COMPARING HYSTERESIS MODELS FOR RELATIVE PERMEABILITY IN WAG STUDIES

Abstract

Immiscible WAG have been simulated by use of different relative permeability hysteresis models. The oil relative permeabilities were generated by a modified Stone I model. Experimental results especially from intermediate wetting systems have shown a significant drop in gas relative permeability between primary processes (gas saturation increasing first time) and tertiary processes ( gas saturation increasing after an increasing-de creasing sequence). Numerical simulation of both ID core-floods and 2D reservoir cross-sections were used in order to investigate WAG. The input to each model is selected from experimental measured data. The results show that standard hysteresis models for non-wetting phase relative permeability is lacking flexibility to describe experimental data important for the performance of the WAG process. INTRODUCTION Relative permeabilities are generally functional dependent of saturation and saturation history. The second dependency is in literature described as relative permeability hysteresis. The hysteresis behaviour in non-wetting phase (gas) relative permeability differs significant depending on wetting preferences of the system being investigated. Strongly water-wet systems show drainage-imbibition hysteresis as documented by many investigators in the literature1'2. In addition, intermediate-wetting systems show a complicated hysteresis behaviour depending on saturation cycle history3. Many reservoirs have intermediate-wetting properties, and a detailed study of the relative permeability hysteresis is important in processes involving saturation oscillation during three-phase flow. The experimental investigation3, which forms a basis for this paper, numbered the displacements processes with primary, secondary and tertiary depending on the number of forgoing displacements cycles. In this way a tertiary drainage is an increasing gas saturation process after a secondary imbibition (decreasing gas saturation) and primary drainage. Oil saturation was decreasing in all displacements and water, and gas was the injected fluids. This investigation shows that in order to describe the WAG process, a set of both primary and secondary drainage water and primary and secondary drainage-imbibition gas relative permeabilities have to be measured. Then an ambiguous problem arises, because standard two-phase hysteresis models can not reproduce all the experimental information. The scope of this paper has been to show that the information lost when using two-phase hysteresis models can be important for the performance of the WAG process. A hysteresis loop exists of a drainage process with a following imbibition process. It is frequently observed higher relative permeability to gas from a primary drainage than in a tertiary drainage process. The reduction of endpoint permeability can be as large as a factor of ten. These two processes are separated by a water-flood, indicating that trapping of non-wetting phase can have an important place in the hysteresis observed in intermediate wetting systems. The trapping