Use of a void network model to correlate porosity, mercury porosimetry, thin section, absolute permeability, and NMR relaxation time data for sandstone rocks

Abstract

The Pore-Cor void network model is used to construct stochastic realizations of the void structures of five sandstone samples of varying lithography. A close match was achieved to experimental porosity and mercury intrusion curves. The samples were resin impregnated and the fragments of voids revealed in thin sections photographed by backscatter electron microscopy at two magnifications. The sizes of these pore fragments matched those derived from a simulated microtoming of the network model much more closely than the sizes derived from the traditional capillary bundle approximation. Absolute permeabilities of the network were calculated by finding the flow capacity of the entire flow network, based on parametrized Navier Stokes equations with Klinkenberg correction, applied to each pore-throat-pore arc. A match to the experimental trend was obtained, although the network model considerably underestimated the experimental values. The results were also compared with the semiempirical equations of Thomson et al. and Kozeny and Carmen modified to accept thin section image analysis. Finally, the simulated pore and throat size distributions were compared to proton NMR transverse (T2) spin-echo relaxation times. Although the shapes of the distributions differed markedly, the mean values trended together. The capillary bundle approximation, however, gave a poor match to the NMR data.