The generated network structures created by the software are constructed under the criterion that they hold the same percolation characteristics as those derived from experimental data. The experimental data is fitted using an iterative fitting of pore and throat sizes until a “closest fit” is found. The fitting process is undertaken using an annealed simplex algorithm, which works to find the global minima for a five dimensional surface.

 

To simulate mercury intrusion and liquid expulsion, a computational representation of fluid is applied to the top face (maximum z) of the unit cell only, and percolates in the –z direction. The throat skew, throat spread, pore skew, connectivity and short range size auto correlation are adjusted by the Boltzmann-annealed amoeboid simplex (Press and Teukolsky, 1991, Johnson et al., 2003) to give a close fit to the entire mercury intrusion curve or liquid expulsion curve. There are three additional Boolean constraints on the simplex: it rejects structures in which the network is fragmented, in which voids overlap, or which cannot be adjusted to give the experimental porosity without contradicting the experimental percolation characteristics.

 

The void network model can generate different types of structures and these structures are described on the subsequent pages. The different structure types together represent different arrangements of pores and throats found in natural and man-made porous materials. If you know your material that has been analysed contains distinct layers of different sizes of pores and throats, a vertically banded or horizontally banded structure would be more suitable than a bundle of capillaries with no interconnections between the capillaries.