Synthesis of sand particles from 3D shape descriptors using tomographic reconstruction techniques

Shear behavior and liquefaction susceptibility of geomaterial aggregates have been determined to be dependent on particle shape and angularity. However, quantification of this dependence is a challenging task owing to a dearth of quantitative models for particle shape and the difficulty of modeling angular particle assemblies. The situation becomes more complex when the quantitative models are required to synthesize arbitrary 3-D particle shapes that are representative of specific sand mixtures. The authors have recently described a method for quantitative identification of 3-D particle morphology estimated from projective two-dimensional representations. This paper extends prior work and describes the design and development of an automated three-dimensional particle synthesis algorithm using tomographic reconstruction techniques. A database of 2-D and 3-D images has been generated by optical and X-ray scans of the following sands: #1 Dry Sand, Daytona Beach, Standard Melt, Rhode Island, Hawaii Kahala Beach, Michigan Dune and Hawaii Ala Wai Surfer's Beach. The algebraic reconstruction technique (ART) has been used to characterize 3-D particle shapes from this database. This paper demonstrates the consistency, separability, and uniqueness of the 3-D shape descriptor algorithm as well as its ability to synthesize 3-D particle shapes representative of the respective aggregate mixtures. Copyright ASCE 2006.