Abstract:
In the Gongshanmiao area of the Sichuan Basin, the channel sandstones of the Second Member of the Lianggaoshan Formation are severely affected by the strong shielding effect of overlying thick lacustrine shales, which heavily suppresses the weak seismic reflection signals. Consequently, conventional seismic attributes and reservoir inversion methods fail to effectively characterize these channel sandstone reservoirs, posing a critical challenge to tight oil exploration and development in this area. To address this issue, this paper proposes a quantitative prediction method for subtle channel sandstone reservoirs by integrating Adaptive Variational Mode Decomposition (AVMD) and two-stage facies-controlled inversion. Based on far-angle stack attributes, the method employs AVMD to extract dominant frequency attributes, achieving precise decoupling of strong shale reflections from weak channel sandstone signals. A two-stage facies-controlled inversion workflow consisting of sedimentary-facies and lithofacies constraints is established to perform joint inversion of gamma-ray (GR) and P-impedance (IMP) in different facies zones, enabling fine characterization and quantitative prediction of subtle channel sandstone reservoirs. Application to real data shows that the prediction coincidence rate for sand bodies reaches 82.6%. The findings of this study have supported the design of 9 pilot test well schemes. The first deployed Well X001-H1 achieved a sandstone drilling ratio of 100% and a reservoir drilling ratio of 62.67%, demonstrating remarkable exploration and development effectiveness. This method can be extended to reservoir prediction under strong shielding backgrounds such as shale, limestone, and coal seams, demonstrating favorable practical value and application potential.