Application of multi-attribute neural network inversion in gravity flow reservoir prediction—a case study of Es1 Formation in Qinan Slope,Qikou Sag
Es1 Formation of Qinan Slope in Qikou Sag is a typical gravity channel deposit,which is affected by factors such as deep basin expansion and steep sedimentary slopes.The sand bodies in each layer are generally developed,but the vertical overlapping and horizontal changes are large.While the seismic attrib-ute analysis has poor ability and accuracy in distinguishing vertically overlapping sand bodies.The multi-attribute neural network inversion technique based on feature curve reconstruction first improves the ability to identify thin sandstone layers through frequency processing of stacked seismic data.Then,using time-fre-quency analysis techniques based on high-order statistics,a fifth-order sequence framework is construc-ted to refine the geological research object.Finally,the multi-attribute neural network inversion based on GR curve reconstruction is used for reservoir prediction.This technology makes the resolution of sand body overlapping area stronger,the boundary characterization clearer,and improves the accuracy and precision of inversion results.Practice has shown that the accuracy of reservoir prediction can reach over 80%,provi-ding a new technical approach for the prediction of sand body distribution with large vertical and horizontal changes.
gravity flowhigh order statisticsfifth-order sequencecurve reconstructionneural network