低阻油层综合分类及支持向量机评价——以Muglad盆地Fula凹陷AG1层为例  

Comprehensive categorization and appraise with support vector machine for low resistivity pays——A case of AG1 formation in Fula Sag Muglad Basin, Sudan

作　　者：韩如冰[1] 

机构地区：[1]中国石油勘探开发研究院

出　　处：《国外测井技术》2016年第6期10-14,3,共5页World Well Logging Technology

基　　金：国家科技重大专项"复杂油气藏精细表征与剩余油分布预测"(2011ZX05009-003)。国家科技重大专项"高含水油田提高采收率新技术"(2011ZX05010)

摘　　要：Muglad盆地Fula凹陷内大部分油田进入开发中后期,低阻油层成为重要的剩余资源。低阻油层预计的采油强度和累产油量等是在油井措施实施中的关键参数,直接决定经济效益。本文对凹陷内4个主力油田AG1层已生产的51层低阻油层的储层特征、渗流特征和产能特征进行分析,并以采油强度和累产油量为核心,将低阻油层分为三类。然后利用支持向量机对低阻油层所属类型进行判别,研究其对低阻油层类型及采油强度和累产油量等指标的预测能力。结果表明,以密度、声波时差、中子孔隙度、地层深电阻率、自然伽马、自然电位幅度比等作为模型变量的支持向量机分类器检验样本判别准确率为75%,可对低阻油层采油强度和累产油等进行一定程度的预测。Most of the oilfields in Fula Sag of Muglad Basin have entered mid-late development stage and low resistivity pays become important residual resources.The estimated oil production per unit thickness and accumulative production volume are very critical parameters in the plan and implementation of workover. The reservoir characteristics, percolation characteristics and production characteristics of 51 perforated low resistivity pays in AG1 formation in 4 main oilfields of Fula Sag were studied and the low resistivity pays were categorized into 3 types. Then the types of low resistivity pays were studied with support vector machine, to test its ability of differentiate different low resistivity pays types and their production ability. The results showed that the support vector machine model with density, acoustic transit time, neutral porosity, deep resistivity, natural gamma ray, spontaneous potential being used as model variables has the accuracy as high as 75% in the study of test data. The model could estimate the production parameters to some extent.

关 键 词：低阻油层 综合分类 采油强度 累产油 支持向量机 

分 类 号：P618.130.2[天文地球—矿床学]