Application Practice of Intelligent Reservoir Prediction Research in Bohai L Oilfield

The prediction of oil and gas reservoirs, especially thin layer prediction, has always been one of the key and difficult points in oil and gas exploration and development. The development of thin interbedded reservoirs in the Bohai L oilfield urgently requires high-resolution reservoir prediction to meet the demand for precise exploration. On the basis of summarizing the principle of improving resolution through machine learning, the combination of earthquake and machine learning methods is applied to carry out reservoir prediction in this oilfield, fully exploring the subtle information hidden by earthquakes. Data samples are established through reverse engineering, tomographic velocity field building, seismic attribute extraction and optimization, and data structuring processing; Based on tag data extraction, machine learning algorithms are selected to establish a data-driven model and carry out intelligent reservoir prediction based on machine learning. The application practice of Bohai L Oilfield has shown that the machine learning reservoir prediction results have greatly improved the accuracy of main sand body characterization, interlayer and thin-layer sand body identification. After verification by new wells, the reservoir identification compliance rate exceeds 80%, which is consistent with production dynamics. The fine reservoir description based on this method can effectively guide the development and production work of the research area.