井震数据联合驱动下砂体叠置模式构建技术及应用——以WX油田东北部姚家组葡萄花油层为例  

Construction technology of superimposition patterns of sandbodies driven by well-seismic data and its application:Taking the Putaohua reservoir of Yaojia Formation in the northeastern WX Oilfield as an example

作　　者：徐世东 陈书平[1,2] 薛佳雯 孔令华 XU Shidong;CHEN Shuping;XUE Jiawen;KONG Linghua(State Key Laboratory of Petroleum Resources and Prospecting,China University of Petroleum(Beijing),Beijing 102249,China;College of Geosciences,China University of Petroleum(Beijing),Beijing 102249,China;Beijing Bright IP AgencyCo.,Ltd.,Beijing 102200,China;PetroChina Xinjiang Oilfield Company,Karamay,Xinjiang 834000,China)

机构地区：[1]中国石油大学(北京)油气资源与探测国家重点实验室,北京102249 [2]中国石油大学(北京)地球科学学院,北京102249 [3]北京布瑞知识产权代理有限公司,北京102200 [4]中国石油新疆油田分公司,新疆克拉玛依834000

出　　处：《石油地球物理勘探》2023年第1期178-189,共12页Oil Geophysical Prospecting

基　　金：国家自然科学基金项目“济阳坳陷新生界走滑构造及与油气关系”(42172138)资助。

摘　　要：葡萄花油层是松辽盆地WX油田重要的含油层系之一,该油层砂体层数多、厚度小(1~2 m);局部井间距较大,井间沉积相变快,砂体展布特征复杂,精细地刻画油层段各小层沉积微相特征难度大,目前砂体叠置模式尚未建立。为此,以测井、录井和三维地震数据为基础,采用优化的随机森林算法和数据挖掘技术建立沉积微相概率预测模型;再通过机器学习和模糊判别方法分别识别WX油田东北部姚家组葡萄花油层各小层的沉积微相类型;最终建立四种砂体叠置模式,即平面连接式、平面分隔式、垂向接触式和垂向分离式。通过钻井验证,该技术具较高的储层预测精度,对于河道沉积微相,样本井的准确率平均值可达88.8%,各小层(除PI11和PI3外)检验井预测准确率均可达80.0%以上。精细刻画的沉积微相和建立的砂体叠置模式可为后期储层综合评价、开发方案优化、调整和井位合理部署等提供依据。Putaohua reservoir is one of the important oil-bearing layers in the WX Oilfield of Songliao Basin.Sandbodies in the reservoir have a large number of layers and small thicknesses(1~2 m).In addition,fast change in the sedimentary facies between wells makes sandbody distribution characteristics complex,and well spacing is large in some areas.As a consequence,it is difficult to describe the sedimentary microfacies characteristics of each sublayer precisely by traditional technologies and methods.So far,the superimposition patterns of sandbodies have not been established,which brings difficulties to oilfield development.In response,utilizing well logging,mud logging,and 3D seismic data,this paper builds a probability prediction model of sedimentary microfacies with the optimized random forest algorithm and data mining technology driven by well-seismic data.Further,the sedimentary microfacies types of each sublayer of the Putaohua reservoir in the Yaojia Formation in the northeastern WX Oilfield are identified using machine learning and the fuzzy identification method.At last,the paper estab-lishes four superimposition patterns of sandbodies in this area:plane connection,plane separation,vertical connection,and vertical separation.The technology has high reservoir prediction accuracy according to drilling verification in the study area.For sedimentary microfacies in channels,the average accuracy of sample wells can reach 88.8%,and the prediction accuracy of test wells in each sublayer(except the PI11layer and the PI3 layer)is up to more than 80.0%.The finely described sedimentary microfacies and the established superimposition patterns of sandbodies can provide a basis for the comprehensive and effective evaluation of reservoirs,the optimization and adjustment of development programs,and the rational deployment of well locations.

关 键 词：葡萄花油层 井震联合驱动 机器学习 概率预测模型 沉积微相 砂体叠置模式 

分 类 号：P631[天文地球—地质矿产勘探]