Investigation of the VAPEX Process Using CT Scanning and Numerical Simulation

Abstract

Abstract The "VAPEX" process, a solvent analogue of Steam Assisted Gravity Drainage, has attracted considerable attention as a recovery method for heavy oil. However, to date, there are still many questions about the nature and magnitude of basic process mechanisms, and whether the process can produce economic oil rates. The experiments discussed in this paper were aimed at quantifying some of the basic mechanisms, in particular the dispersive mixing mechanism. We have performed a series of topdown solvent injection experiments under varying conditions, utilizing a CT scanner to monitor fluid movements. All of the displacements we have observed are gravity-unstable in the early stages, and characterized by viscous fingering of the solvent into the 5,500 cP oil. After solvent breakthrough, the displacements become stable, dominated by a single solvent finger which has many of the features of a VAPEX solvent chamber. The "mixing parameter" we infer for these experiments using the Butler/Mokrys analytic model is higher than that reported for Hele-Shaw VAPEX experiments. An analysis of localized fluid velocities in the experiments using numerical simulation shows that the enhanced mixing parameter can be understood as a consequence of convective dispersion in the porous medium. By adjusting the amount of physical dispersion, the simulations can match breakthrough time, post-breakthrough oil rates, and the general character of the fingering. A novel type of "quasi-pore scale" simulation grid appears to provide advantages in simulating the unstable period at the beginning of the displacements. Introduction Compared with steam-based processes such as Steam Assisted Gravity Drainage (SAGD) for recovery of heavy oil, solventbased processes offer the possibility of reduced energy consumption and greenhouse gas production. However, they are mechanistically complex, and questions remain regarding their expected performance. To date, no field data are publicly available to answer these questions. One solvent-based process that has been proposed is the Vapour Extraction (VAPEX) process(1, 2). This solvent analogue of SAGD utilizes gravity as the driving agent, and solvent dilution of the heavy oil as the mobilization mechanism. The concept of the process is illustrated in Figure 1. A practical, solvent-based recovery process will depend for its success on the interplay of a number of phenomena. Some of the most important of these are: diffusion/dispersion, viscous fingering, capillary-driven mixing (in the case of a gaseous solvent), and the effects of reservoir heterogeneity. The first three are accessible for study in the laboratory, and understanding their interplay at the laboratory scale is a first step toward predicting their effects in a field process. Numerical simulation is required both to extrapolate laboratory experience to the field scale, and to incorporate the effects of reservoir heterogeneity. The study described in this paper addresses the phenomena of diffusion/dispersion and viscous fingering based on a series of laboratory experiments, combined with numerical simulation. Our initial experiments utilized a liquid solvent; therefore capillary mixing effects were absent. Future work will extend the results to gaseous solvents.