Image-based Characterization and Discrete Element Modeling of Granular Materials

This seminar presents an integrated framework for characterizing and modeling granular materials through advanced imaging techniques, discrete element method (DEM) simulations, and machine learning (ML) applications. First, we introduce nano-focus X-ray computed tomography (CT) to investigate particle morphology and soil properties, enabling high-resolution 3D visualization and quantitative analysis of granular microstructures. Experimental results from CT scanning reveal correlations between particle geometry, packing behavior, and macroscopic mechanical responses. Next, we explore DEM modeling to simulate soil behaviors at the particle scale. A novel one-to-one modeling approach is developed to replicate real granular assemblies captured from CT imaging, while breakage analysis of irregularly shaped particles provides insights into failure mechanisms under dynamic loading. Finally, we integrate machine learning to bridge micro- and macro-scale phenomena. Particle tracking algorithms are employed to monitor deformation patterns, and ML models are trained to predict mechanical behaviors using microstructural descriptors. By synergizing image-based characterization, DEM simulations, and data-driven techniques, this work advances the understanding of granular materials and offers practical tools for optimizing geotechnical designs.