined version of Maxwell's criteria is provided for cellular metamaterials.

Conclusions:

• The proposed framework offers a powerful tool for understanding and designing cellular metamaterials.
• Key morphological features like strut orientation and cell compactness significantly dictate material properties.
• The framework's generality allows for extension to other architected materials and properties.