Petrofacies types and evolution characteristics of tight sandstone gas reservoirs of Jurassic Shaximiao Formation, Tianfu gas area, central Sichuan Basin

The classification of rock types is important for characterizing the heterogeneity of tight sandstone reservoirs and revealing their differential evolution process. The Middle Jurassic Shaximiao Formation in the Tianfu gas area of the central Sichuan Basin was selected as the study object, and petrofacies types and diagenetic evolution process were investigated using thin-section observation, scanning electron microscopy (SEM), and stable carbon and oxygen isotope analysis. The gas-bearing reservoirs of the Shaximiao Formation in the Tianfu gas area exhibit strong heterogeneity in physical properties, pore types and distribution. Based on the differences in petrographic composition and texture, diagenetic patterns and processes, and pore structure characteristics, four types of petrofacies were classified: ductile-lean sandstone, ductile-rich sandstone, tightly calcite-cemented sandstone, and tightly laumontite-cemented sandstone. During the reservoir evolution process, ductile-lean sandstone experienced moderate compaction, active fluid-rock interactions, and strong dissolution, and generally underwent multiple stages of dissolution and cementation, resulting in effective reservoir petrofacies. Ductile-rich sandstone underwent strong mechanical compaction, became dense in the early diagenetic stage, and exhibited weak fluid activity and slight dissolution in the late stage. In tightly calcite- and laumontite-cemented sandstone, calcite and laumontite were interlocked, resulting in early-stage densification and late-stage weak fluid modification. The original composition and texture of sediments controlled the differences in the influencing patterns and degrees of diagenesis within reservoirs. The concept of petrofacies provides a systematic framework for identifying key petrographic parameters that influence reservoir porosity and permeability. In combination with well logging data, this approach can effectively guide reservoir property modeling.