ИСТИНА

There are three main approaches to model or reconstruct (using some limited information or properties) some porous media: 1) statistical methods (correlation functions and simulated annealing, multi-point statistics, entropy methods), 2) sequential methods(sphere or other granular packs), and 3) morphological methods. In this contribution we analyze all these methods and show that correlation functions have some substantial advantages. Main results and discussions include: 1) characterization and classification of soil 2D and 3D structure using different correlation functions (two-point probability, linear and cluster functions), 2) reconstruction ofsoil structure based on its correlation functions,3) reconstruction of 3D soil structure from 2D cut(s),4) reconstruction quality control using morphological and other statistical methods, 5) approaches to deal with soil anisotropy and hierarchical levels. All examples come from real soil samples obtained via conventional thin-sectioning or X-ray microtomography.