More Insight Into the Pore-Level Physics of the Solvent-Aided SAGD (SA-SAGD) Process for Heavy Oil and Bitumen Recovery

摘要

The pore-level recovery mechanisms of the SA-SAGD process have been recently studied in the Porous Media Lab at the University of Waterloo using glass-etched micromodels. The experiments were conducted at controlled environmental conditions of an inverted-bell vacuum chamber to reduce the excessive heat loss to the surroundings. Different chemical additives (n-pentane and n-hexane) were added to steam prior to injecting into the models. Local temperatures along the model's height and width were measured and collected on a real time basis using a data acquisition system. An integrated data acquisition and control system was used to control, monitor and adjust the environmental vacuum pressure. The pore-scale events were videotaped and the captured snapshots were analyzed thoroughly using image processing techniques. The relevant pore-scale mechanisms responsible for the in-situ oil mobilization and drainage in a SA-SAGD process were addressed; transport and capillary phenomena at the pore- level were qualitatively documented including fluid flow, and heat and mass transfer aspects of the process. The pore-scale visualizations revealed that the gravity drainage process takes place within a thin layer of pores, composed of 1-5 pores in thickness, in the direction of gravity parallel to the apparent oil-vapour mixture interface in a so-called SA-SAGD mobilized region. The interplay between gravity and capillary forces results in the drainage of the mobile oil, whose viscosity is significantly reduced as a result of combined heat and mass transfer at the micro-scale. Heat transfer is believed to take place by conductive and convective mechanisms at the pore-level. The solvent content of the injected vapour mixture diffuses into the oil phase, hence reduces its viscosity following dilution as a result of molecular diffusion as well as convective mass transfer. The visualization results demonstrated the formation of water-in-oil emulsions at the interface because of the condensation of steam. The extent of emulsification depends on the temperature gradient between the gaseous mixture and the mobile oil phase. Water in oil emulsion is formed due to the non-spreading nature of water over the mobile oil phase in the presence of a gas phase. Asphaltene precipitation was observed when the condensed solvent reached the bitumen interface. Other pore-scale phenomena include localized entrapment of steam and solvent vapour within the continuum of the mobile oil at the interface due to capillary instabilities followed by subsequent condensation, relatively sharp temperature gradient along the SA-SAGD mobilized region, and snap-off of liquid films. In the absence of direct measurement of production data, the average horizontal advancement velocity of the apparent SA-SAGD interface was measured and was correlated with system parameters such as operating temperature, macroscopic and pore-scale properties of porous media, and heavy oil properties within the range of experimental conditions. This average sweep rate of the SA-SAGD process, along with the ultimate recovery factor values at the end of each particular test were considered as representatives of the SA-SAGD process performance at the pore-scale. Normal hexane was found to be a more effective steam additive compared to n-pentane at similar operating conditions. Increasing the solvent content in the injecting vapour mixture accelerates the recovery process at the pore-scale, and results in greater ultimate recovery factor values. When all other experimental variables are remain unchanged, the smaller the in-situ oil viscosity is, the greater would be the horizontal sweep rate and the ultimate recovery factor value. The pore-level interface advancement velocity was found to be a function of the pore-scale characteristics of the porous media. Different pore-scale properties such as pore-to-pore distance, pore body width, pore throat width, and diffusion distance affect the measured horizonta