|
Abstract
|
Based on world experience, determining the best enhanced oil recovery (EOR) method on a specific oil field can save time, reduce the costs of a project, and also increase the production rate and thus increase the profitability of an oil project. EOR methods screening is considered a type of technical and economic feasibility to choose the best option for the EOR method in hydrocarbon reservoirs. This study proposes EOR screening to determine the best EOR methods for the candidate field. Several new models are introduced based on the database collected from the literature along with rock and fluid properties and some other relevant reservoir characteristics. In this regard, different machine learning (ML) techniques like support vector machine (SVM), random forest (RF), LogitBoost, neural network (NN), and fuzzy logic (FL) algorithms have been used for modeling. In addition to the coding done, EORgui software has also been used to check the validity of the research. As a result of these investigations, the comparison of ML methods and the proposal of the most promising method with high performance in determining the most appropriate EOR method are provided according to the characteristics of the studied field which includes four reservoirs: X1, X2, X3, and X4.
|