| Oil producers are beginning to consider solvent-based or solvent-assisted processes to recover heavy oil. To design these processes, it is necessary to predict the amount and composition of the phases that form when the heavy oil and solvent mix. In particular, the formation of asphaltene-rich phases can strongly affect the recovery process.;A new experimental approach was developed to calculate the composition of solvent in the asphaltene-rich phase. The experiments were based on the difference between the evaporation rate of the free or entrained solvent versus that of the solvent dissolved in the asphaltene-rich phase. The mass of a mixture of asphaltenes and solvent was measured over time and there was a clear and consistent change in slope when the entrained solvent evaporation ended. The amount of solvent in the asphaltene-rich was determined at this point.;A previously developed regular solution model was modified to include all components in the asphaltene-rich phase to take part in the phase equilibrium. The model was retuned using n-heptane/toluene/Athabasca asphaltene data. The model predicted reasonably good results for fractional precipitation of asphaltenes having different average molar mass, except for the low molar mass asphaltenes. The predictions for different n-alkane solvents were less accurate particularly for heavy solvents n-octane and n-decane. The model did not handle the effect of asphaltene concentration well. The revised model also significantly over-predicted the solvent content of the asphaltene-rich phase.;A previously developed regular solution model successfully predicted the amount of asphaltenes in dense phases formed in solutions of asphaltenes, toluene, and n-alkanes as well as in heavy oils diluted with n-alkanes. However, this model assumed that there is no solvent partitioning to the asphaltene-rich phase, which is thermodynamically incorrect. Furthermore, no compositional data for the asphaltene-rich phase was available with which to tune the model. |
