石油学报 ›› 2023, Vol. 44 ›› Issue (4): 626-635.DOI: 10.7623/syxb202304005

• 地质勘探 • 上一篇    下一篇

基于特征点的测井曲线对比拾取方法及其在自由水界面确定中的运用

邓亚1, 李勇1, 张文旗1, 胡丹丹1, 肖聪2, 许家铖1, 刘达望1, 张磊夫1, 王宇宁1   

  1. 1. 中国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
  • 收稿日期:2022-04-29 修回日期:2022-08-18 出版日期:2023-04-25 发布日期:2023-05-05
  • 通讯作者: 邓亚,男,1990年8月生,2016年获中国石油勘探开发研究院硕士学位,现为中国石油勘探开发研究院工程师,主要从事碳酸盐岩油藏综合地质研究以及储层三维地质建模工作。Email:dengya@petrochina.com.cn
  • 作者简介:邓亚,男,1990年8月生,2016年获中国石油勘探开发研究院硕士学位,现为中国石油勘探开发研究院工程师,主要从事碳酸盐岩油藏综合地质研究以及储层三维地质建模工作。Email:dengya@petrochina.com.cn
  • 基金资助:
    中国石油天然气股份有限公司"十四五"前瞻性基础性科技重大项目(2022DJ3209)资助。 第一作者及

Well log correlation and extraction method based on feature points and its application on determining free water level

Deng Ya1, Li Yong1, Zhang Wenqi1, Hu Dandan1, Xiao Cong2, Xu Jiacheng1, Liu Dawang1, Zhang Leifu1, Wang Yuning1   

  1. 1. PetroChina Research Institute of Petroleum Exploration and Development, Beijing 100083, China;
    2. College of Petroleum Engineering, China University of Petroleum, Beijing 102249, China
  • Received:2022-04-29 Revised:2022-08-18 Online:2023-04-25 Published:2023-05-05

摘要: 为方便准确地确定自由水界面的变化规律,提出了一种基于特征点的测井曲线对比拾取方法和基于该方法的自由水界面确定模型及算法流程。首先以测井曲线为对象,通过分析寻找测井曲线特征点,并对比追踪井间地层,得到一系列具有相同或者相似岩石物性的数据集合。然后联合J函数和阿尔奇公式,推导相同特征点下自由水界面海拔深度、距离自由水界面深度与电阻率之间的理论数学模型,并同时建立了水平和倾斜两种自由水界面分布模式。在理论模型的基础上,提出了以回归方程判定系数为优化目标的算法,通过预设实验数据以及添加高斯白噪音方法验证了算法的合理性和鲁棒性,最后使用该方法确定了伊拉克生物碎屑灰岩油藏的自由水界面。研究结果表明,某一特征点下电阻率与距离自由水界面高度在双对数坐标下具有线性关系。通过理论模型可以表征自由水界面水平和倾斜情况下的分布模式。基于优化算法求解该模型的精度取决于非均质性程度及数据质量控制,当没有观察误差时,曲线完美拟合,曲线拟合的质量随观察误差的增大而降低,因此运用模型前必须做好质量控制,保证模型精度。实际应用中,可以利用降噪技术对数据进行前处理,从而提高模型精度。新方法对于分析地质现象、进行参数计算具有重要意义,还可以推广应用到测井解释、数据质量控制、沉积相分析、地质建模等方面。

关键词: 生物碎屑灰岩, 自由水界面, J函数, 倾斜界面, 特征点对比拾取

Abstract: To conveniently and accurately determine the change characteristics of free water level (FWL), this paper proposes a well log correlation and extraction method based on feature points, a new model for determining free water level based on the aforesaid method, and relevant algorithm procedures. Taking logging curve as the target, a series of datasets with the same or similar petrophysical properties were obtained by searching for the feature points of well log and comparing and tracing cross-well strata. Then a mathematical theoretical model between the elevation depth of free water level, the height above free water level and the electrical resistivity under the same feature point was derived using J function and Archie equation, and the horizontal and tilted distribution modes of free water level were established. Based on the above theoretical model, an algorithm for optimizing the coefficient of determination (R2) in regression equation was proposed, of which the rationality and robustness was verified through presetting experimental data and adding Gaussian white noise. The aforesaid method was applied to determine the free water level of a bioclastic limestone reservoir in Iraq. The research shows that the electrical resistivity under a certain feature point has a linear relationship with the height above free water surface in a log-log coordinate. The horizontal and tilted distribution modes of free water level can be characterized by the theoretical model. The result of model accuracy obtained based on the optimization algorithm depends on the degree of heterogeneity and the control over data quality. In the case of no observation error, the well log can achieve perfect fitting, and the quality of curve fitting decreases with the increase of observation error. Therefore, quality control must be well performed prior to the actual application of the theoretical model to ensure the accuracy. In practical application, noise reduction technology can be used to pre-process data and thus improve model accuracy. The new method introduced herein is of great significance for analyzing geological phenomena and calculating reservoir parameters. It can also be further promoted and applied in the fields such as logging interpretation, data quality control, sedimentary facies analysis, and geological modeling.

Key words: bioclastic limestone, free water level, J function, tilted free water level, well log correlation and extraction based on feature points

中图分类号: