Abstract
A modelling framework for the prediction of effective properties in random packings of particles is presented. Random packings of spheres and agglomerates of spheres are numerically generated by using packing algorithms. Effective properties of both the types of packings are evaluated through a Monte-Carlo random-walk (a.k.a. mean square displacement) method, which allows the calculation of both geometrical parameters (e.g., pore size distribution, specific surface area) and transport properties (e.g., effective gas diffusivity, permeability). The results are reported as a function of porosity in dimensionless form, in order to obtain scale-independent information. The effective properties obtained for random packings of spheres are compared with independent experimental data showing a satisfactory agreement. Effective properties of packings of agglomerates are also evaluated, showing that particle agglomeration significantly increases the mean pore size while reducing the effective gas diffusivity and the specific surface area. The results show that agglomerates can not be generally assimilated to spheres with an equivalent diameter. The modelling approach presented in this study may improve the quantitative characterization of porous media composed by aggregates of particles.
