Determination of gas–oil miscibility conditions by interfacial tension measurements

Abstract

Processes that inject gases such as carbon dioxide and natural gas have long been and still continue to be used for recovering crude oil from petroleum reservoirs. It is well known that the interfacial tension between the injected gas and the crude oil has a major influence on the efficiency of displacement of oil by gas. When the injected gas becomes miscible with the crude oil, which means that there is no interface between the injected and displaced phases or the interfacial tension between them is zero, the oil is displaced with maximum efficiency, resulting in high recoveries. This paper presents experimental measurements of interfacial tension between crude oil and natural gases (using a computerized drop shape analysis technique) as a function of pressure and gas composition at the temperature of the reservoir from which the crude oil was obtained. The point of zero interfacial tension was then identified from these measurements by extrapolation of data to determine minimum miscibility pressure (MMP) and minimum miscibility composition (MMC). The gas–oil miscibility conditions thus obtained from interfacial tension measurements have been compared with the more conventional techniques using slim-tube tests and rising-bubble apparatus as well as predictive correlations and visual observations. The miscibility pressures obtained from the new VIT technique were 3–5% higher than those from visual observations and agreed well with the slim-tube results as well as with the correlations at enrichment levels greater than 30 mol% C₂₊ in the injected gas stream. The rising bubble apparatus yielded significantly higher MMPs. This study demonstrates that the VIT technique is rapid, reproducible, and quantitative, in addition to providing visual evidence of gas–oil miscibility.

Introduction

The recent worldwide survey of enhanced oil recovery (EOR) projects conducted by the Oil and Gas Journal [1] indicates that even with the uncertainty caused by the oil price collapse prior to mid-1999, operators still see EOR processes as opportunities for increasing recovery factors from known oil accumulations. This survey lists 18 fieldwide or pilot gas injection projects planned to start between 2000 and 2001, which include 7 CO₂ miscible projects, 7 hydrocarbon miscible projects, 2 hydrocarbon immiscible WAG projects, and 2 nitrogen projects. The size of these projects ranges from 10 acres for the Sprayberry Trend in Midland County in Texas to 52,000 acres for PanCanadian's CO₂ flood in the Weyburn Unit in Saskatchewan. The survey concludes that EOR processes have weathered the low oil price environment of the last two years and continue to contribute significant oil production.

Terra Nova, discovered in 1984, is the second largest oil pool in the Grand Banks of the Canadian east coast. It is located 350 kilometers south-southeast of St. Johns, Newfoundland in 95 m of water depth. The initial production rate from this reservoir is anticipated to be about 115,000 barrels per day. The production strategies considered for Terra Nova included waterflood, solution gas injection for gas storage, and enriched gas injection for miscible flood.

In order to evaluate the potential for carrying out a miscible flood in Terra Nova, miscibility conditions were determined for Terra Nova live oil and several field gas streams of differing compositions using the slim-tube displacement test, the rising bubble apparatus, and the recently developed vanishing interfacial tension (VIT) technique. The results from these experimental investigations are presented and discussed in this paper.

Miscible gas injection has long been recognized for its ability to minimize the oil trapping effect of capillary forces and continues to be practiced as an economic enhanced oil recovery process in many parts of the world. By optimizing the enrichment of the injected gas and by operating close to the minimum miscibility pressure (MMP), the economics of the gas injection projects can be further improved. However, this requires accurate determination of miscibility pressure and composition in the laboratory. In order to attain accuracy in laboratory measurements, it is instructive to revisit the definition of miscibility. The following are some definitions of miscibility appearing in the literature. “Miscible displacement is a process where there is an absence of phase boundary or interface between the displaced and displacing fluids” [2]. “Two fluids are miscible when all mixtures of these two fluids in all proportions remain in a single phase without any interfaces, and consequently with no interfacial tension, between the fluids” [3]. “Miscibility is that physical condition between two or more fluids which permits them to mix in all proportions without the existence of an interface” [4]. “Two fluids that mix together in all proportions within a single fluid phase are miscible” [5].

The absence of an interface, or consequently the condition of zero interfacial tension, is evidently the fundamental criterion for attaining miscibility between the injected gas and the reservoir crude oil. However, the currently used experimental techniques, such as the slim-tube, the rising bubble apparatus (RBA), or the pressure–composition (P–X) diagram, do not directly answer this specific requirement of zero interfacial tension. On the other hand, the vanishing interfacial tension (VIT) technique [6], as the name implies, is based entirely on measurements of interfacial tension and therefore attempts to satisfy the fundamental definition of miscibility.

Several studies [7], [8], [9], [10], [11], [12] have reported on the comparison of slim-tube, RBA, and P–X techniques of determining gas–oil miscibility. Yurkiw and Flock [13] have provided a comparison of various correlations available in the literature for estimating miscibility pressure. In this paper, we compare the miscibility pressures and enrichments obtained from the VIT technique with those from the RBA, and slim-tube tests as well as the predictions from Benham et al. [2] and Kuo [14] correlations.