Mechanisms of Miscibility Development in Hydrocarbon Gasdrives: New Interpretation

Abstract

Summary This study concerns the mechanism of miscibility development in multiple-contact-miscible (MCM) gasdrives. Many low-pressure, rich-gas floods, traditionally assumed condensing processes, are shown to be vaporizing or (upper) liquid-extraction drives. It is shown that the solvent phase behavior at reservoir conditions determines the mechanism of miscibility development and that the rich-gas drives in low-temperature reservoirs are mostly condensing processes, while the liquid-extraction mechanism prevails in high-temperature reservoirs. The equation-of-state (EOS) simulations support this new interpretation of rich-gas-drive mechanism. Direct experimental evidence is also provided. The misuse of pseudoternary diagrams and failure to recognize their limitation when applied to multicomponent systems are shown often to lead to a false indication of a condensing processing. This adversely affects interpretation of experimental data (such as interpretation of miscibility in slim-tube displacements), and thus miscible solvent design. Introduction A state-of-the-art review of MCM (dynamic) gasdrives has been presented by Stalkup. Conventionally, the processes are presented as two extremes. On one hand, there are lean-gas drives with natural gas or N2 applicable at high pressures to crude oils rich in intermediates and involving mass transfer (vaporization) of intermediates from crude oil into the solvent. The miscible zone thus forms in forward contacts. On the other hand, the floods with rich solvents are assumed to be condensing gasdrives, where the miscible zone forms in swept-zone contacts and where the intermediate components transfer from rich solvent into the intermediates-lean crude oil. There has been some indication in the past that the mechanism of rich-gas drives may be somewhat more complex. More recently, these processes have been described as "combined condensing/vaporizing" drives. The latter study concluded that true miscibility (in a phase-behavior context) may not actually be developed and that pseudoternary diagrams cannot be used to explain or to predict miscible displacements. Our previous investigation of the mechanism of rich-gas drives has shown that true miscibility does develop, but often through a mechanism different from the two conventional processes described above, both of which involve mass transfer of intermediates. The miscibility development is often through an (upper) liquid-extraction mechanism, loosely referred to as a vaporizing drive. Our study also identified factors controlling the mechanism of miscibility development and explained why the liquid-extraction mechanism was more prevalent than the traditional condensing gasdrives. This was the first time that the liquid-extraction mechanism of miscibility development was identified in rich-gas drives. This paper complements our earlier study. It offers further insight into the phase behavior aspects of MCM rich-gas drives and also presents direct experimental evidence of the liquid-extraction mechanism. Contrary to the recently published study, the pseudotenary diagrams are found to be a valuable tool in analysis of multiple-contact processes and in phase behavior interpretation in general, provided the diagram limitations are recognized. Furthermore, analysis of the effect of crude oil light ends on miscibility development under different mechanisms clarifies why incorrect conclusions regarding process mechanism and ternary diagram usefulness were reached in the recent study. The Peng-Robinson EOS was used throughout this investigation to analyze phase-behavior trends. While we recognize that any EOS fluid model is nonunique, the conclusions of the study are independent of EOS model assumptions. The main result, frontal miscibility development, was predicted regardless of what EOS was used, as long as the pseudoternary envelope was constructed in a multiple-contact path and the C6+ fraction was not lumped into a single component. We were not concerned with the EOS's ability to describe the system phase behavior in detail, but focused on the basic understanding of mass-transfer mechanism because, without this understanding, the EOS would be a meaningless tool. The agreement between the EOS model and the experimental data is briefly addressed in the Discussion section. Liquid-Extraction Mechanism In Rich-Gas Drives Design of a rich solvent for a miscible flood in an Alberta reservoir is used to highlight the novel interpretation of the mechanism of some rich-gas drives. Table I shows the composition of the reservoir fluid (Fluid A). Also shown are compositions of the dry-gas and natural-gas-liquids (NGL) streams blended to form a solvent. The reservoir temperature is 210 degrees F [372 K], and the design pressure is 3,300 psi [22.75 MPa]. The slim-tube displacements of Fluid A with solvents of various NGL contents conducted at the design pressure indicated that to obtain oil recoveries of more than 90% of oil in place required a solvent NGL content well over 50 %. Thus, given the low content Of C2 through C5 intermediates in the crude oil (28 %), the relatively rich solvent, and the relatively low design pressure, the process would appear to be a condensing gasdrive. The miscibility development between Fluid A and solvents of various NGL contents was investigated with the EOS. The C6+ fraction of the fluid was described in terms of four pseudocomponents, with the weight fraction of the heaviest pseudocomponent approximately equal to the weight fraction of asphaltenes. Both the extended chromatographic analysis and Fluid A data in Table 1 were used in the EOS model development. The model was tuned by matching PVT and vapor/liquid equilibrium data. In investigation of the miscibility development, a series of chain flash calculations was performed where equilibrium vapor after each flash. step was contacted. with intact crude oil (forward contacts) and the equilibrium liquid in each flash step was similarly contacted with fresh solvent (swept-zone or backward contacts). Such a series of contacts approximates batchwise what happens during displacement at the front and back of the transition zone. The calculated equilibrium phase compositions were plotted as tie-fines in a pseudotenary diagram. Fig. 1 shows the calculated two-phase envelopes for three increasingly rich solvents (the solvents are referred by their NGL/dry gas molar ratio; the 75/25 solvent thus contains 75% NGL). The ternary envelopes for the 50/50 and 70/30 solvents (Figs. 1a and 1b) indicate immiscible behavior. Constriction in the two-phase envelopes of the richer 70/30 solvents, however, indicates a region of lowered liquid/vapor interfacial tension (IFT) and thus the possibility of increased displacement efficiency. (The IFT's were calculated with the parachor method.) It is important to note that most of the two-phase region shown was developed in forward contacts; swept-zone contacts provided only a small part of the phase envelope close to the limiting tie-line passing through the solvent composition and corresponding to liquid stripping. The IFT vanished for the 75/25 solvent (Fig. 1c), and miscibility was achieved. The two-phase envelope in Fig. 1c, however, appears to be in a different area of the ternary space than usually shown for either a condensing or a high-pressure vaporizing miscible drive. The phase envelope represents development of miscibility in forward contacts. SPERE P. 341^