No abstract

The diagenesis and porosity evolution of the Middle Jurassic Shaximiao sandstones were analyzed based on petrographic observations, X-ray diffractometry, scanning electron microscopy observations, carbon and oxygen stable isotope geochemistry, fluid inclusion microthermometry, and thermal and burial history modeling results. The point count data show that secondary pores (av. 5.5 %) are more abundant than primary pores (av. 3.7 %) and are thus the dominant pore type in the Shaximiao sandstones. Analysis of porosity evolution indicates that alteration of sandstones mainly occurred during two paragenetic stages. Mechanical compaction and cementation by early chlorite, calcite, and quartz typically decrease the depositional porosity (40.9 %) by an average of 37.2 %, leaving porosity of 3.7 % after stage I (\85°C, 175-145 Ma). The original intergranular porosity loss due to compaction is calculated to be 29.3 %, suggesting that mechanical compaction is the most significant diagenetic process in primary porosity destruction. Stage II can be further divided into two sub-stages (Stage II a and Stage II b ). is characterized by late dissolution, which enhanced porosity by 8.8 %, and the porosity increased from 3.7 % to 12.5 %. During stage II b ([120°C,, the precipitation of late chlorite, calcite, quartz, and kaolinite destroyed 3.3 % porosity, leaving porosity of 9.2 % in the rock today.

The tight sandstones of the Upper Triassic Xujiahe Formation (T 3 x) constitute important gas reservoirs in western Sichuan. The Xujiahe sandstones are characterized by low to very low porosity (av. 5.22% and 3.62% for the T 3 x 4 and T 3 x 2 sandstones, respectively), extremely low permeability (av. 0.060 mD and 0.058 mD for the T 3 x 4 and T 3 x 2 sandstones, respectively), strong heterogeneity, micronano pore throat, and poor pore throat sorting. As a result of complex pore structure and the occurrence of fractures, weak correlations exist between petrophysical properties and pore throat size, demonstrating that porosity or pore throat size alone does not serve as a good permeability predictor. Much improved correlations can be obtained between permeability and porosity when pore throat radii are incorporated. Correlations between porosity, permeability, and pore throat radii corresponding to different saturations of mercury were established, showing that the pore throat radius at 20% mercury saturation (R 20 ) is the best permeability predictor. Multivariate regression analysis and artifi cial neural network (ANN) methods were used to establish permeability prediction models and the unique characteristics of neural networks enable them to be more successful in predicting permeability than the multivariate regression model. In addition, four petrophysical rock types can be identifi ed based on the distributions of R 20 , each exhibiting distinct petrophysical properties and corresponding to different fl ow units.