• Digital rock and DEM were combined to analyze shale pore structure evolution under coupled stress–seepage conditions. • Macropores show the highest pressure sensitivity, dominating pore structure evolution under loading. • NMR T₂ measurements validate numerical evolution trends of different pore-size pores. • An improved NMR-based permeability model incorporating pore-size-segmented pressure sensitivity is proposed. • Field application demonstrates improved permeability prediction accuracy compared with the conventional SDR model. Shale reservoirs are characterized by strong heterogeneity and complex pore systems with wideranging size distributions, and these structures undergo pronounced evolution under coupled mechanical loading and fluid flow. In this study, we developed an integrated framework combining CT-based digital rock reconstruction, Discrete Element Method (DEM) simulations, and Nuclear Magnetic Resonance (NMR) validation to investigate these pore-scale responses. The novelty of this work lies in leveraging the particlescale capabilities of DEM to elucidate the mechanical mechanisms behind pore-structure evolution. Our results indicate that higher initial porosity leads to more pronounced pore structure deterioration and enhanced heterogeneity under stress perturbations. Specifically, through particle-scale mechanical analysis, we observed that macropores exhibit the highest pressure sensitivity and are prone to significant compression or structural rearrangement, whereas smaller pores remain relatively stable. The simulated pore-size responses show strong consistency with NMR T 2 measurements across different pore-pressure levels. Building upon these insights, we proposed an improved NMR-based permeability formulation incorporating pore-size-segmented pressure sensitivity. Validation against both laboratory and engineering field data confirms that the proposed model significantly improves prediction accuracy in stress-sensitive shale reservoirs compared to conventional unified models. Overall, this research provides a physics-consistent basis for understanding permeability degradation and supports more reliable evaluation in unconventional reservoirs.
Wang et al. (Sun,) studied this question.