Research On Comprehensive Explanation Method Of Horizontal Well Geosteering

The purpose of Horizontal Wells Geosteering is enhancing production of single well and recovery rate.Geosteering is based on the geological research on the establishment of predrilling geological model,however,due to the complex and varied formation,it is necessary to use the LWD、MWD and logging data to verify the structure and reservoir information in real time,revised geosteering models constantly,real-time predicting the depth of the landing point and the target,timely adjustment of the well trajectory.Therefore,this paper makes a comprehensive study on the whole process of horizontal well geological guidance by using LWD and logging.First,the pre-drilling reservoir structure model is studied.A two-dimensional reservoir boundary prediction model is established based on the offset data and the mathematical method of curve fitting,it is a binary function of the horizontal displacement and the reservoir boundary.This model is applicable to strata with a large range of ups and downs and can also be applied to horizontal wells through faults.The reservoir interface prediction model can be used to determine the inclination of the reservoir,calculate the viewing angle of the reservoir and determine the relative position of the bit and the reservoir boundary.The reservoir boundary prediction model is continually modified to improve the accuracy of the model based on real-time data before drilling the reservoir.The safe and accurate landing of the build-up section in the process of horizontal well geological guidance is a key step in drilling horizontal wells,so the location of the forecast landing point is the key to this part.This paper mainly studies the predicted landing point before drilling and the predicted landing point.The key to predicting the landing point in drilling is to compare the real-time identification marker layer with the adjacent well,therefore,based on CBR(case-based reasoning)theory for artificial intelligence identification marker layer research,that is by analyzing the similarity of marker layer characteristics of current well and offsets to determine whether the drilling marker layer.And the geologic model is continually updated before landing.In order to determine the position of the drill bit in the reservoir in real time during the drilling process of the horizontal segment,there is a certain distance between the tool of LWD and drill bit is considered,therefore,a drilling predictive model based on BP neural network is established,that is,using real-time logging data to predict the LWD data at the drill bit,and the predicted data were corrected to improve the accuracy,to ensure the maximum degree of drilling in the reservoir.The reservoir geologic model is updated with the real-time data of the drilled section to ensure the subsequent drilling rate.In this paper,the Horizontal Well Geosteering Comprehensive Analysis Software was developed by using VS 2013 and SQL Server 2008.The software can intelligently identify the logo layer,intelligent to determine whether the landing and predict while drilling,and used the data collected at the drilling site to verify its feasibility.