This research focuses on the critical task of facies classification, involving the characterization of various types of rocks and sedimentary structures based on well log data. The study introduces a novel approach that utilizes machine learning techniques in the Fandango Field. Initially, zones with anomalous productivity are isolated, followed by the application of unsupervised machine learning for outlier detection. These outliers are then replaced with synthetic data derived from multiple linear regression. Additionally, the study applies agglomerative hierarchical clustering to perform facies analysis on a selected set of wells, successfully identifying six geologically significant classes. Finally, a random forest model is used for supervised facies classification, demonstrating notable improvements by conditioning well log data using machine learning. The results are validated through a cross-sectional stratigraphic analysis.
Fill out the registration form to watch the video recording!
