Answers are grounded in this paper and cite their sources.
Demonstrates enhanced lithofacies prediction in a deepwater system, indicating improved reservoir characterization.
Accurate lithofacies prediction in deepwater turbidite systems remains a persistent challenge because of lateral and vertical heterogeneities that operate below conventional seismic resolution. This study presents a novel unsupervised machine-learning workflow applied to the Simian Field, West Delta Deep Marine (WDDM) Concession, offshore Nile Delta, Egypt. The workflow integrates multi-attribute seismic classification using Self-Organizing Maps (SOMs) with K-means clustering of petrophysical well-log data to generate a high-resolution, three-dimensional lithofacies prediction volume. Several instantaneous and spectral decomposition seismic attributes were computed from 3D post-stack depth-migrated (PSDM) data; Principal Component Analysis (PCA) was then used to select the six attributes carrying the greatest discriminating power, which were subsequently ingested by a 7×7 SOM (49 neurons). Independently, K-means clustering of volume shale (Vsh), effective porosity (ϕe), and water saturation (Sw) logs from six wells defined three lithofacies: Gas Sand, Wet Sand, and Shale. SOM neurons were cross-tabulated with well-derived lithofacies to build conditional probability relationships that were propagated across the full seismic volume. Statistical robustness was assessed through Cramér's V and a percent-fit metric evaluated at blind-well locations. Results demonstrate that the SOM-based approach resolves stratigraphic detail comparable to conventional seismic interpretation while offering significantly reduced computational cost and subjectivity. The workflow provides a practical, reproducible framework for lithofacies prediction in complex deepwater settings where well control is sparse.
No takes yet. Share an insight, caveat, or question.
Hussein et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: