石油学报 ›› 2013, Vol. 34 ›› Issue (1): 133-139.DOI: 10.7623/syxb201301016

• 油田开发 • 上一篇    下一篇

基于空间矢量的点坝砂体储层构型建模

李宇鹏 1  吴胜和 1  耿丽慧 1  Mehran M.Hasanpour 2  印森林 1  陈玉琨 1   

  1. 1.中国石油大学地球科学学院 北京 102249; 2.University of Alberta Edmonton T6G 1H7
  • 收稿日期:2012-05-12 修回日期:2012-08-09 出版日期:2013-01-25 发布日期:2013-04-09
  • 通讯作者: 李宇鹏
  • 作者简介:李宇鹏,男,1976年8月生,2011年获阿尔伯塔大学博士学位,现为中国石油大学(北京)讲师,主要从事储层建模、地质统计学方面的教学和科研工作。
  • 基金资助:

    国家重大科技专项(2011ZX05009-003)资助。

Spatial-vector-based reservoir architecture modeling of point-bar sand

LI Yupeng 1  WU Shenghe 1  GENG Lihui 1  Mehran M.Hasanpour 2  YIN Senlin 1  CHEN Yukun 1   

  • Received:2012-05-12 Revised:2012-08-09 Online:2013-01-25 Published:2013-04-09

摘要:

提出了一种基于空间矢量的曲流河点坝砂体构型建模方法。传统基于目标的方法是在建模的初始阶段定义网格,而基于空间矢量的储层构型建模方法直接在研究区内投放空间矢量定义的构型要素。该方法借鉴了计算机图像学中矢量图形存储的思路,采用了基于实数集的点、线、面所限定的体来定义构型要素。由于在模拟阶段模型不定义网格,因此模型不受网格尺寸的限制,模型模拟结果可以刻画不同规模层次的构型要素。并且由于在模拟过程中只是对构型要素的空间位置和形态定义参数进行调节,与传统的基于目标的模拟相比,更易满足井的条件约束,模拟收敛速度更快。笔者以中国东部油田某区块点坝砂体构型建模为例,给出了本方法的基本原理和具体的建模流程。

关键词: 储层构型, 随机模拟, 矢量存储, 目标模拟, 曲流河, 点坝, 侧积层

Abstract:

The spatial-vector-based stochastic modeling of point-bar reservoir architectures is proposed in this paper. Compared with traditional object-based modeling approaches, this proposed modeling approach has no cell or grid definition at the simulation stage. Architecture elements are simulated directly by dropping them into the modeling domain. Borrowing the idea from the vector image format in computer graphics, we used spatial lines, points and surfaces to form simulated architecture element bodies. The architecture elements are defined with spatial vectors which are expressed using their spatial distribution parameters. And the shape parameters are in real data set. Thus, different scale of heterogeneities can be reproduced from the realization. Because the well data conditioning is obtained through the modification of the architecture elements’ spatial locations and their shape parameters, it is easier to be satisfied. And the simulation has a faster convergent speed compared with traditional object-based approaches because of grid free and parameter adjustment in simulation. A developed block of one oilfield in eastern China, which has a geological background of meandering-river point-bar sand, is used to illustrate the theory and modeling procedures of this approach. 

Key words: reservoir architecture, stochastic simulation, spatial-vector-based storage, object-based simulation, meandering river, point bar, lateral accretion layer