Permeability prediction of numerical reconstructed multiscale tight porous media using the representative elementary volume scale lattice Boltzmann method

Tight oil sandstones have obscured flow characteristics because of their complex pore structure and ultralow permeability. In this study, representative elementary volume (REV) scale lattice Boltzmann method (LBM) was applied to simulate fluid flow in the multiscale tight porous media generated by coupling digital cores from two scales, and the effects of clay permeability, clay volume fraction and interparticle pores (interP) volume fraction on the permeability of the multiscale tight porous media were analyzed. The tight multiscale porous media were generated by assigning the nanometer scale digital core of clay to each pixel of clay constituent in the micron scale digital cores containing three constituents of matrix, interP and clay. The porosity of the generated multiscale tight porous media ranges from 11.12% to 14.98%, among which the porosity contributed by the interP is 4.06-5.00%, and the porosity contributed by the clay is 6.40-9.98%. The pore size distribution curves are bimodal with the pore size ranging from 0.05 to 100 μm, and show good agreement with our previous experimental results. The porosity and permeability inputted to the REV scale LBM were from pore scale simulation results of fluid flow in the nanometer scale digital cores. The simulated permeability of the multiscale tight porous media is 0.05 × 10−3-0.23 × 10−3 μm2, which is close to our experimental results. The permeability of the multiscale tight porous media shows positive linear correlation with the clay permeability and clay volume fraction, but is unrelated to the interP volume fraction in the reasonable porosity range. It is hoped that this study can provide some new insights into the core analysis of tight oil sandstones.