Development of New Models for Predicting the Solution Gas Oil Ratio of Rmelan Crude Oils
Abstract
Solution Gas Oil Ratio (SGOR) is an important reservoir fluid property. Many reservoir engineering calculations are relied upon for this property, such as reserve estimation, fluid production and injection rates, secondary and enhanced oil recovery, and prediction of the future production performance of oil and gas reservoirs. In some circumstances, in vitro PVT analysis is unavailable, the experimental-based PVT correlations are the best alternative. In this paper, new models were developed to predict the SGOR for Rmelan crude oil using non-linear multiple regression and artificial neural networks (ANN). 50 PVT datasets were collected from Rmelan oil fields for this purpose. The collected data include the SGOR, oil-specific gravity, gas-specific gravity, reservoir temperature, and bubble point pressure. A new SGOR correlation was developed using a nonlinear multiple regression technique with an R2 of 0.81 Also, a new ANN model was developed for calculating the SGOR with an R2 of 0.97 and an average absolute relative error of 0.04. This study shows that the developed models are matched with Rmelan crude oil better than other published correlations. The developed models represent a road map for PVT properties of Rmelan crude oil saving time, effort, and money.
