Summary
Measurements of the interfacial tension (IFT) of mixtures of a reservoir fluid and injection gas at various pressures have been proposed as an experimental method for predicting the minimum miscibility pressure (MMP) in an experiment referred to as the vanishing-IFT (VIT) technique. In this paper, we analyze the accuracy and reliability of the VIT approach using phase equilibrium and slimtube experimental observations and equation-of-state (EOS) calculations of the behavior of VIT experiments for the same systems.
We consider 13 gas/oil systems for which phase equilibrium and density data and slimtube measurements of the MMP are available. We show that tuned EOS characterizations using 15 components to represent the gas/oil systems yield calculations of phase compositions and densities and calculated MMPs that reproduce the experimental observations accurately. We assume that IFTs can be calculated with a parachor expression, and we simulate the behavior of a series of VIT experiments with different mixture compositions in the VIT cell. We show that compositions of mixtures created in the VIT cell are not, in general, critical mixtures and that calculated estimates of the MMP obtained by the VIT approach depend strongly on the composition of the mixture used in the experiment. We show also that those MMP estimates may or may not differ significantly from values obtained in slimtube displacements. Fortuitously chosen mixture compositions can result in VIT-experiment estimates that agree well with slimtube MMPs, while for other mixtures, the error of the estimates can be quite large. In particular, we show that errors in the VIT-technique estimate of the MMP are often large for gas/oil systems for which the first-contact miscibility pressure (FCMP) is much larger than the slimtube MMP.
We conclude, therefore, that the VIT experiment is not a reliable single source of information regarding the development of multicontact miscibility in multicomponent gas/oil displacements.
Introduction
Many oil fields are now candidates for enhanced-oil-recovery processes such as tertiary gasfloods or miscible water-alternating-gas injection schemes. The MMP is an important parameter in the design and implementation of these displacement processes and, hence, it is equally important that the MMP be determined by a method that is both reliable and accurate.
Several methods have been proposed for measurement of the MMP. The slimtube-displacement experiment is the most commonly used approach (Yellig and Metcalfe 1980; Holm and Josendal 1982; Orr et al. 1982). Because of the time-consuming process of performing multiple slimtube-displacement experiments, alternative experimental approaches have been proposed. Some investigators have suggested use of a rising-bubble experiment, in which observations of bubbles of injection gas rising through oil (Christiansen and Haines 1987; Eakin and Mitch 1988; Novosad et al. 1990; Sibbald et al. 1991; Mihcakan and Poettmann 1994), are a basis of a method for determining the MMP. Zhou and Orr (1988) concluded that the changes in bubble behavior observed in the rising-bubble experiment are caused primarily by changes in IFT as components in the bubble dissolve in the oil and components in the oil transfer to the bubble. They showed that rising-bubble experiments could be used to measure the MMP for vaporizing gas drives, but are less accurate for condensing gas drives, while a drop of oil falling through gas could be used to determine the MMP for condensing gas drives. Whether either a falling-drop or a rising-bubble experiment could be used to determine the MMP accurately in condensing/vaporizing gas drives such as those described by Zick (1986), Stalkup (1987), and Johns et al. (1993) has not been determined.
Rao and coworkers proposed a different use of IFT observations to determine the MMP (Rao 1997, 1999; Rao and Lee 2002, 2003; Ayirala et al. 2003; Ayirala and Rao 2004, 2006a, 2006b; Sequeira 2006). They measured IFTs for pendant drops of oil suspended in a cell containing a two-phase mixture of the injection gas and the oil. In that approach, known as the VIT experiment, the IFT is measured at a sequence of pressures, and the MMP is taken to be the pressure at which the IFT plotted as a function of pressure extrapolates to zero IFT. Orr and Jessen (2007) presented an analysis of the VIT technique based on EOS calculations for well-characterized ternary and quaternary gas/oil systems and demonstrated that the VIT experiment may give estimates of the MMP that differ significantly from the MMP based on critical tie-lines for condensing, vaporizing, and condensing/vaporizing gas drives.
In this paper, we extend the analysis of Orr and Jessen (2007) and calculate the IFT behavior that would be observed in the VIT experiment for gas displacements of multicomponent crude-oil systems. We assess the accuracy of MMP estimated by the VIT approach for 13 multicomponent gas/oil displacements for which experimental phase-equilibrium and slimtube data are available, and we demonstrate that for these multicomponent crude-oil systems, the VIT approach can give estimates of the MMP that are close to the actual MMP or that are significantly in error, depending on the compositions of mixtures created in the equilibrium cell.
Implementation of multilayer perceptron (MLP) and radial basis function (RBF) neural networks to predict solution gas-oil ratio of crude oil systems