A Novel and Robust Mixing Rule Model Coupled with Neural Network for Rapid Determination of Minimum Miscibility Pressure

Abstract

Miscible gas injection is one of the most effective enhanced oil recovery techniques. Minimum miscibility pressure is one of the most important parameters in the gas injecting process in oil reservoirs. Accurate determination of this parameter is critical for an adequate design of injection equipment and project investment prospects. In this study, 128 samples of experimental data are used based on a slim tube test. Effective parameters on minimum miscibility pressure are investigated to define independent variables. The mixing rules method is coupled with an artificial neural network to present a new model for simulating the slim tube apparatus. A comparison between the results of the proposed model and the other conventional methods indicated that it is more accurate and rapid in predicting minimum miscibility pressure. The new model yields the lowest average absolute relative error equal to 2.21%, and the lowest standard deviation of error equal to 3.03%. The proposed model is applicable for various injected gases, such as light hydrocarbons gases, pure and impure CO2, nitrogen, and flue gases.