油气储层勘探建模技术新进展及未来展望  

New progress and future prospects of oil and gas reservoir modeling technology for exploration

作　　者：罗红梅[1] 王长江[1] 张志敬[1] 房亮 管晓燕[1] 郑文召 LUO Hongmei;WANG Changjiang;ZHANG Zhijing;FANG Liang;GUAN Xiaoyan;ZHENG Wenzhao(Exploration and Development Research Institute,Shengli Oilfield Company,SINOPEC,Dongying City,Shandong Province,257015,China)

机构地区：[1]中国石化胜利油田分公司勘探开发研究院,山东东营257015

出　　处：《油气地质与采收率》2024年第4期135-153,共19页Petroleum Geology and Recovery Efficiency

基　　金：中国石化科技攻关项目“地质模式约束的非均质储层精细刻画”(P22161)。

摘　　要：油气储层建模利用地质统计学等方法,综合测井、地质、地震等多学科信息,是油气田开发研究的利器,油藏地质模型可以将油藏各种地质特征在三维空间的变化及分布定量表征出来,是油气藏的类型、几何形态、规模、油藏内部结构、储层参数及流体分布的高度概括,储层地质模型是油藏地质模型的核心,可以对储层的沉积特征、非均质性、物性及流体等特征进行综合表征。但在勘探阶段,面对大尺度沉积体系和稀疏井网条件下的储层展布规律表征的建模难点为:①地质知识的量化表达问题,包括地质专家的经验认识如何数字化表征。②稀疏井网条件下无法直接用钻井资料对地质体的发育规模、展布方向和结构特征准确定量描述及构建地质模式,大尺度空间中复杂沉积体系无法用简单数学函数表征。③传统地质统计学等方法在勘探模型构建中如何实现地震、测井、地质、油藏等多维度数据的融合问题。因此,基于确定性建模和传统地质统计学等随机建模的储层建模理论和技术遇到极大挑战。笔者在系统剖析传统储层建模技术流程和方法的基础上,通过构建涵盖地质、测井、地震、分析化验等信息的多学科地学大数据知识库,开展多维数据凝聚层次聚类的沉积相模式库表征和基于生成式网络的智能建模,提出了多学科协同的油气储层勘探建模技术对策及技术体系,实现了构造、沉积及储层之间匹配关系的定量表征。该技术体系在东营凹陷北部陡坡带、洼陷带勘探部署中开展系统应用,构建融合古地貌、古物源、搬运通道、测井及地震属性等多信息的岩相、物性及油气运聚的地质模型,基于模型新范式指导部署井位,支撑了陆相断陷盆地复杂砂砾岩体、页岩油等勘探实践。笔者通过深度剖析东营凹陷北部陡坡带勘探建模实践难点及精度问题,进一步探讨�Oil and gas reservoir modeling integrates multidisciplinary information from logging,geology,and seismic data through geostatistics and other methods,and it is a powerful tool for oil and gas field development.The geological model of reservoirs can quantitatively characterize the variation and distribution of various geological features of the reservoir in three-dimensional space and is a high generalization of the type,geometry,scale,internal structure of the oil and gas reservoir,reservoir parameters,and fluid distribution.The geological model of reservoirs is the core of the geological model of oil and gas reservoirs,which can com‐prehensively characterize the sedimentary characteristics,heterogeneity,physical properties,and fluid characteristics.However,it is difficult to characterize reservoir distribution under the condition of a large-scale sedimentary system and sparse well pattern in the exploration stage,covering①quantitative representation of geological knowledge,including how to represent the experience of geological experts digitally;②it is impossible to accurately and quantitatively describe the development scale,distribution direc‐tion,and structural characteristics of geological bodies directly with logging data and construct geological models under the condi‐tion of sparse well pattern,and complex sedimentary systems in large-scale space cannot be characterized by simple mathematical functions;③traditional geostatistics and other methods can not realize the fusion of seismic,logging,geological,reservoir,and other multi-dimensional data in the construction of exploration model.Therefore,the theory and technology of reservoir modeling based on deterministic modeling and stochastic modeling,such as traditional geostatistics,have met great challenges.On the basis of systematic analysis of traditional reservoir modeling technology processes and methods,the authors constructed a multidisci‐plinary big data knowledge base of geoscience covering geology,logging,seismic,analytical,and labora

关 键 词：储层勘探建模 地学大数据知识库 相模式库 生成对抗网络 智能建模 

分 类 号：TE19[石油与天然气工程—油气勘探]