3D meso-scale modeling of concrete with a local background grid method

Abstract

Concrete is a three-phase heterogeneous composite material composed of aggregate, cement mortar and its bonding interface. The complex meso-structure of concrete has a direct influence on its macroscopic mechanical properties. In this paper, a new efficient method of concrete meso-modeling is proposed based on the local background grid method. The random polyhedral aggregate is generated according to aggregate gradation and dropped directly into the background mesh one by one. The process of concrete meso-modeling is simplified compared with the traditional meso-scale model of concrete. According to the aggregate shape and its spatial position, the newly placed aggregate is encapsulated by a bounding box, in which the identification of concrete meso-components and intrusion detection of new and old aggregates are carried out. By transforming aggregate intrusion detection during the process of concrete meso-geometric modeling into overlap check of aggregate elements in the local background grid, a large number of disjoint conditions between new and old aggregates are avoided. Thus a large amount of global calculation is greatly reduced, and the efficiency of concrete meso-modeling is obviously improved. The effects of aggregate element content, element mesh size and aggregate particle size distribution on concrete meso-modeling are analyzed. Finally, the reliability and validity of the concrete meso-model are verified by numerical simulation of concrete uniaxial compression and projectile penetration into concrete.