Mechanics of Bubble Flow in Heavy Oil Reservoirs

Abstract

SPE Members Abstract A new mechanism of fluid flow in solution-gas drive heavy oil reservoirs is identified through experimental studies. This paper presents experimental results in order to verify previous hypothesis on fluid flow in heavy oil reservoirs in Canada. It has been postulated that solution-gas drive in these reservoirs involves simultaneous flow of gas and oil. However, gas remains in tiny droplets under reservoir conditions. A new mathematical model was proposed in order to describe peculiar pressure-dependent multiphase flow properties. peculiar pressure-dependent multiphase flow properties. This paper presents experimental validation of some of the hypotheses offered by Smith. More than 40 experimental tests are performed both in a capillary tube and in a core packed with unconsolidated reservoir fluids and sands. The effects of decreasing bubble size on total fluid production and pressure drop across a pipe are observed. Experiments with live oil and reservoir sands enable one to quantify the contribution of bubble flow in solution gas drive process involved in heavy oil reservoirs. Some of these results help explaining the anomalies observed in heavy oil reservoirs in Canada. Introduction Elkins et al hypothesized the existence of 'worm hole' porosity in the unconsolidated Sands held together only porosity in the unconsolidated Sands held together only by the viscous oil. This hypothesis has often been supported by sudden collapse in injection schemes or failure of drilling and workover operations in heavy oil reservoirs of the Lloydminster area. This concept has often been used to introduce a negative skin factors in modelling heavy oil reservoirs. While this technique has given satisfactory results in several occasions, the technique has been criticized to be scientifically inaccurate. In fact, Islam and George have demonstrated through laboratory experimentation that sand redistribution or even sand/fines removal would actually decrease the near wellbore permeability. Smith suggested that solution gas drive in Canadian heavy oil reservoirs involves simultaneous flow of many oil and tiny gas bubbles. He proposed a new model for predicting performance of heavy oil reservoirs in Canada. predicting performance of heavy oil reservoirs in Canada. Kennedy and Olson studied bubbles of methane in kerosene in the presence of silica and calcite crystals. They observed that bubbles formed on silica or calcite surfaces rather than in the oil medium itself. They found that the number of bubbles formed for a given volume of reservoir rock depends on the rate of diffusion of gas through oil and on the rate of pressure declined in the reservoir. One of their important observations was that the variation in gas distribution may lead to different relative permeabilities for the same gas saturation in the reservoir. In an effort to explain discrepancy between laboratory and field observations, Stewart et al. conducted a series of solution gas drive tests in limestone cores. They observed that the oil recovery depends directly on the number of bubbles produced. The number of bubbles, in turn, depends on the rate of decline of pressure. Since, usually laboratory tests are performed at a pressure decline which is much higher than that in the pressure decline which is much higher than that in the field, laboratory tests consistently lead to higher recovery performance. performance. Hunt and Berry presented both experimental and theoretical studies in order to determine the probability function of bubble nucleation time. They found that the mean rate of bubble formation increases rapidly with increasing supersaturation. They presented a theory for predicting bubble size as well as bubble concentration for a given rate of pressure decline. Wieland and Kennedy re-confirmed some of the previous observations. The found that a definite previous observations. The found that a definite supersaturation ranging from 14 to 25 psi can be imposed without forming any bubble. This range depended on the type of rock and fluid used. P. 495

Novel Techniques for Measuring Heavy-Oil Fluid Properties

“…Smith (1988) was the first to develop an apparent in-situ oil viscosity equation for two-phase foamy oil. Later Islam and Chakma (1990) calculated a mixture viscosity to account for microbubbles in the oil using the volume fraction of oil and the single-phase viscosities of the gas and oil. Poon and Kisman (1991) indicated that foamy oil exhibits non-Newtonian flow, and shear thickening effects account for the low viscosity far from the wellbore where the shear rates are low.…”

Venezuela-MEM/USA-DOE Fossil Energy Report IV-11: Supporting technology for enhanced oil recovery - EOR thermal processes

“…The bubble size seems to be smaller (Islam, 1990) and the rate of produced bubbles higher (Kennedy & Olson, 1952, Hunt&Berry, 1956, Moulu, 1989) if the supersaturation is higher. Since both nucleation and diffusion of dissolved gas are driven by supersaturation, this might indicate that when the supersaturation exceeds some limit value, the nucleation rate increases stronger than the diffusion.…”

“…The mechanics of bubble flow in heavy oil reservoirs is discussed by Islam (1990), who performs experiments on the flow of a foamy oil in capillary tubes and core flow experiments. Although supersaturation is not directly discussed in this paper, Islam finds that at higher pressure decline rates, smaller bubbles are found and higher recoveries.…”

Foamy-Oil-Viscosity Measurement

“…This reduction in the asphaltene content in oil reduces the live-oil viscosity (Claridge 1994). However, different experimental studies suggest either an increase or a decrease in the apparent oil viscosity in the presence of bubbles (Bora 1998;Fisher et al 1999;Goodarzi et al 2005;Islam and Chakma 1990;Albartamani 2000). Thus, the effect of dispersed bubbles on heavy-oil flow properties in porous media is still debatable.…”