砂岩敏感识别因子的建立及直接提取方法  

A method for establishing and directly extracting sensitive identification factors of unconsolidated sandstones

作　　者：高刚[1,2] 杨亚华 赵彬[1,2] 段宏亮[4] 王晓阳 魏亚斋 GAO Gang;YANG Yahua;ZHAO Bin;DUAN Hongliang;WANG Xiaoyang;WEI Yazhai(Key Laboratory of Exploration Technologies for Oil and Gas Resources(Yangtze University),Min­istry of Education,Wuhan.Hubei 4301009 China;College of Geophysics and Petroleum Resources,Yangtze University,Wuhan,Hubei 430100,China;School of Geosciences,China University of Pe­troleum(bast China),Qingdao,bhandong China;Jiangsu Oilfield Branch Co.,SINOPC,Yangzhou,Jiangsu 225009,China;Southwest Branch,BGP Inc.,CNPC,Chengdu,Sichuan 610213,China;Dagang Branch,GRI,BGP Inc.,CNPC,Tianjin 300280,China)

机构地区：[1]长江大学油气资源与勘探技术教育部重点实验室,湖北武汉430100 [2]长江大学地球物理与石油资源学院,湖北武汉430100 [3]中国石油大学(华东),山东青岛266580 [4]中国石化江苏油田分公司,江苏扬州225009 [5]东方地球物理公司西南物探分公司,四川成都610213 [6]东方地球物理公司研究院大港分院,天津300280

出　　处：《石油地球物理勘探》2019年第6期1329-1338,1347,I0012,共12页Oil Geophysical Prospecting

基　　金：国家科技重大专项“大型油气田及煤层气开发”(2017ZX05005003-009);国家自然科学基金项目“基于卷积神经网络的压裂微震实时监测参数优选研究”(41604099);中国石油科技创新基金项目“地震波频散AVOZ响应特征分析及其在储层流体识别中应用研究”(2015D-5006-0301)和“基于机器学习的地震解释技术在陆相油藏地质建模中的应用研究”(2018D-5007-0301);天津市企业博士后创新项目“相控约束的开发地震储层预测技术研究与应用”(TJQYBSH2018017);油气资源与勘探技术教育部重点实验室(长江大学)开放基金项目“基于机器学习的地震解释技术应用”(PI2018-07)联合资助

摘　　要：大量文献证明敏感识别因子已成功用于储层的岩性、流体、岩石脆性等的预测,但研究砂岩敏感识别因子的建立及直接提取方法相对较少。为此,根据黄珏地区浅层砂岩纵、横波速度相对较高、密度较低情况,设计砂岩敏感识别因子SF4,避免了大部分常规岩性识别因子(速度与密度乘积形式)预测的砂、泥岩数值重叠现象;为避免二次计算产生的累计误差,推导了包含砂岩敏感识别因子的AVO近似式,引入低频软约束项补偿反演的低频信息,结合Downton的参数协方差矩阵特征分解去相关思路,利用贝叶斯参数反演方法直接估算该因子。数值试验与实际应用均证明了砂岩敏感识别因子建立与直接提取方法效果较好。Many literatures prove that sensitive identification factors are successfully used to predict reservoir lithology,fluids,and rock brittleness.However,very few researches focus on the establishment of sensitive identification factors of reservoir and its direct extraction methods.We demonstrate this kind of research in shallow-sandstone reservoirs in Huangjue area.Firstly,a sensitive identification factor in sand reservoirs is designed for SF4 based on high P-wave velocity(vP),high S-wave velocity(vS),and low density(ρ),which effectively avoids numerical overlaps of conventional sand and mud identification factors(velocity multiplied by density).Then the AVO approximation containing the sensitive sand identification factor is derived which decreases accumulative errors caused by indirect calculations.Finally,the sensitive sand identification factor is directly extracted using the Bayesian-based elastic parameter inversion and the low frequency component of the inversion can be compensated by introducing the soft low-frequency constraint.Numerical tests and real data examples prove the validity and practicability of the proposed method.

关 键 词：直接提取 岩性识别 AVO近似式 贝叶斯反演 

分 类 号：P631[天文地球—地质矿产勘探]