Application of Machine Learning for Predicting Potential Hydrocarbon Zones in Wells – X Field, Northern Red River Basin

The explosion of machine learning (ML) models that are capable of analyzing big data shed new light on oil and gas industry. This paper investigates the application of machine learning algorithms to predict zones with high potential for hydrocarbon reservoirs based on well data. By using three well data of Well-A, Well-B and Well-C in the north of Red River Basin and using ML algorithms including decision tree (DT), random forest (RF), extreme Gradient Boosting (XGB), artificial neural network (ANN), Self-Organizing Map (SOM), the attribute sets of hydrocarbon reservoir are determined by using unsupervised classification. After that, supervised classification is used to predict the potential hydrocarbon zones based on that attribute sets. Mentioned algorithms are compared with accuracy, and the results show that SOM displays very satisfactory results and predicts additional potential hydrocarbon zones. The paper also compares the performance of various models and proposes an optimization approach that integrates geophysical parameters and data preprocessing techniques to improve prediction reliability.