基于高分辨率阵列感应测井的GA-SVM流体识别方法  

Fluid identification method based on high-resolution array induction logging using GA-SVM

作　　者：刘丹[1] 潘保芝[2] 周玉凤[1] 王宇博 黄溯航[1] LIU Dan;PAN Bao-zhi;ZHOU Yu-feng;WANG Yu-bo;HUANG Su-hang(Beijing Urban Construction Exploration and Surveying Design Research Institute Co. , Ltd. , Beijing 100101, China;College of Geo-exploration Science and Technology, Jilin University, Changchun 130026, China)

机构地区：[1]北京城建勘测设计研究院有限责任公司,北京100101 [2]吉林大学地球探测科学与技术学院,长春130026

出　　处：《地球物理学进展》2017年第5期2051-2056,共6页Progress in Geophysics

基　　金：国家科技重大专项<大型油气田及煤层气开发>鄂尔多斯盆地大型低渗透岩性地层油气藏开发示范工程(2011ZX05044)资助

摘　　要：苏里格致密砂岩气藏岩性致密,具有非均质性强、微观孔隙结构复杂、低阻气层和高低阻水层并存的特点.其测井响应关系复杂,微电极电阻率曲线在低渗透层上无明显幅度差,常规测井技术系列和解释方法已呈现出新的不适性.本文综合分析苏里格致密砂岩气藏不同层位下高分辨率阵列感应测井(HDIL)电阻率曲线的差异,应用累积法建立电阻率曲线总差异参数.另外,在无阵列感应测井条件下,通过有高分辨率阵列感应测井(HDIL)数据的井点建立与常规测井之间的关系,进而可以使得常规测井也具备高分辨率阵列感应测井的优势.并在此基础上,将基于遗传算法GA优化的SVM算法(GASVM)引入到测井解释当中,建立了流体识别的非线性模型,解决了常规测井技术系列和解释方法在致密砂岩储层上无法准确识别流体的问题.The tight sandstone gas reservoir in Sulige is characterized by tight lithology,strong heterogeneity,complicated micro pore structure,coexistence of low resistivity gas reservoir and high and low resistivity water layer. The logging response relation is complicated,and the resistivity curve of the microelectrode has no positive amplitude difference in the low permeability layer,and the conventional logging technique series and interpretation method have presented new discomfort. In this paper,the difference of resistivity curve of high resolution array induction logging（ HDIL） in Sulige gas reservoir under different layers is analyzed comprehensively.The cumulative difference method is used to establish the total difference parameter of resistivity curves. In addition, in the absence of array induction logging conditions, through a high resolution array induction logging（ HDIL） data of wells,and to establish a relationship between conventional logging,which can make the conventional logging also has the advantages of high resolution array induction logging. On this basis,the GA-SVM algorithm which is a SVM algorithm optimized by genetic algorithm GA is introduced into the log interpretation, and the nonlinear model of fluid identification is established. It solves the problem that the conventional logging technology and interpretation methods can not accurately identify fluid in tight sandstone reservoir.

关 键 词：致密砂岩 高分辨率阵列感应测井 电阻率曲线总差异参数 GA-SVM算法 流体识别 

分 类 号：P631[天文地球—地质矿产勘探]