利用决策树方法识别复杂碳酸盐岩岩性——以苏里格气田苏东41-33区块为例

发布时间：2018-05-06 12:08

本文选题：复杂碳酸盐岩 + 测井岩性识别　；参考：《油气地质与采收率》2017年06期

【摘要】：苏里格气田苏东41-33区块具有低孔、低渗透、非均质性强的特点,其奥陶系马家沟组碳酸盐岩储层受多期不同类型构造、沉积等作用,岩性复杂多样,对岩性的准确识别成为研究区开发的重点和难点。近年来,决策树方法作为一种机器学习方法,在地质学领域中的运用越来越受到关注,尤其在岩性预测方面。以测井、录井资料为基础,通过岩性参数特征分析,优选出对岩性较敏感的声波时差、自然伽马、光电吸收截面指数、密度、深侧向电阻率和补偿中子6种测井参数,通过分析这6种测井参数特征,构建基于决策树方法的多分类器,将岩性信息与岩石特征信息融合。与测井、录井资料的对比、分析结果表明,利用决策树方法识别复杂碳酸盐岩岩性的准确率超过80%,且相比朴素贝叶斯方法,其岩性识别的准确率提高了13%。
[Abstract]:The Suitou 41-33 block in the Sulige gas field is characterized by low porosity, low permeability and strong heterogeneity. The carbonate reservoir of the Ordovician Majiagou formation is affected by many different types of structures, deposits and so on, and the lithology is complex and diverse. The accurate identification of lithology has become the key and difficult point in the development of the study area. In recent years, the decision tree method has been used as a machine to learn. Methods, more and more attention has been paid to the application of geology in the field of geology, especially in lithologic prediction. Based on logging data and logging data, the acoustic wave time difference, natural gamma, photoelectric absorption cross section index, density, deep lateral resistivity and compensated neutrons are selected to select 6 logging parameters, which are more sensitive to lithology. The characteristics of these 6 logging parameters are analyzed. The multi classifier based on the decision tree method is constructed, and the lithologic information is fused with the rock feature information. The comparison between the logging data and the logging data shows that the accuracy rate of identifying the lithology of the complex carbonate rocks by the decision tree method is more than 80%, and the accuracy rate of the lithology identification is compared with the naive Bayes method. Improved 13%.

【作者单位】：中国地质大学(北京)数理学院;北京中地润德石油科技有限公司;
【分类号】：P618.13

【相似文献】