(24c) Towards a Modeling Framework for Heavy Oil Phase Behavior and Properties

Heavy oil and bitumen are a challenging resource to recover and process mainly due to their high viscosity. Typically, heat is used to reduce the viscosity; for example, thermal processes are used to recover in situ heavy oil and a hot water process is used to extract bitumen from mined oil sands. These methods are energy and water intensive. Solvent-based or solvent-assisted processes are a potential alternative but require the understanding of more complex phase behavior. Data, characterization methods, and modeling methodologies are all lacking for mixtures of heavy oil and solvents; hence, the development of these processes carries significant risk.

A comprehensive characterization and modeling methodology for these fluids is under development at the University of Calgary. The approach involves new characterization methods and an equation-of-state based phase behavior model supported by new property correlations. A synopsis of this approach is presented with a focus on characterization.

A key challenge in characterizing heavy oils is the large non-distillable fraction of the oil. Conventional methods are based on the distillable fraction and properties must be extrapolated over the non-distillable fraction. For heavy oils, liquid-liquid equilibrium calculations are sensitive to the characterization of the medium fraction of the oil and asphaltene precipitation is sensitive to the characterization of the heavy fraction. Therefore, phase behavior predictions require a more accurate fluid characterization than can be obtained from simple property extrapolations originally developed for distillation based methods.

An alternative to extrapolation is to more rigorously characterize the non-distillable fraction of heavy oils. The idea is to fractionate the non-distillable portion of the oil using solvent-based methods and to measure properties of these fractions including vapour pressure, molar mass, density, and heat capacity. This data will be combined with GC, distillation, SimDist, and possibly PNA analysis to create a more complete representation of the oil suitable for modeling with equations-of-state.

All of the measurements except for vapour pressure are straightforward. A deep vacuum apparatus was designed to measure the vapour pressures of low volatility heavy fractions and possible to aid in the fractionation of the lighter end of the non-distillable cut. Preliminary characterization results using this approach are presented and the sensitivity of VLLE predictions to the characterization is examined.