Scholarlink

<p>Facies classification and channel detection remain challenging in geologically complex reservoirs. Faulting, thin-bed effects, and wavelet interference strongly influence seismic responses in such settings. This study presents an integrated multi-attribute seismic workflow utilizing Self-Organizing Maps (SOM) to enhance facies classification and delineate gas-bearing channels within the Permian Rotliegend reservoir of the Groningen Field, the Netherlands. A 3D post-stack seismic volume and four wells were utilized. The workflow integrates seismic interpretation, colored inversion, and unsupervised machine learning. Principal Component Analysis (PCA) was applied to evaluate attribute variability, followed by sensitivity testing to refine the final attribute set used in SOM classification. The SOM workflow successfully delineates NW-SE to NNW-SSE gas sand channels, consistent with the known fluvial-aeolian depositional framework of the Rotliegend Formation. Validation against well-log-derived lithofacies shows good agreement. The results highlight that multi-attribute SOM analysis provides clearer channel geometry and improved facies discrimination compared to individual seismic attributes.</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

Facies classification and channel detection for reservoir characterization using self-organizing maps in geologically complex Groningen gas field, the Netherlands

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月22日

查看全文
  • Sarah El-Attar,
  • Abdel-Khalek El-Werr,
  • Eman Mohamed Abdel-Rahman,
  • Mohamed Nabil Sobhy

摘要

Facies classification and channel detection remain challenging in geologically complex reservoirs. Faulting, thin-bed effects, and wavelet interference strongly influence seismic responses in such settings. This study presents an integrated multi-attribute seismic workflow utilizing Self-Organizing Maps (SOM) to enhance facies classification and delineate gas-bearing channels within the Permian Rotliegend reservoir of the Groningen Field, the Netherlands. A 3D post-stack seismic volume and four wells were utilized. The workflow integrates seismic interpretation, colored inversion, and unsupervised machine learning. Principal Component Analysis (PCA) was applied to evaluate attribute variability, followed by sensitivity testing to refine the final attribute set used in SOM classification. The SOM workflow successfully delineates NW-SE to NNW-SSE gas sand channels, consistent with the known fluvial-aeolian depositional framework of the Rotliegend Formation. Validation against well-log-derived lithofacies shows good agreement. The results highlight that multi-attribute SOM analysis provides clearer channel geometry and improved facies discrimination compared to individual seismic attributes.