The Application of Semi-Supervised Geobody Detection Technique using Multiple Seismic Attributes in Petroleum Exploration
Li, Lei; Ran, Xianhua; Tao, Chun Feng; Wan, Zhonghong; Zhan, Shifan
In last decades, with the progress of exploration and researches, the distribution of structural reservoirs underground has been clearly known on the whole, while complex stratigraphic lithologic reservoirs, represented by small fault blocks, fracture zones, carbonate caves and reefs etc, are not precisely found out. In addition, conventional exploration techniques are facing more and more difficulties in the aspect of finding new reservoirs, since the targeted areas have moved from middle-shallow strata to middle-deep strata, and types of reservoir from simple structural to stratigraphic lithologic. Consequently, the recent research and development of work were focused on the accurate and efficient recognition of special geobodies (such as small fault blocks, fracture zones, carbonate cave s and reefs etc), which aimed at not only heightening production efficiency, but also reducing exploration risks.
This paper presents a semi-supervised geobody detection technique using seismic multi-attribute analysis and shows how this technique can be used to estimate some important reservoir parameters from seismic volume. The technique, as a nonlinear and completely data-driven method, requires no accurate initial model and no priori operator linking the reservoir properties with seismic attributes and inversion data. In practical applications, this technique is firstly used to estimate the space distributions of karst carbonate caves, located in western China. It shows that the predicted dissolved pores and fractures are identified precisely and agree with the logging data near the wells. In the second example, we pay our attention to depict the reefs reservoir from western China. This type of geological bodies is characterized by large buried depth, compact lithological character and strong anisotropy. As shown by the results, the margin forms, sketch structures and inner cavities of the reefs are delineated clearly, respectively. Finally, small fault blocks and fracture reservoirs from eastern China will be identified automatically by this present technique. The results indicate that various scales of fault blocks and fractures are identified clearly, though the thickness and volume of geobodies may be small and change fast laterally. In addition, high productive oil flow was obtained in all of these reservoirs.
AAPG Search and Discovery Article #90163©2013AAPG 2013 Annual Convention and Exhibition, Pittsburgh, Pennsylvania, May 19-22, 2013