A Methodological Analysis of the Mechanisms Associated With Steam/Solvent Coinjection Processes Using Dynamic Gridding
Summary Hybrid steam-solvent processes have gained importance as a thermal-recovery process for heavy oils in recent years. Numerous pilot projects during the last decade indicate the increasing interest in this technology. The steam/solvent coinjection process aims to accelerate oil production, increase ultimate oil recovery, reduce energy and water-disposal requirements, and diminish the volume of emitted greenhouse gases compared with the steam-assisted-gravity-drainage (SAGD) process. Among the identified physical mechanisms that play a role during the hybrid steam/solvent processes are the heat-transfer phenomena, the gravity drainage and viscous flow, the solvent mass transfer, and the mass diffusion/dispersion phenomena. The major consequence of this complex interplay is the improvement of oil-phase mobility that is controlled by the reduction in the oil-phase viscosity at the edge of the steam chamber. It follows that a detailed representation of this narrow zone is necessary to capture the involved physical phenomena. In this work, a study of sensitivity to grid size was carried out to define the appropriate grid necessary to represent the near-edge zone of the steam/solvent chamber. Our results for the steam/solvent coinjection process in a homogeneous synthetic reservoir indicate that a decimetric scale is required to represent with a good precision the heat and mass-transfer processes taking place at the edge of the steam chamber. In addition, we present some numerical results of the adaptive dynamic gridding application. Comparison was performed between the SAGD process and steam/solvent coinjection after the characterization and analysis of the mechanisms that govern oil production under typical Athabasca oil-sand conditions. Finally, in the framework of the proposed numerical methodology, the effect of solvent type and injection conditions on the oil-recovery efficiency is quantitatively illustrated. Published data for similar applications are also discussed. It is expected that this work will provide some insight for the simulation community about methodological aspects to be taken into account when hybrid steam/solvent processes would be modeled.