Well log facies classification for improved regional exploration ^*

Abstract

Well log and rock sample data from fourteen offshore petroleum exploration wells have been successfully up-scaled and integrated using facies classification. Over 100 rock samples were categorised into six petrofacies classes based on their composition and texture. Cluster analysis was used to classify well log data into five electrofacies units, after careful conditioning and selection of input logs. The result is a direct link between well logs and rock samples. Electrofacies profiles clearly illustrate stratigraphic information previously hidden in the well logs. In addition, well log acoustic rock property relationships based on the new electrofacies classes are found to be better constrained than lithology-based models.

Previous facies studies either have been applied at a field scale or have had conventional core for petrofacies calibration. In this paper, I illustrate how facies analysis can be successfully applied at a regional scale with only sidewall sample calibration. Particular attention was given to conditioning the cluster analysis input logs by removing all effects of fluid fill and mechanical compaction, which varied significantly across the study area. A lesson learned from the project was that it is easy to generate misleading results with cluster analysis, so care was taken to select only the most appropriate input logs, and to thoroughly quality control the output electrofacies.