Characterization of the wall effect of concrete via random packing of polydispersed superball-shaped aggregates

Abstract

On the mesoscopic scale, concrete can be treated as a two-phase composite consisting of aggregates and mortar matrix. The geometrical feature of aggregate particles usually has significant impact on the structural evolution of concrete and then affects the macro-mechanical properties. To explore the relation between the aggregate characteristics and the macroscopic properties of material, a variety of packing models of particles such as spheres and ellipsoids have been widely applied in previous studies. In the paper, a family of polyhedron-like particles which are defined as the superball is applied to represent the aggregate particles. Based on the Monte Carlo simulation, the models of concrete mesostructure are generated in a random manner. By combining the semi-periodic models of these particles with the section analysis technique, three parameters (i.e., the volume fraction of aggregates VV, the specific surface area of aggregates SV and the mean free spacing between aggregates λ at the cross-section of the models) can be obtained. On the basis of different particle packing systems, we further study the effect of the aggregate characteristics (e.g., particle size distribution, particle shape, particle packing volume fraction, etc.) on the variation of the interfacial layers (which are caused by the wall effect) near the boundaries of formwork and a series of meaningful results are successively obtained. Based on the simulated results in the work, a more comprehensive insight into the phenomenon of the wall effect in concrete can be gained. It is hoped that the conclusions may provide a helpful guidance for the design of the composite materials.