Grouping method to optimize oil description for compositional simulation of gas injection processes

Numerous grouping methods have been proposed to reduce the computational time required for phase behaviour calculations in a compositional reservoir simulator. In gas injection processes where significant mass transfer occurs between the injection gas and the reservoir oil, the use of pseudo-component properties generated from the original oil composition may be insufficient to accurately predict the phase behaviour at some conditions. The paper presents a method to model the progressive phase behaviour of injection gas and reservoir oil systems, with reduced computational requirements by optimising the compositional description. Multistage experiments simulating progressive compositional variations of injection gases and reservoir oils have been conducted, and compositions, volumes, and densities of both phases at equilibrium measured. The experimental data have been used to evaluate the proposed grouping method. The method allows the retrieval of full compositions with very little effort and almost no extra computational time, to select new groups for further calculations. A new mixing rule, based on the parameters of an equation of state has also been developed and incorporated in the grouping method. It is shown that there exist an optimum number of groups describing a reservoir oil in any gas injection process. Reduction of components above the optimum value does not impair the prediction accuracy, whilst below that the results become highly unreliable.