石油学报 ›› 2010, Vol. 31 ›› Issue (3): 445-448.DOI: 10.7623/syxb201003016

• 油田开发 • 上一篇    下一篇

基于储层静态地质参数的一种随机建模结果排序方法

李少华 1,2   张昌民 1   Clayton V.Deutsch 2   

  1. 1长江大学地球科学学院  湖北荆州  434023; 2阿尔伯塔大学计算地质统计学中心  加拿大埃德蒙顿  T6G2W2
  • 收稿日期:2009-07-10 修回日期:2009-10-24 出版日期:2010-05-25 发布日期:2010-07-28
  • 通讯作者: 李少华
  • 作者简介:李少华,男,1972年8月生,2003年获中国石油勘探开发研究院博士学位,现为长江大学地球科学学院副教授,主要从事地质统计学与储层建模方面的科研与教学工作。
  • 基金资助:

    国家科技重大专项(2008ZX0511-003)和湖北省高等学校优秀中青年科技创新团队项目(T200602)联合资助。

Ranking realizations of stochastic modeling based on reservoir static geological parameters

LI Shaohua 1,2   ZHANG Changmin 1   Clayton V. Deutsch 2   

  • Received:2009-07-10 Revised:2009-10-24 Online:2010-05-25 Published:2010-07-28

摘要:

储层随机建模可以产生大量实现,而油藏数值模拟由于计算量大,通常只能对非常有限的实现进行模拟计算。随机选择实现常常会导致过于悲观或过于乐观的模拟结果,并且无法合理评价地质模型所包含的不确定性。利用与数模响应密切相关的静态地质指标,如连通原油体积对实现进行排序,然后选择典型的实现进行数模,较好地解决这一问题。但是,该指标没有考虑含油网格与生产井之间的距离及渗透率大小,而这两个参数对于流体流动有着重要的控制作用。在连通原油体积的基础上提出了一种新的排序指标。该指标在计算连通原油体积的公式上增加了网格与生产井之间的距离和渗透率两个影响因子。通过实例对这两种指标进行了对比研究,新指标更好地反映了不同实现的差别。

关键词: 油藏数值模拟, 随机建模, 连通原油体积, 流体流动, 静态地质参数, 排序指标

Abstract:

Reservoir stochastic modeling can produce a large number of realizations with modern modeling software. However, the flow simulation of all the realizations is challenging because of computational costs associated with the multiple detailed flow simulations. Only a limited number of realizations are often selected for the flow simulation. Selecting the first realization or randomly selecting them may lead to unusual low or high results and cannot permit an assessment of uncertainty. Ranking realizations according to some simple geological measures and then selecting some typical ones are good way for solving this problem. The simple measures should be highly correlated to the flow response variables of interest. Connected hydrocarbon volume in place (CHV) is a kind of such measure and has been applied in some study cases.  A new static ranking measure was proposed on the basis of CHV. In this measure, the calculation formula for connected hydrocarbon volume includes two affecting factors such as the distance from each cell to the nearest connected production well and the geometric average of permeability from the cell to the nearest production well in the shortest path. The two factors were investigated by the study case. The new measure shows the difference of different realizations.

Key words: reservoir numerical simulation, stochastic modeling, connected hydrocarbon volume, flow fluid, static geological parameter, ranking factor