论文详情
渤海新近系河-湖过渡带复合河道砂体叠置关系复杂、相带变化剧烈, 利用传统地震属性划分沉积微相多解性较强。提出了改进振幅谱距离-K中心聚类方法并将其应用于复合河道砂体沉积微相预测。方法包括两个方面: ①以测井标定沉积微相对应的典型波形作为初始聚类中心以及聚类中心更新过程中的约束条件; ②利用振幅谱距离评价波形之间的差异, 进一步提升波形聚类精度。根据渤海新近系A油田地质条件设计复合河道砂体三维模型, 模型试验结果证实改进后的方法微相预测准确率达95%, 相比传统K均值聚类算法精度提升15%, 较好地区分了泥岩、单期河道边部、单期河道主体、河道边部叠置、河道边部与主体叠置、河道主体叠置(或多期叠置)共6种河道不同部位或叠置样式。利用改进后的方法指导A油田10个砂体沉积微相划分, 将大型砂体S沉积微相划分为分流河道、河口坝、席状砂、决口河道、分流间湾、天然堤6类, 指导开发方案中优先动用分流河道、河口坝等优势微相的地质储量, 证明了方法的实用价值。
Owing to complex Neogene channel sands stacking and abrupt change of sedimentary facies in the river-lake transition zone in Bohai, microfacies classification using seismic attributes may be quite uncertain. We propose an improved amplitude-spectrum-distance K-center clustering method to mitigate the uncertainties. Microfacies calibrated using log data are set as the initial clustering center and constraints of clustering, and amplitude-spectrum distance is used to measure the differences among seismic wave forms for microfacies prediction. In view of the geologic conditions in Field A, Bohai Bay, a 3D geologic model is built with superposed channel sands. The model test shows a prediction accuracy of 95%, 15% higher than that of the K-mean clustering in the time domain. Six channel styles, i.e. mudstones, single-phase channel edge, single-phase channel body, superposed channel edges, superposed channel edge and body, and superposed channel bodies (or multi-phase superposed channels), could be identified, based on which Neogene channels are classified into 10 microfacies. A large reservoir, S, is classified into 6 microfacies: distributary channel, mouth bar, sand sheet, crevasse channel, inter-distributary bay, and natural levee, among which distributary channel and mouth bar are the promising microfacies for production. The case study demonstrates the feasibility of the improved classification method.