Modification of the Relative Permeability Functions through Oil/Gas Displacement Tests under Near Miscible Conditions by Using Inverse Modeling

An important step in the modeling of two-phase gas and oil flow is to have the reliable relative permeability functions. None of the previous conventional methods can predict the consistent and accurate relative permeability values to show the effect of near miscibility on the qualitative and quantitative behavior of relative permeability functions under near miscible conditions. The main contribution of this work is to select the optimum approach to relative permeability determination based on the classification of the available relative permeability methods and comparing their results in order to extract more accurate relative permeability values under near miscible conditions. In this work, the unsteady state displacement experiments were performed on two different reservoir rock samples (i.e. a dolomite and a sandstone core plug sample from the west of Iran). In addition, CO2 and light oil samples as injection fluids were used to decrease the variance error. Also, by the selection of the best relative permeability methods, the bias errors are decreased. In this study, inverse modeling is used with the Civan and Donaldson method as the initial guess for 1-D, two-phase flow simulation. The results show that the history matching as the optimum method presents a reformed relation in order to generate the relative permeability values more consistent to laboratory data under near miscible conditions. Finally, this reformed relation can be embedded in fluid flow simulators to simulate the near miscible gas injection more precisely.