VOX model: application of voxel-based packing algorithm on cementitious composites with 3D irregular-shape particles

Abstract

The internal structure of composite materials is usually composed of a continuous matrix with dispersed particles. To simulate the spatial distribution of particles, the random sequential addition algorithm (RSA) is commonly used. The RSA using continuous geometric function is difficult to achieve a high packing density in allowable computer time and time-consuming in the case of irregularly shaped aggregates, such as in mortar and concrete systems. In this work, a geometrical model, named as VOX, is introduced to deal with 3D irregularly shaped particles. The model applies the voxel-based overlap determination in the RSA to improve the efficiency and packing density. Also, the model is applicable to the complex boundary conditions of any geometry and provides high-quality triangle meshes valuable for finite element analysis. The mortar and concrete cubes with real-shape particles are simulated, in which the packing density is increased to 65% and 51%, respectively. The VOX model successfully packs over 150,000 real-shape aggregates into the steel cage in a full-size reinforced concrete beam model on a desktop machine. Particle shape significantly affects the packing density in the RSA and the sphericity is recommended as an indicator. The particles with higher sphericity can achieve denser packing under the same effort of attempts. Real-shape gravel used in this study behaves similarly with regular octahedron because of the similar sphericity.