An Overview on Applications of Machine learning in petroleum Engineering

Predicting production, reservoir characterization, petrophysical interpretation and drilling operation have always been a challenge to petroleum engineering. Data-driven modelling (DDM), provides the procedures for evaluating and realizing the relationships amongst the state of the system features excluding physics-based model behavior. Intelligent computations use the theoretical basis for developing Genetic Algorithms, fuzzy rules-based systems, and artificial neural networks. The primary phase in a data analytics cycle is data management and collection. For analytical purposes, different data forms are gathered from various sources. This study demonstrates the feasibility study of application of data-driven machine learning algorithms, integrating geoscientific, distributed acoustic sensing (DAS), distributed temperature sensing (DTS) fiber-optic, completions, flow scanner production log, and surface in petroleum engineering. We demonstrated the supervised data-driven machine learning models using Random Forest (RF), Artificial Neural Network (ANN), and Support Vector Machine (SVM) algorithms to understand the well performance and forecast the daily oil and gas production and petrophysical interpretation.

Keywords:

Data-driven modelling , machine learning , Random Forest , Support Vector Machine , Artificial Neural Network , production log , reservoir