Microstructure reconstruction of inhomogeneous materials using elastic waves and deep learning

Accurate characterization of microstructures in inhomogeneous materials is crucial for understanding and predicting their macroscopic mechanical behavior. In this work, we propose a deep learning-based inverse framework for reconstructing complex and realistic microstructures of inhomogeneous materials directly from elastic wave responses. Unlike the conventional machine learning classifiers employed in previous studies, the proposed framework uses advanced deep neural networks capable of capturing highly nonlinear relationships between elastic wave propagation and heterogeneous microstructural features. The elastic wave response is used as the model input, while the corresponding spatial microstructure is predicted as the output. Numerical simulations based on finite element wave propagation are conducted to generate comprehensive training and testing datasets with increased material complexity and spatial heterogeneity. The results demonstrate that the proposed deep learning model can accurately and robustly reconstruct realistic microstructures from elastic wave data, significantly outperforming traditional machine learning approaches in terms of prediction accuracy and generalizability. This study highlights the potential of elastic-wave-driven deep learning as a powerful nondestructive inverse tool for microstructure characterization in complex engineered inhomogeneous materials.