基于支持向量机对夹层定量识别研究  

Study on Quantitative Recognition of Interlayer Based on Support Vector Machine

作　　者：洪伟俊 徐守余[1,2] 杨茜 Hong Weijun;Xu Shouyu;Yang Qian(School of Geosciences,China University of Petroleum(East China),Qingdao 266580,China;Evalution and Detection Technology Laboratory of Marine Mineral Resources,Qingdao 266071,China)

机构地区：[1]中国石油大学(华东)地球科学与技术学院,山东青岛266580 [2]海洋国家实验室海洋矿场资源评价与探测技术功能实验室,山东青岛266071

出　　处：《甘肃科学学报》2019年第4期7-13,共7页Journal of Gansu Sciences

基　　金：国家科技重大专项(2017ZX05009001)

摘　　要：夹层描述与预测为高含水油田剩余油研究的重要技术手段之一。经过数十年的开发,中国东部老油田基本已全面进入高含水、高采出阶段,但实践表明由于夹层分割的作用,仍存在可观的剩余油相对富集区。支持向量机是建立在统计学理论基础上的一种小样本统计学习理论,能够处理模式识别(分类问题、判别分析)及回归问题等诸多问题。以喇嘛甸油田南中西二区SⅢ4-7小层为研究对象,在夹层成因的基础上,结合测井资料对夹层类型进行识别,选取识别参数及识别标准建立样本集,从而建立支持向量机评价模型。利用评价模型对SⅢ4-7小层的夹层进行定量识别研究,并与实际结果对比,最后对模型进一步优化。Interlayer description is one of the important technical methods to study remaining oil in the predicted high water cut oil fields.After exploration for decades,the old oil field in east of China has been high water content,high recovery,but it still has some residual oil relative enrichment area by interlayer segmentation.Support Vector Machine is a small sample statistical learning theory based on statistical theory which can deal with mode recognition (classification and discriminant analysis) and regression problem.Take single layer in SⅢ4-7,middle west-2 district,south Lamadian oil south field as sample,based on interlayer causing,distinguish its type with logging information,select identification parameters and identification criteria to build sample set so as to build evaluation model supporting SVM.Use the model to study interlayer of SⅢ4-7 single layer,comparing to the real results and optimizing the model.

关 键 词：喇嘛甸油田 支持向量机 夹层 定量识别 评价模型 

分 类 号：TE122[石油与天然气工程—油气勘探]