Random pore-network development methodology based on Voronoi and Delaunay tessellations for residual coal under axial stress.

• The particle size distribution law of broken coal bodies was calculated. • A random pore network development method for broken coal bodies is proposed. • The pore network was adjusted using permeability and the conductivity adjustment factor η i was verified to be general. Coal spontaneous combustion (CSC) in the gob areas is considered to be an important safety issue in the coal industries. Numerical simulation is a reasonable tool for predicting and preventing this problem. The commonly applied finite-volume and finite-element techniques are computationally expensive when they must account for pore-scale phenomena. A pore-network model (PNM) has the advantage of lower computational cost while also simulating heat and mass transfers in porous media at scale with reasonable accuracy. This study provides a method for extending pore networks to packed coal particles with a wide spectrum of size distributions, and to generate pore networks of arbitrary size within a range. This method was developed based on the Voronoi and Delaunay tessellations theories. Statistical data on the size distribution, porosity, and permeability of packed coal particles were collected experimentally. This method requires a relatively low network generation cost for the preservation of the voids that are formed among larger particles and characterize the effect of fine particles filling in throat channels by reference to a conductivity adjustment factor η i , which depends on axial stress. The average relative error of predicted and experimental permeability does not exceed 5% when this new pore-network construction procedure is used. The predicted result holds until the volume upscaling ratio reaches 58. [ABSTRACT FROM AUTHOR]