Subtle Faults Characterization Based on Fault Simulation and AI OBN Seismic Attributes Optimization

Abstract Subtle faults are often below seismic resolution, especially in strike slip regimes, it is very difficult to identify them as they have small throw and the seismic attributes do not change significantly across the fault. However, faulting often have an impact in fluid propagation leading to compartmentation. To improve the prediction accuracy of subtle faults, a new AI subtle faults characterization method based on OBN seismic attributes and fault physical simulation was proposed. This method includes 5 steps. 1) Integrated regional stress surveys and seismic AI subtle faults prediction results to describe the geometric characteristics. 2) Performing similar fault physical simulation to analyzing the kinematic characteristics, and record the horizontal and vertical features. 3) Comparing the horizontal features of simulation model with seismic attributes along target layer to analyze the orientations. 4) Comparing vertical fault features of simulation model with seismic section to analyze the fault stages and dip angles. 5) Based on the orientations, dip angles and fault stages, optimizing AI fault prediction results. This method has been successfully applied in the Middle East. AI based fault prediction utilizing high-quality OBN seismic data can clearly demonstrate the geometric features for small faults. The study area located in a strike-slip fault zone, with two sets of faults oriented in the NW-SE and W-E directions which were controlled by the W-E shear stress. Under the similar stress, physical simulation of strike-slip faults was conducted, deepening the geological understanding. Comparative analysis between physical simulation and seismic attributes reveals that the NW-SE faults correspond to the early stage of R shear faults, and W-E faults correspond to the later stage of P shear faults. Comparative analysis of fault physical simulation and seismic attribute sections reveal that strike-slip faults can be divided into two stages. The first stage involves deep faults formed in a compressive-torsional stress with reverse faults. The second stage involves shallow faults formed in a tensile-torsional stress with normal faults. Guided by the geometric and kinematic characteristics of strike-slip faults, AI fault prediction results in deep and shallow formation were optimized based on the understanding of fault orientation, dip, and different stages. This AI subtle faults characterization method efficiently combined high- efficiency AI fault prediction technology with fault physical simulation techniques, deepening the geological understanding of faults system while also enhancing the efficiency and accuracy of subtle fault prediction.