Seismic facies classification using 2‐D and 3‐D multi‐attribute hierarchical clustering algorithms

Abstract

Summary Reflection seismic data interpretation is an applied method in oil and gas exploration industry. Interpretation of seismic facies could help understanding complexity of internal stratigraphic geometries of complex sequences. In this paper, three hierarchical clustering procedures single, complete and average linkage algorithms are utilized to categorize seismic facies based on seismic attributes. In the hierarchical clustering procedure, input data is normalized and entered to the principal component analysis algorithm. Principal component analysis reduces redundant data and random noises. Then, according to dissimilarity matrix, time samples are integrated into clusters of similar patterns. By applying these three algorithms to a synthetic earth model it is concluded that the average linkage algorithm has the best results compared to the others. In order to evaluate the method in 3-D seismic data, nine seismic attributes had been calculated and entered to PCA using Paradigm software. Six principal components have been selected based on PCA. By applying clustering algorithm to these components in two ways, 2-D and 3-D, it is shown that improved resolution in 3-D clustering method appears to be more diagnostic in lateral seismic facies variations.