准噶尔盆地永进油田超深层致密砂岩裂缝测井识别  

Fracture Identification in Super-Deeply Buried Tight Sandstone Reservoirs in Yongjin Oilfield,Junggar Basin,China

作　　者：刘浩杰[1] 盖姗姗 于文政 刘洪平 张昌民[2] 张莉[2] LIU Haojie;GAI Shanshan;YU Wenzheng;LIU Hongping;ZHANG Changmin;ZHANG Li(Geophysical Research Institute,SINOPEC Shengli Oilfield,Dongying,Shandong 257015,China;School of Geosciences,Yangtze University,Wuhan,Hubei 430100,China)

机构地区：[1]中国石化胜利油田分公司物探研究院,山东东营257015 [2]长江大学地球科学学院,湖北武汉430100

出　　处：《测井技术》2022年第5期592-598,共7页Well Logging Technology

基　　金：中国石油化工股份有限公司科研攻关项目“低渗致密油藏地震工程一体化关键技术研究”(P20069-1)。

摘　　要：裂缝是影响永进油田超深层致密砂岩油气层产能的重要因素,但多种类型裂缝的发育导致其测井识别难度大。在研究区裂缝系统描述的基础上,划分裂缝类型,系统分析不同类型裂缝的常规测井特征,采用曲线重构放大裂缝信号,综合利用小波分解和支持向量机回归法开展裂缝识别。结果显示,研究区以水平缝为主,少见垂直缝和高角度斜交缝。裂缝发育段常规测井响应不明显,基于重构曲线的小波分解能够识别薄层水平缝、垂直缝和高角度斜交缝,但对于厚层水平缝发育段识别效果较差,结合支持向量机回归法能够同时识别大段水平缝以及薄层裂缝。该技术思路可有效识别研究区裂缝发育段。Fracture is a key factor that controls the petroleum production of the super-deeply buried tight sandstone reservoirs.However,it is difficult to identify fractures using conventional well logs due to the variable fracture types.Fractures characteristics are described systematically and the fractures are divided into different types.The conventional well logging responses of the different fracture types are analyzed systematically and the logging curve reconstruction is done to enlarge the well log response of fractures.A method combining the wavelet analysis and support vector machine(SVM)is used for the fracture identification.The result shows that low dip-angle fracture is the dominant type,and vertical or high-dip angle fractures are rare.The conventional logs are not sensitive to the fractures.Wavelet analysis can be used to identify vertical fractures,high dip-angle fractures and thin vertical sections with low dip-angle fractures.However,the wavelet analysis shows low identification accuracy for the thick vertical sections with low dip-angle fractures.A combination of SVM and wavelet analysis can be used to identify fractures in the study area.

关 键 词：测井识别 永进油田 水平缝 小波分解 支持向量机 裂缝识别 

分 类 号：P631.84[天文地球—地质矿产勘探]