Determination of permeability transforms from geophysical logs using statistical pattern recognition: The Mardie greensand

Abstract

Many permeability models and equations are empirical in log analysis and evaluation. Therefore, their application is restricted to a given set of lithology conditions. It is essential to select the correct transform relationships for each type of lithology in order to determine accurately permeability from well logs. This is especially true in the Mardie Greensand, a lithologically complex formation. There are four porosity and permeability transforms determined based on the core data in the sandstone formations of the Mardie 'B' and 'C' intervals in the Barrow Field, Carnarvon Basin, Western Australia. The aim in this study was to determine which porosity and permeability transform is applied at which formation from well logs, using statistical pattern recognition of electrofacies. In this paper a number of electrofacies corresponding to these porosity and permeability transforms were determined in the complex formation. The application of pattern recognition to this complex lithology makes it possible to identify electrofacies and select corresponding transforms to calculate permeabilities from well logs. Results indicate that it can reduce the uncertainty of log-derived permeability and help to simplify the problem of complex lithologies in log interpretation by recognising known lithologies within a formation, thereby improving the precision of permeability determination in this complex formation