Automatic parameter identification of discrete mesoscale models with application to the coarse-grained simulation of reinforced concrete structures

In this study, the analysis of reinforced concrete structures is performed by adopting a coarse-grained approximation of a recently developed meso-scale model for concrete, the Lattice Discrete Particle Model (LDPM). The coarse-grained system is obtained by scaling the particle size with a "coarsening factor" > 1. The coarse system cannot reproduce all aspects of the mechanics of the fine system, consequently, the coarse-scale model parameters need to be calibrated from the behavior of the fine-scale system by matching some aspects, e.g. the macroscopic stress-strain/load- displacement curves under different loading conditions, of the fine scale response. The coarse-scale parameters are optimized by the least-square minimization procedure. The effectiveness of the procedure is verified by comparing full fine-scale solutions and coarse-grained solutions for cases of interest in which the full meso-scale problem can be actually solved.