Sensitivity to material contrast and scaling measures in statistically-based homogeneization procedure for random composite materials

A statistically based homogenization procedure, previously developed by some of the authors [1‐3], is adopted. This procedure allow us to obtain the effective elastic properties of micropolar continua able to describe composite materials exhibiting an internal microstructure characterized by random distributions of particles inside a matrix. The so‐called Representative Volume Element (RVE) used to perform the homogenization procedure is obtained by increasing a scale factor representing the ratio between the size of a control window, referred as Statistical Volume Element (SVE), and the particle size until the statistical convergence is reached. In this article, an evaluation of the sensitivity to material contrast (ratio between inclusion and matrix moduli) in the identification of the classical and micropolar constitutive coefficients of the homogenized continuum is investigated, and the convergence trend of properly defined scaling‐measures is analyzed. To this aim a series of parametric analyses on two dimensional samples of composites with disk‐shaped inclusions are performed by varying both the inclusion density and the material contrast. The results obtained for different kind of composites ‐ ranging from metal or ceramic matrix composites up to concrete, masonry‐like and geo‐materials ‐ highlight the importance of taking into account the spatial randomness of inclusions in identifying the bulk, shear and bending behavior of composites as well as the effectiveness of the micropolar continuum modeling.