The Study On Fracture Prediction In The Thin Interbed Mud-dolomite Reservoir,Jiuquan Basin

Based on the analysis of different scale fractures characteristics of thin interbed mud-dolomite reservoir in Qingxi sag,two geological models of different lithologic combinations are established.The lithology combination,the attitude and thickness of stratum,and the distance of the faults were closely related to the fracture development degree.Through petrophysical modeling,the seismic responses of different scale fractures were obtained.The results confirmed that the anisotropy of frequency gradient and attenuation could be used to identify and quantitative describe the sub-seismic fractures.Meanwhile,some poststack seismic attributes are respectively used to make comparisons in different sensitivity of the fracture detection,including structure-oriented coherence,dip and azimuth,maximum structure-oriented curvature and eigenvalue.The geological and geophysical parameters related to fractures were used to optimize the inversion model,and a non-linear neural network method was applied to determine the weight of different poststack seismic attributes in the fusion process.The strength and average density of fracture of the Cretaceous thin interbed mud-dolomite reservoir in Qingxi depression was quantitatively predicted.Research results show that the coincidence rate of multi-attributions fusion is up to 82% compared with FMI recognition result.