基于多属性的断溶体特征提取与分带性自动识别

断溶体作为顺北油气田的主要储集体类型,具有储层非均质性强,地震波场特征复杂的特点,如何精细描述断溶体空间形态和内部结构是断溶体油藏勘探开发的关键。基于二维经验模式分解(BEMD)和纹理分析方法开展了断溶体描述方法研究,首先采用BEMD信号分解技术对地震数据进行预处理,通过压制地层结构突出断溶体地震反射特征;然后利用优化梯度结构张量属性和纹理属性算法,提高断溶体空间形态检测精度;最后基于多种优势属性聚类分析,实现断溶体横向分带性的自动划分,形成了满足实际生产应用需求的断溶体描述方法技术流程。将该方法技术应用于顺北油气田断溶体识别和描述,得到的断溶体形态和分带性的认识与井资料吻合度较高。

In the oil and gas fields in Shunbei,reservoir type identification methods often return broken solutions,which are indicative of strong reservoir heterogeneity and complex seismic wavefield characteristics.The description of the spatial morphology and internal structure of these reservoirs is essential for their exploration and development.To this aim,bidimensional empirical mode decomposition (BEMD) and texture analysis were used in this study.In particular,BEMD was used to preprocess the seismic data and highlight the seismic reflection characteristics of the broken solution by suppressing the formation structure.Then,the detection accuracy of the spatial morphology of the solution was improved by optimizing the gradient structure tensor attribute and the texture attribute algorithm.Finally,by clustering a number of advantageous attributes,a horizontal zonation of the broken solution was automatically achieved.The proposed method was applied for the identification and description of broken-solution reservoirs in the study area,providing results that were in good agreement with well data.

国家重大专项(2016ZX05014-001)、中国石化油气重大项目(P17049)和中国石化科技部项目(P18040)共同资助。