The characterization of interwell section architecture is critical for revealing reservoir lateral heterogeneity and connectivity. This process integrates well and seismic data with geological knowledge yet faces inherent multiple solutions. Current characterization methods remain hampered by high levels of manual intervention, insufficient automation, and difficulties in evaluating the uncertainty of interwell section architecture. To address these challenges, this study presents an intelligent method for the automated characterization of reservoir architecture along section directions based on a Bayesian expert system. The approach quantifies domain knowledge via prior normal distributions. By utilizing well and seismic data, Bayesian probabilistic reasoning infers the guiding influence of each individual piece of domain knowledge on predicting the interwell distribution of architectural elements. A weighted ensemble decision framework then integrates these inferences to determine the interwell distributions of architectural elements and associated uncertainties. Case studies demonstrate that the method effectively evaluates uncertainty, generates geologically consistent section characterizations, achieves 81% consistency in blind well sand body predictions, and excels in delineating the lateral boundaries and contact relationships of architectural elements.

The depositional architecture of submarine fans in rift basins is significantly controlled by complex geomorphology created by widespread normal faults, presenting a key challenge in deep-water sedimentology. Although extensive previous studies have established depositional architectures and sand body distribution patterns of transverse submarine fans controlled by graben boundary faults and interbasin transfer zones, research on how axial submarine fan architecture responds to intra-graben slope gradients and evolving transverse confinement remains inadequate. This study takes the Upper Jurassic B4 oil group in the North Sea X Oilfield as an example. By using 3D seismic data, cores and logging data to restore paleo-geomorphology and dissect depositional architecture, and further reveal the controls of intra-graben paleo-geomorphic variations on axial submarine fan depositional architecture. Our analysis shows that the paleogeomorphology in the study area features axial stepped slope breaks and phased evolution in transverse confinement. As the axial slope transitions from extremely steep slope segments to steep slope segments, slope transition zones and gentle slope segments, turbidity currents evolve from supercritical to subcritical states through hydraulic jumps, while generating divergent flows. This progression drives architectural transformation from sediment bypass, incised channel–overbank systems, distributary channels and channelized lobes to lobes. Concurrently, phased transverse confinement evolution controls vertical stacking characteristics of individual channel–lobes: early asymmetric stages produce lateral migration stacking; middle symmetric stages develop unordered compensational stacking; and late locally confined stages form deflected retrogradational stacking. We propose a dynamic submarine fan depositional architecture response model, which emphasizes how evolving paleo-geomorphology directly controls spatiotemporal configurations of architectural elements by altering gravity flow pathways and energy distribution, and is further modified by feedbacks where the deposits themselves become influencing topographic elements. It provides a new perspective for deep-water depositional models and reservoir prediction in areas with similar geomorphic settings.

High-precision sand prediction is fundamental to improving the efficiency of oil and gas exploration and development. To address the limitations of traditional fixed-weight fusion strategies, particularly under conditions of significant lateral variation in sand body distribution, this study proposes a dynamic weighting–deep neural network (DW-DNN) for adaptive frequency-decomposed attribute fusion. The approach integrates physical constraints with deep learning and introduces two innovations: (ⅰ) a priori weight matrices derived from the amplitude–frequency and tuning thickness relationship (amplitude variation with frequency, AVF) are embedded into the attention mechanism to adaptively allocate multiband seismic attributes, emphasizing high-frequency features for thin sands and low-frequency features for thick sands; and (ⅱ) a deep neural network with a composite loss function combining mean squared error (MSE) and AVF-based constraints is designed to jointly optimize weight allocation and prediction accuracy. The method was applied to the Xi 233 area of the Qingcheng Oilfield in the Ordos Basin and compared with conventional approaches. DW-DNN achieved high accuracy and generalizability, with an R² of 0.92 in the 30% blind-well test, 24.3% higher than conventional methods. In addition, 91% of well-point errors were within 0–3 m, while prediction accuracies for thin (≤3 m) and thick (>3 m) sands reached 88% and 91%, respectively. The model also maintained stable performance under low well-control conditions (training–test ratio 5:5). Predicted sand distributions exhibited improved continuity and geologically plausible geometries, clearly delineating channels, lobes, and estuary bars. The results demonstrate that DW-DNN enhances frequency-decomposed attribute fusion through adaptive weight allocation, providing a robust tool for predicting sand body distributions in complex reservoirs.

Reservoir physical parameters serve as fundamental quantitative indices for characterizing the storage capacity and fluid percolation potential of subsurface reservoirs. Well logging interpretation, a critical methodology for accurately estimating these parameters, constitutes a sophisticated nonlinear regression challenge. To address the inherent limitations of existing petrophysical parameter interpretation techniques, particularly their inadequate generalization performance under few-shot learning conditions, this investigation systematically devises a dual-framework analytical approach. This study initially proposes a sample optimization methodology based on cluster analysis. The spatial configuration of samples is partitioned through the implementation of the K-means clustering algorithm, followed by selective sample curation according to spatial distribution characteristics to maximize learning sample diversity. Building upon this optimized sample architecture, the study further introduces a hierarchical residual neural network-based interpretation framework for petrophysical parameter estimation. The proposed methodology enhances conventional fully connected neural architecture through four innovative mechanisms: (1) Integration of cross-layer residual connections facilitates progressive refinement of residual mappings between multivariate logging inputs and target petrophysical outputs, thereby enabling hierarchical abstraction of complex petrophysical relationships from limited training instances.(2) The integration of ensemble learning paradigms amalgamates diverse machine learning methodologies, effectively mitigating overfitting risks through algorithmic diversity. (3) The implementation of a multi-task learning framework establishes intrinsic correlations between porosity and permeability interpretation tasks via shared latent representations, thereby enhancing individual task generalizability under data scarcity constraints. (4) The introduction of a quadratically weighted root mean square error loss function preferentially reduces interpretation errors in high-permeability reservoir intervals. Results from 90 rigorously designed comparative experimental configurations in the study area demonstrate that the cluster-based sample optimization methodology effectively enhances generalization performance across multiple machine learning models under few-shot learning constraints. Application of the proposed hierarchical residual neural network framework for well-logging interpretation of reservoir porosity and permeability within the investigated reservoir area achieves coefficients of determination of 88% and 94%, respectively, demonstrating statistically significant superiority over conventional methodologies in both petrophysical interpretation accuracy and generalization capability. Blind testing validation on cored wells reveals 12 and 20 percentage point improvements in predictive precision compared to other various existing methodologies, the proposed approach in this study demonstrates substantial advancements in addressing few-shot learning challenges through algorithm optimization strategies encompassing distribution-based sample selection and multi-task collaborative frameworks. This methodology significantly enhances feature representation fidelity in petrophysical datasets, exhibiting superior petrophysical interpretation accuracy and enhanced generalization capabilities.

The submarine fan is an important reservoir for oil and gas in deep water areas. The differences in reservoir quality have a significant impact on the differential accumulation and exploitation of oil and gas. Previous studies have conducted extensive research on the differences in reservoir quality of submarine fans. However, the characteristics and distribution patterns of reservoir quality differences within submarine fans under a steep continental slope background are still unclear. This paper takes the Oligocene submarine fan reservoir in the X gas field of the Rovuma Basin in East Africa as the research object. By integrating core, well logging and seismic data, an in-depth study has been carried out on the control of reservoir quality differences and distribution patterns of submarine fan sedimentary microfacies and lithofacies under the steep continental slope background. The results show that the changes in reservoir quality within the submarine fan are mainly controlled by rock texture, lithofacies (association) and sedimentary microfacies under the circumstance of weak diagenesis. Grain sorting and clay content mainly control the porosity and permeability of the reservoir, respectively, but the relationship between grain size and reservoir properties is very complex. In sand-rich lithofacies, fine sandstones have the highest porosity due to their good sorting, and massive gravel-bearing coarse sandstones have the highest permeability due to their low clay content. Under the steep continental slope background, the submarine fan sedimentary microfacies are arranged in the order of muddy channel-sandy channel-lobe main body-lobe edge along the source direction, resulting in the source-directional differences in reservoir quality in the order of “poor, good, and poor”. The proximal muddy channel consists of fine-grained and clay-rich lithofacies, with overall poor physical properties. In the middle position, the sandy channel and lobe main body change to massive gravel-bearing coarse sandstone lithofacies and medium-coarse sandstone lithofacies, with low clay content and improved to good physical properties. Among them, the reservoir quality of the sandy channel is better than that of the lobe main body. The internal high porosity and high permeability zones of the sandy channel are in the form of elongated lenses, while the relatively high porosity and high permeability areas of the lobe main body are in the shape of lobes. The distal lobe edge changes to fine-grained lithofacies (fine-medium sandstone, fine sandstone) with increased clay content and gradually deteriorated physical properties.

Seafloor topography plays an important role in the evolution of submarine lobes. However, it is still not so clear how the shape of slope affects the three-dimensional (3-D) architecture of submarine lobes. In this study, we analyze the effect of topography factors on different hierarchical lobe architectures that formed during Pliocene to Quaternary in the Rovuma Basin offshore East Africa. We characterize the shape, size and growth pattern of different hierarchical lobe architectures using 3-D seismic data. We find that the relief of the topographic slope determines the location of preferential deposition of lobe complexes and single lobes. When the topography is irregular and presents topographic lows, lobe complexes first infill these depressions. Single lobes are deposited preferentially at positions with higher longitudinal (i.e. across-slope) slope gradients. As the longitudinal slope becomes higher, the aspect ratio of the single lobes increases. Lateral (i.e. along-slope) topography does not seem to have a strong influence on the shape of single lobe, but it seems to affect the overlap of single lobes. When the lateral slope gradient is relatively high, the single lobes tend to have a larger overlap surface. Furthermore, as the average of lateral slope and longitudinal slope gets greater, the width/thickness ratio of the single lobe is smaller, i.e. sediments tend to accumulate vertically. The results demonstrate that the shape of slopes more comprehensively influences the 3-D architecture of lobes in natural deep-sea systems than previously other lobe deposits and analogue experiments, which helps us better understand the development and evolution of the distal parts of turbidite systems.