Self-organizing mapping neural network analysis on seismic-sedimentary facies

Abstract

The basic research on structural style and sedimentary system is weak, which will seriously restrict the exploration and development process of the new study area. Taking the strata of Bayingobi formation of Lower Cretaceous in L sag of Y basin as the research object, relying on seismic and a small amount of well data, this paper introduces self-organizing mapping neural network analysis into the process of seismic attribute clustering, fully excavates the classification information of seismic facies reflected by post stack seismic data, and carries out the study of seismic- sedimentary facies. The results show that the seismic facies distribution of Bayingobi formation in L sag conforms to the structural zoning, and the boundary of facies zone is clear. There are mainly four types of sedimentary facies: Fan Delta, braided river delta, shore-shallow lake and deep lake. The application of this technical method reduces the unreliability of seismic facies analysis results and provides a new basis for sedimentary facies analysis of oil and gas exploration in the new study area.