| During the past several years there has been a sudden and intense interest in the use of artificial intelligence techniques in petroleum industry. This thesis explores the use of machine learning approach, specifically decision tree learning, as a means to identify geological formation facies from well logs. Identifying geological formation facies is critical for economic successes of reservoir management and development. Formation facies usually influence the hydrocarbon movement and distribution. The identification of various facies, however, is a very complex problem due to the fact that most reservoirs show different degree of heterogeneity. In this thesis, the existing methods are surveyed, and we propose a new methodology. The current conventional process to solve this problem is tedious and time consuming. We propose the use decision tree learning (DTL) approach as a means to predict facies from well logs. We report a six-step case study on a real oil field. (Abstract shortened by UMI.) |