Abstract
The goal of this paper is to predict formation permeability from readily available wireline well log data measurements. The proposed method has many advantages. It is cost effective compared to permeability obtained from core analysis, and most importantly it provides a continuous permeability profile throughout a particular formation horizon as opposed to the permeability obtained from well test analysis, which provides an average value for the entire drainage area of the well.
The method presented in this paper delineates a general methodology for permeability prediction from well log data. First, principal components analysis, cluster analysis, and discriminant analysis are consistently used to identify and categorize clusters based on the unique characteristics of well log measurements reflecting reservoir mineralogy, fluid content, and hydraulic properties within the logged interval; then, fuzzy-logic model is applied to each cluster to determine the formation permeability within the specified interval. Actual field data obtained from a highly complex heterogeneous sandstone reservoir are used to demonstrate the applicability and competence of the proposed technique.
