The solvent based recovery process "Vapex" has a great potential for the recovery of heavy oil and bitumen resources, due to the low energy intensity and reduced GHG emission associated with the process. The process has been extensively investigated in the laboratory models and through numerical simulation studies. This has indicated the technical viability of the process as an alternative to thermal recovery processes, viz. SAGD. In the last few years several proprietary pilot has been developed to establish the commercial viability of this concept.

The most important uncertainty about this process is the rate of mixing of solvent molecules with high viscosity hydrocarbons. Due to low molecular diffusivity the theoretical predictions of extraction rates are significantly lower than the SAGD process. However, the results of physical model experimental carried out by this author showed a considerably higher mass transfer rate at the solvent bitumen interface. The process, whether in the laboratory model or in the reservoir porous media, takes place in a microscopic level. An experimental evidence of this phenomenon will be presented in the paper.

Transferring this microscopic phenomenon into a macroscopic simulation model presents a serious challenge. Artificially higher diffusion or dispersion coefficients are used to match the experimental data. Even with that both the production rate and the solvent saturation profile can not be matched simultaneously. For example the higher dispersion coefficient results in deep penetration of the solvent, resulting in a diffusion zone, thicker than the experimentally determined value. Lower dispersion coefficient results in a lower production rate. Some of these simulation results are presented in this paper. A recent development in the simulation model, Dynamic Grid Refinement, improves the simulation match by allowing the use of smaller grid blocks at the diffusion boundary layer.

You can access this article if you purchase or spend a download.