Toward an Intelligent Approach for Estimating Solution Gas Oil Ratio

One of the critical steps of evolutionary calculations of oil and gas reservoirs such as reserve and production, inflow performance and material balance calculations is the PVT properties calculation of reservoir fluids. Since laboratory-measured properties of fluid samples are only available from limited and isolated well locations and/or intervals, determination of these parameters is complicated task. Several models including empirical correlations, statistical regression and intelligent systems have been developed to relate these properties. In this study, a new model for prediction of solution gas-oil ratio (Rs) has been developed using more than 1100 experimental data series collected from literature sources. Least Square Support Vector Machine (LSSVM) modeling optimized with coupled simulated annealing (CSA) tool was used to develop a model for prediction of solution gas-oil ratio. Statistical and graphical error analyses indicated that the robustness and accuracy of the developed model are higher than other explicit and implicit published models.

کلیدواژه ها:

solution gas-oil ratio ، least square support vector machine ، crude oil