Perpustakaan Digital ITB

In the oil and gas industry, knowledge of reservoirs is crucial. Reservoir characterization is performed to determine the performance of a reservoir in storing and producing fluids by determining the reservoir's information and properties. Simulating reservoir models for field development can be aided by formation type, porosity, permeability, and water saturation data. Conducting formation, porosity, permeability, and water saturation analyses can be time-consuming and expensive. An effective and efficient method is required to obtain formation, porosity, permeability, and water saturation data. This case study predicts formation, porosity, permeability, and water saturation using well log data and core analysis from 12 wells containing six distinct formations. Several models will be trained using machine learning to predict formation, porosity, permeability, and water saturation. In practice, using all the data may result in suboptimal predictions, so feature selection and data preprocessing are crucial prior to training the models with the data. In addition, hyperparameter tuning will be performed to optimize the training performance of the model. The leave-out well method is also employed to evaluate the model's ability to predict new data. After evaluating the entire model, the best model will be used to predict specific properties. According to the findings of this study, the accuracy of predictions made using machine learning techniques is satisfactory. In this study, three scenarios will be compared: making predictions using only well log data, using well log data in addition to formation data, porosity, permeability, and water saturation from interpretation and measurement, and making integrated predictions using well log data in addition to formation data, porosity, permeability, and water saturation from machine learning model predictions. Using machine learning techniques, the results of this comparison will be used to determine the optimal scenario for predicting formation, porosity, permeability, and water saturation.