石油与天然气地质 ›› 2020, Vol. 41 ›› Issue (4): 852-861.doi: 10.11743/ogg20200417

• 方法技术 • 上一篇    下一篇

基于CT扫描图像的碳酸盐岩油藏孔隙分类方法

廉培庆1(), 高文彬2, 汤翔2, 段太忠1, 王付勇3, 赵华伟1, 李宜强2   

  1. 1. 中国石化 石油勘探开发研究院, 北京 100083
    2. 中国石油大学(北京) 石油工程学院, 北京 102249
    3. 中国石油大学(北京) 非常规油气科学技术研究院, 北京 102249
  • 收稿日期:2018-05-30 出版日期:2020-08-01 发布日期:2020-08-11
  • 第一作者简介:廉培庆(1983-),男,高级工程师,油气田开发。E-mail:lianpq.syky@sinopec.com
  • 基金项目:
    国家自然科学基金青年科学基金项目(41702359);国家科技重大专项(2016ZX05033-003)

Workflow for pore-type classification of carbonate reservoirs based on CT scanned images

Peiqing Lian1(), Wenbin Gao2, Xiang Tang2, Taizhong Duan1, Fuyong Wang3, Huawei Zhao1, Yiqiang Li2   

  1. 1. Petroleum Exploration and Production Research Institute, SINOPEC, Beijing 100083, China
    2. College of Petroleum Engineering, China University of Petroleum(Beijing), Beijing 102249, China
    3. The Unconventional Oil and Gas Institute, China University of Petroleum(Beijing), Beijing 102249, China
  • Received:2018-05-30 Online:2020-08-01 Published:2020-08-11

摘要:

碳酸盐岩油藏具有复杂的储集空间和油气渗流特征,定量描述油藏中孔隙、裂缝、孔洞等储集空间的大小、形状及连通性难度较大。提出了一种基于扫描图像判断碳酸盐岩孔隙类型方法,可定量表征孔隙参数,并对岩心样品进行自动分类。该方法首先对碳酸盐岩的岩心扫描图像进行灰度转换和提高信噪比的预处理,然后对图像进行分割,区分出孔隙区域与基质区域。在此基础上,通过形态学处理和特征参数计算等步骤提取出孔隙特征参数,根据特征参数建立特征向量,采用支持向量机方法对CT图像中的孔隙、孔洞和裂缝进行自动识别并分类。在对岩心所有截面孔隙识别的基础上,提出了判断岩心孔隙类型的分类指数。T油田M油藏和F油藏应用结果表明:该方法识别精度较高,有效确定了油藏中占主导地位的孔隙类型,对油田有效开发具有一定的指导意义。

关键词: CT扫描图像, 图像分割, 支持向量机, 分类指数, 孔隙类型, 碳酸盐岩储层, 油田开发

Abstract:

Carbonate reservoirs are characterized by storage space of high heterogeneity and complex seepage mechanism.It is difficult to quantitatively characterize the size, shape and connectivity of pores, fractures, and vugs in the carbonate reservoirs.In this study, a new workflow is proposed to identify pore types, quantitatively characterize pore structure, and automatically classify core plugs based on CT scanned images.The CT scanned images of carbonates were pre-processed to perform gray conversion and increase signal-to-noise ratio, and further segmented to binary images containing pores and matrix.Then, the characteristic pore structure parameters were extracted through a series of procedures including morphology analysis and parameter calculation.The support vector machine method was adopted to automatically identify pores, vugs, and fractures in the CT images.Based on the pore identification of all cross-sections through CT images of the core, the coefficient was proposed to identify pore types.Application of the proposed workflow to the M and F reservoirs in the T oil field indicates that the workflow shows excellent accuracy in pore identification, can effectively determin the major pore types in reservoirs, thus is of guiding significance to effective petroleum development.

Key words: CT scanned image, image segmentation, support vector machine, classification index, pore-type, carbonate reservoir, oilfield development

中图分类号: