A sequential workflow using fast Fourier transform, Weibull-based updating and lattice Boltzmann method for pore-structure and permeability evolution in digital rocks under hydrostatic strain

Mechanical loading alters rock permeability through changes in pore geometry, but voxel-based digital-rock models lack a direct route from local mechanical fields to an updated pore network. This study develops a one-way sequential workflow integrating fast Fourier transform (FFT)-based homogenization, Weibull-based pore-structure updating, and the lattice Boltzmann method (LBM) for digital rocks under prescribed hydrostatic strain loading. FFT-based homogenization first resolves the local stress and strain fields on the initial pore–solid geometry. Candidate interface voxels are ranked by mean compressive stress for closure and maximum principal stress for opening, while Weibull-type cumulative functions prescribe the number of voxel updates along selected porosity-evolution paths. LBM then evaluates the permeability of each updated pore structure. The implementation describes a dry-skeleton, strain-controlled problem: updated geometries are not fed back to the mechanical solver, and pore-pressure/effective-stress coupling is not included. For a 255 3 -voxel Berea sandstone subvolume, the calculated fields localize near pore–matrix interfaces and narrow solid ligaments, producing spatially selective pore updates. The normalized permeability decreases to 0.8314 at ɛ 0 = − 0 . 0 1 and increases to 1.2199 at ɛ 0 = 0 . 0 1 . Comparison with a normalized Kozeny–Carman-type relation gives mean absolute percentage errors of 0.64% for compression and 1.17% for tension. The remaining differences reflect three-dimensional pore-throat and connectivity changes that cannot be represented by porosity alone. The workflow therefore links prescribed hydrostatic strain, mechanically ranked pore updates, and permeability without claiming iterative hydromechanical coupling or direct simulation of damage and fracture.