A Novel Shrink–Expand–Shrink Method for Modeling Composites with Ultrahigh Volume Fractions of Pre-Graded and Gradient-Distributed Particles

Abstract

An advanced shrink–expand–shrink method is proposed in this paper for efficiently modeling concrete-like particle-reinforced composites with ultrahigh volume fraction of aggregates. The gradation of the aggregates can be pre-given and the aggregate spatial distribution can be nonuniform. By this method, the shrunk aggregates are first generated in the model space, and then expanded to jostle each other, afterwards they are shrunk again to normal size to obtain the final mesostructure models. Any high volume fractions of particles even more than 90% can be easily achieved by adjusting the shrinkage level during this process. Besides, a gradient distribution algorithm is established to conform to the aggregate segregation during the actual pouring process, and the corresponding periodic boundaries can also be generated to quickly create large specimens. Finally, the compression processes of polymer concrete with different aggregate packing densities are calculated via a finite element method. The shrink–expand–shrink method and the corresponding numerical models have memorable importance in the performance analysis and material design of particle-reinforced composites.