基于FCM算法的多属性分析技术在河道砂体精细刻画中的应用——以西湖凹陷T气田为例  

Application of multi-attribute analysis technology based on FCM algorithm in fine characterization of sedimentary microfacies:take T Gas Field in Xihu Sag as an example

作　　者：王凯 刘东成[3] 刘华峰 黄德榕 储飞跃 WANG Kai;LIU Dongcheng;LIU Huafeng;HUANG Derong;CHU Feiyue(School of Geosciences,China University of Petroleum(East China),Qingdao 266580,China;Shandong Provincial Key Laboratory of Reservoir Geology,Qingdao 266580,China;Exploration and Development Research Institute of PetroChina Dagang Oilfield,Tianjin 300280,China)

机构地区：[1]中国石油大学(华东)地球科学与技术学院,青岛266580 [2]山东省油藏地质重点实验室,青岛266580 [3]中国石油大港油田勘探开发研究院,天津300280

出　　处：《海洋地质前沿》2023年第9期55-67,共13页Marine Geology Frontiers

基　　金：校企合作项目“西湖凹陷中南部区域地震沉积学研究”(CCL2020SHPS022RSI);国家自然科学基金“河流辫-曲转换机制、沉积响应及其地震反射模式”(42172144)。

摘　　要：西湖凹陷T气田经过十多年的勘探与开发,亟需在主力层花港组内寻找潜力目标。该区为浅水三角洲沉积体系,岩性组合在空间上变化快,为了精确识别河道砂体及其边界,在海上少井条件下利用三维地震资料识别并刻画河道砂体。在等时地层划分的基础上,对目的层段进行岩石物理性质分析,通过地震沉积学的技术方法结合岩芯及测井等资料,对沉积微相做出初步判断,在此基础上提取6类48种地震属性,与砂厚及各属性之间进行相关性分析,对地震属性进行优选,将优选出的3种反映地质体边界、岩性较好的地震属性采用基于模糊C-均值(FCM)算法的多属性聚类分析,以达到数据降维、减少冗余的效果,研究分流河道沉积体系的整体展布规律。再进行多属性RGB融合显示,增强河道砂体边界的刻画,结合构造特征以及预测的砂体厚度综合分析,提出有利目标区,为后续油田滚动开发及井位部署提供依据。After 10+years of exploration and development of the T Gas Field in the Xihu Sag in East China Sea,it is urgent to find potential targets in the Oligocene Huagang Formation.To identify channel sand body and its boundary accurately,3D seismic data were used based on data of several offshore wells.The petrophysical properties of the target stratum were analyzed based on isochronous stratigraphic division,and the sedimentary microfacies were preliminarily specified by seismic sedimentology and logging data.In total,48 seismic attributes were extracted in six categories.Correlation analysis was conducted on sand thickness and each attribute,from which three most meaningful seismic attributes were determined and selected.The three seismic attributes could well reflect the geological body boundaries and ideal lithology.In addition,multi-attribute clustering analysis based on the FCM(Fuzzy C-Means clustering)algorithm was performed,by which the effect of data dimension and redundancy were reduced.In addition,the distribution of distributary channel depositional system was studied,and RGB fusion display was carried out to highlight channel sand body boundaries.By combining the comprehensive analysis on geological structure and predicted sand body thickness,the favorable target area was proposed,which provides a basis for the subsequent rolling development of oilfield and the well location deployment.

关 键 词：地震属性 模糊C-均值算法 多属性聚类 砂体预测 花港组 

分 类 号：P744.4[天文地球—海洋科学] P736