A NOVEL APPROACH BASED ON FEATURE FUSION FOR FRACTURE IDENTIFICATION USING WELL LOG DATA

Abstract

Accurate identification of fractures is necessary and complex for carbonate reservoir exploration. Using conventional well logs and geological data, we identify various fracture identification methods based on depth point information and waveform processing. The results show that the method based on equivalent medium theory maintains high stability and accuracy in reflecting the secondary pores in cases of unfavorable borehole environments. Both the acoustic log and dual lateral difference fractal dimensions increase in line with the degree of fracture development. The high-frequency energy information shows significantly high values in the fractured zone on a suitable scale. Finally, the fractures are characterized by a novel approach based on feature fusion. The linear predictive relationship for fracture identification via proposed comprehensive factor scores (CFS) avoids the influence of the deviation of a few variables on the stability of the overall results. Our study offers a new framework for fracture identification in the exploration and evaluation of carbonate reservoirs.