Fracture Detection Using 3D Seismic Data: A Case study

Oil production from fractured, self-sourced lacustrine shale occurs in many areas of Jiyang Depression, Eastern China. Many information indicated Hydrocarbons within shale is controlled by a fracture system with possible significant lateral connectivity. Characterizing the fractures in this field is of great importance, So the seismic data from the Luo42 3D, acquired over a full 360 degrees azimuthal range, is used to determine the orientation and density of fractures. The basic theory is that P-wave attributes (traveltime, stacking velocity and amplitude) show elliptical variations along the azimuthal direction in an anisotropic medium induced by vertical fractures. The orientation of the long ellipse indicates the fracture strike and the ratio of the long to short axes indicates the fracture intensity. Two methods are used to extract the fracture information: surface fitting and narrow-azimuth stacking. The first method fits an elliptical surface to all-azimuths and all-offsets by least-square fitting. The second method divides the data into six narrow-azimuth volumes, with 30 degrees (-150 ~150) azimuthal bins. Corresponding to these two methods, there are five possible attributes which may be used to extract the fracture information. Surface-fitting has at least three attributes since a surface can be fitted to either bottom-travel time, or interval travel time and/or amplitudes. Two attributes (stacking velocity and AVO gradient) are commonly used for analyzing the narrow-azimuth stacked data. A full test of all five attributes over Target are performed to identify the most reliable and robust attributes, the results show that surface fitting of interval times and narrow-azimuth stacking velocity are two most robust attributes. These two attributes are not only very robust, but also allow easy compensation of the overburden effects and dip-layer effects, and the amplitudes are generally scattered and less reliable.