Automated Missed Pay Zones Detection Method Based on BV10 MemberData of Samotlorskoe Field

摘要

Nowadays in the oil and gas industry more and more tasks are related to the analysis of big volume ofdata.Successful solutions to this type of problem are based on machine learning algorithms.The aim of thisstudy is to create an automated model for predicting pay zones based on the analysis of well log data.Theobject of research are sedimentary deposits of the BV10 object of the Samotlorskoe oil and gas condensatefield.In this project an approach for the gradual interpretation of lithotypes,prediction of the saturationtype for potential reservoirs and also the prediction of porosity and permeability was chosen.The articlediscusses the application of three machine-learning algorithms:the support vector machine method,theneural network and the random forest algorithm.The existing problems of the studied data are describedas well as methods for solving them and the stages of data preparation.As a result of the analysis it wasdecided to use a neural network to create a model for predicting pay intervals.Additionally,the accuracyof the model predictions was estimated from two wells excluded from the analysis.