APPLICATION OF MULTIPLE ATTRIBUTES CLUSTER ANALYSIS TO PERMIAN DEPOSITS IN THE SOUTH YELLOW SEA BASIN

The change in formation parameters causes the change in seismic attributes. But it is not a one-to-one relationship between the changes. Multi-solution problem comes from the reservoir parameters inversion of seismic data, and multi-attribute integration is one of the important means to solve the problem. In the South Yellow Sea Basin, drilling holes are rare and the single well control area is very large. Therefore, it becomes a necessity to make full use of seismic attributes information for facies analysis of the Permian Longtan and Dalong Formations. In this paper, the RMS amplitude, reflection intensity, relative acoustic impedance, dominant frequency, and average instantaneous frequency are all used to cluster analysis of Permian Formations by neural network algorithm. The results show the distribution pattern of sandstone and limestone lithofacies. The process not only avoids the multi-solution problem of a seismic attribute, but also reveals the characteristics of the clastic rock above the lower Permian Qixia limestone. Combined with well-seismic analysis, it is concluded that various facies, such as continental shelf facies, tidal flat-lagoon-swamp facies and fluvial facies are well developed during the Longtan-Dalong sedimentary period of Permian