Efficient approaches for modeling and simulating the mechanical behavior of concrete using lattice discrete particle models

Abstract

Simulating the quasi-static mechanical behavior of concrete at the micro- or meso-scale, considering its heterogeneous nature, quickly becomes impractical in terms of computational cost. This manuscript explores efficient computational strategies in numerical modeling by means of the Lattice Discrete Particle Model (LDPM), a state-of-the-art approach for simulating concrete at the coarse aggregate level, emphasizing three interaction approaches. Whereas the original formulation of LDPM employs a 12-facet formulation, this research proposes a simplified interaction approach for LDPM, based on either 6-facet or edge-based interactions, designed to significantly reduce computational costs while maintaining precise predictions of the concrete fracture behavior. This approach is systematically applied to a variety of standard concrete tests, including unconfined compression, biaxial compression, triaxial compression, torsional-compressive, three-point bending, and cyclic compression loading in order to assess the predictive capabilities of the model. The efficiency and accuracy of the reduced number of interaction surfaces are critically discussed in both tensile and compressive loading conditions. The results indicate that approaches based on edge-based and 6-facet interactions substantially reduce computational costs and memory usage while providing similar results to the 12-facet model, except for unconfined compression simulations based on edge-based interaction. This research opens a promising avenue for advancing the utilization of LDPM in concrete mechanics simulations.