Intelligent Classification of Sedimentary Microfacies in Delta-Front Tight Sandstone

Tight sandstone reservoirs in delta-front settings represent a critical target for hydrocarbon exploration and development, where the identification and classification of sedimentary microfacies serve as the foundation for detailed reservoir characterization. This study focuses on the tight sandstone reservoirs within the lower sub-member of the third member of the Shahejie Formation (lower Es3) in the Linnan Sag, Jiyang Depression, Bohai Bay Basin. A novel intelligent sedimentary microfacies classification model was established to subdivide the delta-front subfacies into four microfacies: interdistributary bays, subaqueous distributary channels, mouth bars, and sheet sands. By integrating six primary well-log parameters (CAL, GR, DEN, CN, AC, RD) and eight geologically sensitive transformed parameters, the model addresses imbalanced sample distribution and complex feature coupling through the Random Forest algorithm and SMOTE oversampling technique. Validation across four wells demonstrated classification accuracy exceeding 90%. This study enriches the exploration and development framework for delta-front tight sandstone reservoirs and provides a robust methodology for sedimentary microfacies delineation using well-log data.