The importance of relative permeability curves in calculations of reservoirperformance can hardly be overstated. Both oil-water and oil-gas two-phaserelative permeability functions are crucial parts of the input data for heavyoil reservoir simulation studies of primary depletion and thermal recoveryprocesses. The simulation results are often very sensitive to small changes inthe relative permeability curves. Therefore, substantial effort and resourcesare spent in obtaining laboratory measurements of relative permeability. However as is often the case, when the history match is unsatisfactory, theparameter which is most likely to be adjusted is the relative permeability. Isit, then, futile to measure relative permeability in heavy oil systems?
This paper will attempt to answer this question by focusing on theoretical andexperimental problems in applying conventional relative permeability conceptsand methods to heavy oil systems. The experimental difficulties can start inthe very first step of obtaining a representative core sample. The core damageoccurring in coring and core handling operations can have significant effect onthe measured relative permeabilities.
A brief review of how heavy oil/water (or heavy oil/gas) relative permeabilityof unconsolidated sand formations is measured in the laboratory will bepresented with emphasis on possible sources of errors. The role of viscousfingering in relative permeability measurements will be explained. Conditionsunder which the laboratory measurements may not represent field behaviour willbe discussed.
It will be shown that the laws governing simultaneous flow of heavy oil andwater (or heavy oil and gas) through porous media have not yet been clearlyformulated. The flow description based on the relative permeability concept, although not as convenient and powerful as it is in the conventional oilsystems, is still the best available alternative.
The concept of relative permeability stems from the empirical extension of Darcy's law to two-phase (or multiphase) flow in porous media. It is the ratioof the effective permeability of a porous medium for a given fluid to theabsolute permeability of the same porous medium. Thus the relative permeabilityis a measure of the fractional loss of conductivity' for a given fluid due tothe presence of other immiscible fluids in the porous medium. The two-phaserelative permeabilities are assumed to depend on the saturation, wettabilityand pore structure but not on the fluid viscosities, densities or flow rates.Bear1 examined the experimental and theoretical evidence supportingthis concept and concluded that it was a good approximation for all practicalpurposes. This conclusion is also supported by the extensive, and generallysuccessful, application of this flow model in reservoir engineeringcalculations. Numerical simulation of reservoir processes requires relativepermeability curves among other data. These curves are generally estimated fromlaboratory experiments involving two-phase flow in small samples of thereservoir rock. In view of the heterogeneous nature of petroleum reservoirs, the application of the experimental data measured on a small samples of therock to represent the fluid flow behaviour of the whole reservoir represents agiant leap of faith. Therefore, it is not surprising that many simulationengineers do not trust the laboratory measured relative permeabilitycurves.