Well log correlation using artificial intelligence

Abstract

Computer-assisted approaches to well log correlation are of interest to petroleum engineers and geologists for two reasons, in large field studies a computer can be used to simply reduce the time required to correlate zones of interest. It is also possible that computer-assisted correlations may suggest zonal matches of interest and originality that night not have been considered. The objective of this study is to develop a new approach to computer-assisted log correlations. The new approach is based on a zonal correlation, methodology and an artificial intelligence technique -- rule-based systems. A prototype rule-based computer software system is developed to implement this method. The system correlates logs in a fashion parallel to human experts. With this system, correlations are directly derived from the shapes of log traces. Results obtained from applying field data show that this new approach can correlate distinct geological horizons with a high degree of success.