Three-Dimensional Characterization of Polydisperse Particulate Composites from Microtomography

This work contributes to a larger effort to analyze the mechanical and transport properties of polydisperse particulate composites. In order to improve the understanding of the morphology of such systems, specimens are analyzed from tomographic images. The material is first characterized by idealized shapes, and the representative pack is analyzed using high-order statistics in order to gain knowledge about the microstructure of the pack. Using this information, the effort proceeds by either constructing a representative unit cell RUC on which numerical analysis can be effectively performed, or performing computations using variational methods based on the statistical data to determine properties of interest. In this work, surrogate packs of rice, mustard, salt, and mixtures of rice and mustard are analyzed. Specimens are scanned using X-ray microtomography. A three-dimensional voxel pack is then reconstructed from the tomographic images. The voxel pack is analyzed in the image analysis software, Amira, to identify individual particles and their centroids, volumes, surface areas, and inertia tensors. Next, an algorithm is developed which reduces the particles of the voxel pack to ellipsoids or cuboids. Next, an algorithm is developed which reduces the particles of the voxel pack to ellipsoids or cuboids. An objective function is used to choose characteristic lengths of the idealized shapes to match the surface area and volume of the corresponding voxel particles.