分频智能融合属性在渤海P油田三角洲相储层预测中的应用OA

The Neogene Guantao Formation in Oilfield P of the Bohai Bay Basin is characterized by shallow-water braided river delta deposits,where sand bodies exhibit rapid lateral variations and substantial thickness heterogeneity.Constrained by the paucity of drilling data and the low dominant frequency(37 Hz)of seismic datasets,the correlation between seismic attributes and sand body thickness in the study area proves to be unsatisfactory.To address this challenge,an intelligent fusion method of frequency-divided seismic attributes was proposed to enhance the identification accuracy of sand body thickness,thereby underpinning fine-scale sand body characterization and optimal well placement.Aiming at the limited resolution of single-frequency seismic attributes,seismic data corresponding to high,medium,and low frequency bands were carefully selected to extract attributes with high correlation to sand body thickness.Subsequently,the support vector regression(SVR)algorithm was employed to establish a quantitative relationship between seismic attributes and sand body thickness,which further enabled the generation of fused attributes reflecting the spatial distribution of sand body thickness.Guided by the sedimentary model,fine-scale prediction of sand bodies was implemented based on the fused attributes.The results demonstrate that the accuracy of sand body prediction is improved,with the coincidence rate of thickness prediction for sand bodies exceeding 5 m reaching 91%.This method has facilitated fine-scale sand body characterization and optimal well deployment in the study area,and can serve as a reliable reference for fine reservoir prediction in analogous oilfields.