PREDICTIVE MODELING OF PETROLEUM RESERVOIRS USING MACHINE LEARNING AND SIMULATION

ABSTRACT The predictive modeling framework presented in this study has the potential to revolutionize the way reservoir behavior and production rates are predicted and optimized in the oil and gas industry. By leveraging the strengths of machine learning and simulation, the framework can provide more accurate and robust predictions than traditional methods, enabling better decision-making and optimization of reservoir management strategies. The study's results demonstrate the effectiveness of the framework in predicting production rates and reservoir behavior, with a high degree of accuracy and reliability. The machine learning models used in the study, including SVMs, Random Forest, and GBMs, are shown to be effective in capturing complex relationships between input features and production rates. The framework's ability to optimize production rates and reservoir management strategies is also demonstrated, with significant potential for improving the efficiency and effectiveness of reservoir operations. By applying the predictive modeling framework, reservoir engineers and managers can make more informed decisions about well placement, production rates, and reservoir management strategies, leading to improved production rates and reduced costs. Keyword: Artificial Neural Networks (ANNs), Support Vector Machines (SVMs), Random Forest, XGBoost (Extreme Gradient Boosting), Linear Regression, Gradient Boosting Machine (GBMs).