No abstract

An experimental study of Athabasca bitumen/water relative permeabilities revealed little or no temperature effect on the relative permeabilities to water and bitumen over a range of 100 to 250°C [212 to 482°F]. Comparable results were obtained with both steadyand unsteady-state relative permeability measuring techniques. It was determined that the oil-phase relative permeability curve was convex. Measured curves were also compared with those obtained by history matching. IntroductionRelative permeability is one of the most important input variables for numerical reservoir simulation models. Knowledge of relative permeability curves and residual fluid saturations is needed to determine the oil production rate and ultimate recovery, respectively.Relative permeability generally has been represented as a function of saturation. Experimental measurement techniques developed in the 1950's and early 1960's for light-oil reservoirs include the steady-state technique and the unsteady-state, or dynamic, method. The latter has received more attention because of its faster turnaround time. Neither method seems superior, however, because of limitations inherent in each.Because of the experimental difficulties encountered in determining relative penneabilities, a suitable laboratory procedure that eliminates inlet and outlet end effects and resulting saturation gradients must be used. The experimental conditions usually are idealized and simplified by the use of small core plugs or reconstituted sand cores and rermed oils rather than native reservoir materials.The objective of this study was to measure experimentally the relative permeability to water and bitumen in unconsolidated sand cores. The difficulties of this task were compounded by the system's high temperatures and pressures and the bitumen's high viscosity . No published experimental data on residual fluid saturations, relative permeability curves, and the effect of temperature on these parameters exist for the Athabasca oil-sand system.'The temperature and pressure conditions encountered in the recovery of oil-sand bitumens (1.02 to 0.97 g/cm3 [8 to 14° API]) from deeply buried formations were reproduced in this study. Because of conflicting results in the literature regarding the effect of temperature on relative permeability, a simple system focusing on the fundamental properties that might cause temperature effects was used. The selected system consisted of unconsolidated silica sand, deionized water, and solvent-extracted Athabasca bitumen. Measurements were performed in the 100 to 250°C [212 to 482°F] temperature range. These idealized conditions minimized clay migration and high-temperature corrosion. The use of crude oil more closely represented the field conditions. In this way, experimental diffiulties were alleviated and the interpretation of relative permeability determination was eased.

The vapour-liquid equilibrium properties of methane-Cold Lake bitumen and ethane-Cold Lake bitumen mixtures were measured using a modified Ruska rocking cell apparatus. Data measured at three isotherms for these two pseudo-binary systems were used to develop a predicting method by means of equations of state. Two equations of state, the modified Soave-Redlich-Kwong and the Peng-Robinson, were chosen in this study. With an appropriate choice of bitumen characterization parameters and binary interaction coefficients, both of the equations of state can adequately represent the vapour liquid equilibrium properties of the two systems studied. Binary interaction coefficients of the modified Soove-Redlich-Kwong equation of state for the two systems were determined and correlated with temperature. Introduction Methane and ethane are commonly found in native Cold Lake oil sands bitumen, and are considered as possible additives for steam-based in situ bitumen recovery methods. Thus, the phase behaviour of methane-Cold Lake bitumen and ethane-Cold Lake bitumen mixtures under in situ conditions is important for reservoir engineers to determine the recovery of bitumen from Cold Lake deposits, as well as for process engineers to develop an adequate numerical simulation model Experimental data for these two pseudo-binary systems are not reported in the literature, and therefore one of the major objectives of this study was to measure their vapour-liquid equilibrium (VLE) properties. Experimental measurements are time-consuming and costly for these bitumen-containing systems. For the purpose of data reduction it is desirable to find methods which can predict their VLE properties with high accuracy. In this investigation, two cubic equations of state were selected to represent phase equilibrium properties. These are the modified Soave-Redlich-Kwong (MSRK) equation of state(1,3) and the Peng-Robinson (PR) equation of state (4,5). A commercially available "EQUI- PHASE" software package developed by DB Robinson and Associates was applied in vapour-liquid equilibrium calculations for the PR equation of state. Calculations were also performed with the PR equation of state using bitumen characterization parameters developed by Fu et al.(10) and binary interaction coefficients determined in this paper. The VLE properties measured in our laboratory are compared to the values calculated for the two gas-bitumen systems using both equations of state. Experimental Aspects Apparatus A schematic diagram of the experimental apparatus, which was verified for VLE measurements in a previous study(7), is shown in Figure 1. It consists of a charging and discharging unit, a constant temperature bath with a rocking equilibrium cell, and a sampling and analysis unit. The heart of the apparatus is the equilibrium cell which is located in a constant temperature bath container as shown in Figure 2. During measurements, the rocking cell is driven by a motor while a stirrer circulates the oil in the bath and keeps it homogeneous. Using this design, temperature was tested up to 423.2 K with an accuracy of ± 0.01 K, and pressure was tested up to 13.8 MPa with an accuracy of ± 1.0 kPa.

In the recovery of bitumen, viscosity reduction becomes important, both below and above the ground. The addition of a liquid diluent is thought to break down or weaken the intermolecular forces which create high viscosity in bitumen (1) . The effect is so dramatic that the addition of even 5% diluent can cause a viscosity reduction in excess of 80%; thus, facilitating the in situ recovery and pipe line transportation of bitumen.The knowledge of the bitumen-diluent viscosity is highly important, since without it, calculations in upgrading process, in situ recovery, well simulation, heat transfer, fluid flow, and a variety of other engineering problems would be difficult or impossible to solve. This paper presents the development of a simple correlation to predict the viscosity of binary mixtures of bitumen-diluent in any proportion. AbstractThe viscosity model is an important component in enhanced oil recovery packages and, for pure bitumen, several accurate models are available. In this study, a simple correlation presented in an earlier publication is extended to predict the viscosity of bitumen-diluent mixtures, as well as the mass fraction required to reduce bitumen viscosity to pumping viscosity.In developing the viscosity model, viscosities of pure bitumen and diluent were used as the endpoints, and the diluent mass fraction was raised to a power of "n" (a viscosity reduction parameter) to account for the sharp drop in bitumen viscosity with increase in diluent mass fraction. The model was developed with 99 data points from three different bitumens and five diluents; spanning a viscosity range of 10 -1 to 10 6 mm 2 /s. The model was used to recalculate the viscosity and mass fraction values, and results compared with similar correlations by Cragoe and Chirinos. The best match was obtained with our correlation, with overall average absolute deviations of 12% and 5% for viscosity and mass fraction predictions, respectively. Predictions on data not used in developing the model showed an excellent match between experimental and predicted values, with an overall average absolute deviation of below 10% for viscosities of mixtures at 25˚ C, 60.3˚ C, and 82.6˚ C.

New scaling criteria for steam and steam additive recovery experiments are presented in this paper. The mathematical development of different scaling criteria for a variety of scaling options and their relative merits are discussed. Past papers have reported similarity groups governing the scaling of steam additive processes to be incompatible. The scaling criteria presently available for steam processes, which permit the use of the same fluids in the model as those found in the prototype, require high pressure models and different porous media. This causes difficulties in scaling properties which depend on pressure or the porous media. Methods are presented which, by relaxing the requirement of geometric similarity, allow the same fluids and the same porous media to be used to scale steam or steam additive processes/or horizontal reservoirs. These methods allow scaling of all properties which depend on pressure and temperature, such as the saturated steam properties. A set of similarity groups is derived by inspectional and dimensional analysis for the steam additive process. Relaxed sets of scaling criteria, particular to the major mechanisms of a process, are then determined. The relative merits and potential applications of the various approaches are compared with (hose published m the literature. A means of selecting, or developing. If necessary, an approach which best scales the major aspects of a particular recovery process IS outlined. Introduction Steamflooding is a known method for recovering oil from some heavy oil reservoir For a number of years, emphasis has been placed on improving the steam flooding process. Many laboratory studies have been conducted in an effort to achieve this goal. Some of the experiments were designed to study the various mechanisms of a process and to extend the results to make field predictions by using numerical simulators. Other experiments, referred to as scaled experiments, were designed allow the relative influence of the various mechanismsobserved in an experiment to be similar to that expected in the field and to permit interpretation of the results to predict field performance. It is difficult to satisfy all of the criteria required to design scaled experiments of steamflooding processes. Consequently, some of the scaling requirements must be relaxed. The choice of which requirements to relax will depend on the particular process being modelled. Scaling of the phenomena considered to be least Important to a particular process might be relaxed without significantly affecting the major features of the process. Background to Scaling Steam Processes Scaled model experiments have been used to predict the effect of various parameters on the production response of a reservoir undergoing steam injection. The parameters include injection rate, production pressure, slug size, completion intervals, well pattern geometry, steam additives, bottom water, reservoir heterogeneities, and steam quality. Scaled models are also used to calibrate numerical models and provide insight into the effects of mechanisms which may not be properly incorporated in numerical simulators. A number of approaches to scaling steam processes have been used by previous investigators.