Study On The Rock Fractures Prediction Methods From Seismic Data

The thesis presents the strategies to extract fracture information from seismic data. Fracture is the important spaces and migration channels of low permeable reservoir. Fracture detecting is one of the focus research topics in oil and gas exploration field. Seismic exploration is the best way to explore rock fractures.There are many ways to extract fracture information from reflected seismic records. As seismic data can be converted into digital image, the edge detecting method can be used to predict rock fractures; as rock fractures is often of fractal characteristic, the fractal method can be used to detect rock fractures; as seismic record is a discrete time sequence, the wavelet multiple resolution analysis method can be used to detect rock fractures. By theory analyses, numerical experiment and application test, the follows conclusions are attained:1. The outstanding role of edge detecting algorithm is that it can be used to find out the boundary between different areas. The edge detecting method can be used to recognize developed area of fractures.2. Multiple scale fractal dimension method detects the rock fractures according the fractal dimension of the seismic amplitude data in different scales. The method developed avoids the difficulty in computing fractal dimension in single scale case. The method is good at recognizing the grades of the fractures. However, to determine the grades of the fractures it is need to set suitable thresholds based on the drilling and well logging.3. The fractal compressing method based on iterative function system is good at detecting the extend of fractures.4. The application results of wavelet multiple resolution analysis are depend on the wavelet base chosen. The multiple resolution analysis using db3 wavelet is good at detecting the strike and grades of fractures.5. Rock fractures is often of variable size, density, geometry and so on. It is recommended to adopt jointly the method based on wavelet and fractal theory to detectfractures. It is also recommended that we should try to extract fracture information using the method based on wavelet packet analysis...