An introduction to wireline log analysis by integration of ascendant hierarchical clustering and k-nearest neighbor methods for permeability prediction using conventional well logs and core data

Abstract

Permeability is a key parameter in the evaluation of hydrocarbon reservoirs. Conventionally, the core data analysis is utilized to determine the permeability. However, core analysis method is time-consuming and expensive. In this paper, we present an approach for permeability prediction from conventional well logs and core data. Our approach integrates two methods, the ascendant hierarchical clustering, which is used to classify the well log responses into relatively homogeneous subgroups based on electrofacies, and the k-nearest neighbor method, to predict the permeability from the nearest neighbors that belong to a similar subgroup. The procedure is applied on one of the Iranian South West oil reservoirs. The results of this study show that integration of well logs and core data based on electrofacies can reliably predict permeability.