Characterization of Macro-Micro Heterogeneity in Shale Reservoirs Based on Full-Diameter Core CT Scanning─A Case Study of Jurassic Lianggaoshan Formation Shale in Well Ping’an 3, Sichuan Basin

Continental shale reservoirs exhibit strong vertical heterogeneity in lamellar fracture development and pore–fracture systems, yet continuous core-scale characterization remains limited. This study presents a hierarchical characterization of four Jurassic Lianggaoshan Formation intervals from Well Ping’an 3, Sichuan Basin, based on full-diameter core computed tomography (CT) scanning combined with convolutional neural network (CNN)-based image segmentation, 3D reconstruction, and pore-scale seepage simulation. The results show that (1) reservoir heterogeneity increases vertically from siltstone to shale intervals, with average lamina/interlayer volume percentages of 11.12%–12.81% and lamina densities of 130–2185 strips/m in shale intervals; (2) lamellar fractures predominantly develop along bedding interfaces, with average fracture porosity of 0.4%–0.7% and strong vertical heterogeneity; (3) within the CT resolution limit, pore–fracture seepage assemblages are classified into pore-dominated, transitional, and fracture-dominated types, with reservoir connectivity jointly influenced by fracture density, aperture, and matrix pore coupling. This study provides a continuous macro–micro integrated characterization workflow for micron-scale pore–fracture features at the CT-resolvable scale (>150 μm), bridging the gap between plug-scale microanalysis and interval-scale geological evaluation. The nanopore system requires complementary methods such as scanning electron microscopy (SEM) and mercury intrusion capillary pressure (MICP) for full characterization.