Predicting Minimum Miscible Pressure in Injection of Gas to Oil Reservoirs by Data-Based Intelligent Methods

CO2 injection processes are among the effective methods for enhanced oil recovery. A key parameter in the design of CO2 injection project is the minimum miscibility pressure (MMP). From an experimental point of view, slim tube displacements test routinely determines the MMP. Because such experiments are very expensive and time-consuming, searching for fast and robust methods for determination of MMP is usually requested. The Neural Network (NN) and Support Vector Machine Regression (SVM) were used to designs networks for estimation MMP. The Networks Trained by trusted data including independent variables. The validity of these new models were successfully approved by comparing the models results to the pure and impure experimental slim-tube CO2-oil MMP and the calculated results for the common pure and impure CO2-oil MMP correlations. The average accuracy of the predicted values for the neural network in terms of coefficient of determination (R2) and mean square error (MSE) are 0.9863 and 0.0018. These values for the SVM regression are 0.9870 and 0.0017, respectively. In addition, the new models could be used for predicting the impure CO2-oil MMP at higher fractions of non-CO2 components (Up to 100% for methane and 50% for hydrogen sulfide).