Research and Application of Wide-Azimuth Seismic Multi-scale Fracture Prediction Technology: A Case Study of Triassic Carbonate Reservoirs in P Oilfield

The Mangyshlak Basin in Kazakhstan is rich in oil and gas resources, and the hydrocarbons are predominantly distributed in Middle Jurassic sandstone and Middle-Lower Triassic carbonate formations. The Middle Jurassic sandstone reservoir is the most important in the basin, characterized by shallow burial depths and predominantly medium-to-low porosity and permeability. Most of these oilfields have entered mid-to-late development stages, experiencing increasing production decline. The Middle-Lower Triassic carbonate formations, adjacent to source rocks, exhibit favorable hydrocarbon accumulation conditions and significant exploration potential. However, these reservoirs are deeply buried and exhibit relatively poor petrophysical properties, with well-developed fractures and vugs. Development practices indicate that the degree of fracture development has an important impact on production levels. Yet, conventional fracture characterization methods have limited accuracy, failing to meet the demands of efficient exploration and development. To address this, a multi-scale fracture prediction study was conducted on the Middle Triassic (T2) formation in the P Oilfield using “wide-azimuth, wide-bandwidth, and high-density” seismic data to enhance fracture identification accuracy. First, post-stack seismic data were subjected to fracture enhancement processing. Subsequently, large-scale fault prediction was performed using coherence, curvature, and discontinuity attributes. Further refinement was achieved by integrating maximum likelihood attributes and an intelligent fault detection method based on a U-net network to predict minor faults, significantly improving prediction accuracy over conventional approaches. Offset Vector Tile (OVT) processing and azimuthal velocity analysis were then performed. Based on anisotropy theory in horizontally transverse isotropic (HTI) media, azimuthal velocity and amplitude anisotropy inversion were conducted for fracture prediction. The direction of the fast P-wave effectively indicated fracture orientation, while the ratio of fast-to-slow P-wave velocities was used to calculate anisotropy parameters, enabling the prediction of spatial fracture distribution in the T2 formation. The fracture prediction results are in good agreement with geological laws and drilling results, and the effect is obviously better than that of the anisotropic fracture prediction result of azimuth amplitude. Fracture prediction results provide strong support for well placement and target optimization.