Answers are grounded in this paper and cite their sources.
The deepwater area of the Qiongdongnan Basin is characterized by sparse well control, low-porosity, and low-permeability reservoirs, weak seismic responses, and small elastic-parameter contrasts. These conditions increase the non-uniqueness of conventional reservoir prediction. To reduce this uncertainty, this study proposes a multi-level facies-constrained reservoir prediction workflow that couples sedimentary facies, seismic facies, and logging lithofacies. The workflow first constructs a facies prior from seismic waveform classification, well-log interpretation, and sedimentary constraints. It then performs Bayesian pre-stack inversion based on the exact Zoeppritz equation, followed by semi-supervised porosity prediction and frequency-dependent amplitude variation with offset flowability analysis. Finally, the P/S velocity ratio, porosity, and water-saturation-related attributes are integrated using a competitive neural network to delineate favorable reservoir zones. Synthetic tests and a Stanford reservoir model test show that the proposed method improves facies agreement from 78.57% to 92.86% relative to a conventional facies-constrained workflow. Because the field area contains only one drilled well, the field results are interpreted as probabilistic exploration targets rather than blind-well proof of deterministic accuracy. The workflow provides a reproducible and geologically constrained approach for reservoir prediction in sparse-well deepwater settings.
No takes yet. Share an insight, caveat, or question.
Sun et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: