关键词: 深薄致密砂岩, 储层预测, 波阻抗反演, 地震属性, 深度学习, 渤南油田

Abstract: With the geological characteristics of large burial depth,thin single-layer and low reserves abundance,the quantitative description of reservoirs is a key constraint to the effective production of deep and thin tight sandstone reservoirs. Taking the Y 184 wellblock of Bonan Oilfield as an example,by developing high-resolution inversion technology of facies-controlled decompaction,the difference of sand and mud wave impedance in the middle and shallow layers is introduced into the deep layer to realize the effective distinction of sand and mudstone wave impedance and determine the development range of reservoirs. Fine description of reservoir physical properties is realized and favorable reservoir distribution range is clarified through a multi-seismic attribute fusion technology based on deep learning,fully exploitation of the nonlinear relationship between seismic attributes and reservoir parameters and new reservoir-sensitive attributes through fusion. Applied in the unproduced reserves of Y 184 wellblock of Bonan Oilfield,this set of technical sequences finely depicts the development boundary of Y 184 well fan delta and quantitatively describes the thickness and physical properties of reservoirs,solving the unclear understanding of the development boundary between fan delta and beach-bar sandstone in this block for a long time,and providing technical support for the block benefit construction and production.

Key words: deep and thin tight sandstone, reservoir prediction, wave impedance inversion, seismic property, deep learning, Bonan Oilfield