Application of a dynamic optimization-based multi-attribute fusion method for fault detection

Abstract

The focus of oil and gas exploration in the Tarim Basin has shifted from interlayers to fracture-controlled karsts. A significant oilfield characterized by strike-slip faults was discovered in the Yueman area. However, identifying such fault zones is challenging because of the complex and chaotic seismic reflection characteristics, as well as the presence of seismic noise and other discontinuities. To improve oilfield production, the accurate identification of strike-slip fault zones in ultradeep tight limestone is a critical issue. The seismic anomalies of such fault zones exhibit diverse characteristics, with the low-velocity zone of the fault causing a “beaded” reflection pattern. Traditional coherent and curvature attribute methods have large errors in identifying strike-slip faults and cannot adequately characterize the contour features of fracture-karst traps. To address these challenges, this study proposed a multi-attribute optimal surface-based fracture identification technology based on forward simulation records. Seismic attributes that were sensitive to different types of strike-slip faults were selected, and multiple attributes were merged to obtain a fracture distribution map using the best surface voting algorithm. This method effectively suppresses noise that is irrelevant to fractures and is sensitive only to fracture information, allowing for the identification of subtle waveform changes caused by strike-slip faults. Thus, the accuracy and continuity of fracture identification were significantly improved.