Study On Real-time Correction Method Of Reservoir Geological Model

Three-dimensional reservoir geological model is an important means to study reservoirs.In recent years,a lot of research work has been carried out to improve the accuracy of models and reduce the uncertainty of underground space.Random modeling,phased modeling and isochronous modeling have emerged,multi-step modeling and other methods improve the accuracy of the model.Based on the research of various geological modeling methods,integrating 3D geological modeling,logging while drilling,machine learning lithology identification and interpretation methods,horizontal well geosteering and horizontal well optimization,a set of real-time correction process methods for the geological model while drilling was established.Taking a certain block in the Weibei Oilfield as an example,based on the existing development characteristics and the existing geological understanding,an advanced geological model is established,and the well logging interpretation is performed according to the machine learning method,and the facies model of the study area is updated.As a geosteering model,it guides horizontal well drilling and improves the drilling rate of oil layers.The real-time correction method of the geological model while drilling can make full use of the new drilling feedback information,reduce the model uncertainty,improve the reservoir understanding,and make the established facies model gradually approach the underground real situation.The results and understanding of this paper are:1.Through the investigation of the geologic guidance technology while drilling and the real-time correction method of geological model,a set of real-time correction process of geological model based on the interpretation while drilling is established;2.Based on the reservoir geological characteristics area and the adjacent well data of the study,a multi-step modeling method is used to establish the advanced geological model.According to the advanced geological model,the horizontal well drilling trajectory is designed,and the horizontal well length optimization design is carried out by numerical simulation;3.Collect logging while drilling data and standardize it.Write machine learning lithology identification code in Python language,and use random forest algorithm to explain the rapid lithology of LWD data;4.On the basis of the lithology interpretation of the logging while drilling,the workflow is automatically updated,the rolling model of the facies is updated,and the pre-designed well trajectory is optimized according to the model correction results.