石油学报 ›› 2020, Vol. 41 ›› Issue (11): 1406-1415.DOI: 10.7623/syxb202011010

• 油田开发 • 上一篇    下一篇

基于荧光薄片的剩余油赋存形态分类方法

高文彬1,2, 李宜强1,2, 何书梅3, 潘登1,2, 刘明熹1,2, 管错4   

  1. 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)石油工程学院 北京 102249;
    3. 中国石油大港油田公司勘探开发研究院 天津 300280;
    4. 中海油研究总院有限责任公司 北京 100028
  • 收稿日期:2020-05-14 修回日期:2020-08-10 出版日期:2020-11-25 发布日期:2020-12-11
  • 通讯作者: 李宜强,男,1972年1月生,1993年获大庆石油学院学士学位,2006年获中国科学院渗流流体力学研究所博士学位,现为中国石油大学(北京)研究员、博士生导师,主要从事提高采收率、油气渗流物理模拟实验等方面的研究工作。Email:lyq89731007@163.com
  • 作者简介:高文彬,男,1992年9月生,2014年获长安大学学士学位,现为中国石油大学(北京)博士研究生,主要从事化学驱提高采收率方面的研究工作。Email:irvinggao0702@163.com
  • 基金资助:

    国家科技重大专项"海上油田化学驱油技术"(2016ZX05025-003-010)资助。

Classification method of occurrence mode of remaining oil based on fluorescence thin sections

Gao Wenbin1,2, Li Yiqiang1,2, He Shumei3, Pan Deng1,2, Liu Mingxi1,2, Guan Cuo4   

  1. 1. State Key Laboratory of Petroleum Resources and Prospecting, China University of Petroleum, Beijing 102249, China;
    2. College of Petroleum Engineering, China University of Petroleum, Beijing 102249, China;
    3. Research Institute of Exploration and Development, PetroChina Dagang Oilfield Company, Tianjin 300280, China;
    4. CNOOC Research Institute Co., Ltd., Beijing 100028, China
  • Received:2020-05-14 Revised:2020-08-10 Online:2020-11-25 Published:2020-12-11

摘要:

随着油藏开发进入高含水阶段,剩余油分布愈加复杂,掌握孔隙内剩余油赋存特征是油田深度开发及提高采收率的基础和保证。荧光薄片法是微观剩余油研究中的基本实验手段,但以往荧光薄片的量化分析多依靠人工经验。通过结合彩色图像分割、分水岭颗粒分割、支持向量机和分类树算法,实现了荧光薄片的剩余油自动分类,从而提高了工作效率。按照荧光颜色,将剩余油划分为弱波及、中波及和强波及3个级别,将剩余油划分为簇状、孔表薄膜状、狭缝状、角隅状和粒间吸附状5种类型,并通过权重赋值的方式,将各种类型微观剩余油的相对比例与含油饱和度之间建立关系。将该分类方法应用于大港油田高含水砂岩油藏取心样品的分析结果表明:簇状剩余油是高含水砂岩油藏的主要类型,不同类型剩余油占比排序规律为簇状 > 孔表薄膜状 > 角隅状 > 粒间吸附状 > 狭缝状,该方法可以为后续的油田开发方案部署提供一定依据。

关键词: 荧光薄片, 微观剩余油, 分类识别, 赋存形态, 图像处理

Abstract:

As reservoir development enters the high water cut stage, the distribution of remaining oil becomes more and more complicated. Recognizing the occurrence characteristics of remaining oil in the pores provides the basis and guarantee for the deep development and enhanced oil recovery of oilfields. The fluorescent thin section method is the basic experimental method in study of microscopic remaining oil, but in the past, the quantitative analysis of fluorescent thin sections mostly relied on artificial experience. By combining color image segmentation, watershed granulometry segmentation, support vector machine and classification tree algorithm, this paper realizes the automatic classification of remaining oil of fluorescent thin sections, thereby improving work efficiency. According to the fluorescence color, the remaining oil is divided into three levels, i.e.weak swept, medium swept and strong swept, and into five types,i.e.cluster-shaped, pore surface film-like, slit-like, cube corner-shaped and intergranular adsorption-like. Additionally, a method of weight assignment is proposed to establish a relationship between the relative proportions of various types of microscopic remaining oil and oil saturation. The results of the above classification method applied to the sampling samples of high water-cut sandstone reservoirs in the Dagang oilfield show that the cluster-shaped remaining oil is the main type of high water cut sandstone reservoirs. The order of the proportions for different types of remaining oil is cluster-shaped > pore surface film-like > cube corner-shaped > intergranular adsorption-like > slit-like, which provides a basis for the subsequent deployment of oilfield development plans.

Key words: fluorescent thin section, microscopic remaining oil, classification and recognition, occurrence mode, image processing

中图分类号: