Scholarlink

<p><b>Abstract</b>—Determining the properties of heterogeneous reservoirs from microseismic evolution data is an important problem in field development. Analyzing the propagation of microseismic events occurring during fluid injection/withdrawal provides valuable information about permeability and stress state of the reservoir. In this paper, we consider the inverse problem of determining reservoir filtration properties from microseismic event propagation data. For this, the influence of various geological factors on the distribution of microseismic event sources is investigated. Machine learning methods were used to identify correlations between geological model parameters and evolution of microseismicity. Due to the insufficient variability of in situ data, an artificial database of catalogs of microseismic events containing the coordinates of sources and their occurrence times was created to train the model. For this, numerical modeling of fluid injection and generation of microseismic events in synthetic models of permeable media with different geological structure was carried out. Thus, a comprehensive approach to the reconstruction of filtration properties of heterogeneous reservoirs from microseismicity evolution data using machine learning methods is proposed. This methodology can be applied to optimize field development, improve the efficiency of fluid recovery, and reduce the risks associated with the occurrence of undesirable anthropogenic seismic activity.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Reconstruction of the Spatial Distribution of Filtration Properties of Heterogeneous Geological Media Based on Variations of Microseismicity Resulting from Fluid Injection

  • 期刊论文

  • 发表日期：

    2025年4月1日

查看全文
  • E. V. Novikova,
  • N. A. Barishnikov,
  • S. B. Turuntaev,
  • M. A. Trimonova

摘要

Abstract—Determining the properties of heterogeneous reservoirs from microseismic evolution data is an important problem in field development. Analyzing the propagation of microseismic events occurring during fluid injection/withdrawal provides valuable information about permeability and stress state of the reservoir. In this paper, we consider the inverse problem of determining reservoir filtration properties from microseismic event propagation data. For this, the influence of various geological factors on the distribution of microseismic event sources is investigated. Machine learning methods were used to identify correlations between geological model parameters and evolution of microseismicity. Due to the insufficient variability of in situ data, an artificial database of catalogs of microseismic events containing the coordinates of sources and their occurrence times was created to train the model. For this, numerical modeling of fluid injection and generation of microseismic events in synthetic models of permeable media with different geological structure was carried out. Thus, a comprehensive approach to the reconstruction of filtration properties of heterogeneous reservoirs from microseismicity evolution data using machine learning methods is proposed. This methodology can be applied to optimize field development, improve the efficiency of fluid recovery, and reduce the risks associated with the occurrence of undesirable anthropogenic seismic activity.