12. Imaging/Interpretation Frontiers
Fault Detection using Coherency Attribute in Khangiran Gas Field, Iran
Ali Reza Javahery and Abdolrahim Javaherian
University of Tehran, Iran.
Abstract
The final step of a seismic exploration project is data interpretation which is a blend of art and science. In modern interpretation procedures use of attributes are inevitable. In other words the interpreter cannot extract full information from seismic data without use of these kinds of tools. Today, attributes have a proven role in recognition of buried deltas, river channels, reefs, fans, faults and similar structures. Generally, seismic attributes describe relation between seismic data and measure of seismic characteristic of interest. The growth of attributes is very high and many attributes have been introduced by researchers so far. Each one of these attributes work on different mathematical basis and can show special aspects of data. One of modern seismic attributes is coherency which has an important role in interpretation of structural discontinuities and stratigraphy features in 3D data. This attribute measures the similarity of traces in three dimensions and therefore present interpretable changes in these cases. The similar traces are mapped with high coherence coefficients while anomalies as well as discontinuities have low coefficients. Coherency attribute shows evaluation criterion of lateral changes in the seismic response, caused by variation in structure, stratigraphy, lithology, porosity and the presence of hydrocarbon. The output of this algorithm is a coherence cube. In this cube structural discontinuity and stratigraphic faces will be depicted with a better resolution. The coherency cube can be computed using three approaches. 1-cross-correlation, 2-semblance and 3-eigenstructure. In this paper, we examined first and third approaches and compared the results. The third method is a matrix based approach and is faster. The objective of this study was faults detection. We showed the output of these algorithms on 3D synthetic models of low SNR. Also, we applied this attribute on one of the gas field data in Iran. As the results show this attribute can facilitate faults recognition. The comparison of cross-correlation and eigenstructure method shows the latter one can map faults with a better lateral and vertical resolution.
Last modified: Wed May 10 08:27:34 2006