Seismic-Print Analysis Enhanced Multiattribute Fusion for Fracture Distribution Prediction in Tight Reservoirs

Predicting fracture development and distribution is crucial in unconventional hydrocarbon exploration and development, especially for sweet spot evaluation and well location optimization in tight reservoirs. Compared to faults, fractures are smaller in scale, have weak seismic reflection signals, and are characterized by irregularity and complexity, making accurate prediction challenging. Conventional fracture prediction methods fail to accurately characterize fracture development zones and their distribution characteristics, which severely restricts the efficient exploration and development of tight reservoirs. To address this limitation, this study proposes a seismic-print analysis enhanced multi-attribute fusion approach for fracture distribution prediction in tight reservoirs. First, considering the inherent consistency between seismic waves and acoustic waves, the voiceprint analysis method from the field of speech recognition is borrowed to perform seismic-print analysis on seismic data to enhance the seismic response characteristics of fractures. On this basis, the gradient structure tensor algorithm is used to extract stratigraphic dip and azimuth information, and then combined with the eigenvalue coherence algorithm and ant-tracking technique to calculate the coherence and ant-tracking attribute volumes to characterize the fracture characteristics at different scales. Finally, the attribute proportional integration method is employed to integrate the calculated coherence and ant-tracking attribute volumes to predict fracture distribution. The theoretical model validation results show that the proposed method is effective and applicable. The field data application in the Sichuan Basin show that the fractures predicted by this method are clearer in shape, smaller in scale, and richer in detail, which are consistent with the actual geological understanding. This study provides a novel approach for multi-scale fracture prediction in tight reservoirs, which can characterize the fracture distribution at different scales more comprehensively, especially for small fractures with weak seismic response, it can achieve better results than conventional methods.