EUVIMEDEUVIMED entdecken
Crossref

A deep learning identification method for the core of fracture–cavity bodies based on angle-domain data

Zhipeng Gui, Junhua Zhang, Yu Wang, Xingliang Deng, Yintao Zhang, Peng Wang

Interpretation · 01.08.2026 · 14 (3) · T303-T316

DOI: 10.1190/int-2025-0050 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Interpretation, online erschienen am 10.07.2026. Verfasst von Zhipeng Gui und 5 weiteren Autor:innen. Die Arbeit nennt 36 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Deep carbonate fracture–cavity reservoirs represent a class of hydrocarbon-rich reservoirs and are currently a focal area in exploration. Due to their considerable burial depth, seismic reflection structures are often coupled with sedimentary strata, making accurate identification challenging using conventional methods. Furthermore, fracture–cavity reservoirs differ from horizontally layered sedimentary reservoirs, being primarily characterized by vertical strike-slip faults and dissolution features. In seismic exploration, signals acquired at different angles contain distinct reservoir information, providing an advantage for the detailed characterization of fracture–cavity systems. To address the challenge of identifying fracture–cavity core zones, a data-driven, angle-domain deep learning method was proposed. Unlike previous studies, the approach began by analyzing variations in angle-domain seismic data: small-angle data (0°–6°, not fixed) exhibited strong responses to horizontal strata, whereas large-angle data (27°–33°, not fixed) showed weaker horizontal stratigraphic influence, thereby mitigating its interference. Additionally, a neural network model incorporating multiscale convolution, channel attention mechanisms, and spatial attention mechanisms was introduced. Comparative testing demonstrated that the method effectively resolved the mismatch between forward-modeled labels and actual samples — an issue unresolved by conventional models. Validation on field data confirmed that the above approach significantly outperformed second-generation coherence (C2) and the latest coherence (C4) methods in identification accuracy while also enhancing noise suppression and prediction reliability compared to small-angle and conventional migration data. In summary, the method improved the precision of fracture–cavity core prediction by optimizing data sources and the predictive model, demonstrating broad applicability in analogous reservoirs.

Zitieren

APA

Gui, Z., Zhang, J., Wang, Y., Deng, X., Zhang, Y., & Wang, P. (2026). A deep learning identification method for the core of fracture–cavity bodies based on angle-domain data. Interpretation, 14(3), T303-T316. https://doi.org/10.1190/int-2025-0050

Vancouver

Gui Z, Zhang J, Wang Y, Deng X, Zhang Y, Wang P. A deep learning identification method for the core of fracture–cavity bodies based on angle-domain data. Interpretation. 2026;14(3):T303-T316. doi:10.1190/int-2025-0050

BibTeX
@article{Gui2026,
  author = {Gui, Zhipeng and Zhang, Junhua and Wang, Yu and Deng, Xingliang and Zhang, Yintao and Wang, Peng},
  title = {A deep learning identification method for the core of fracture–cavity bodies based on angle-domain data},
  journal = {Interpretation},
  year = {2026},
  volume = {14},
  number = {3},
  pages = {T303--T316},
  doi = {10.1190/int-2025-0050}
}
Kostenlos starten →