Abstract
Two-phase relative permeability (Kr) measurements (i.e., gas-oil and water-oil) are key input data to construct three-phase relative permeability models. These models are important for the understanding of many enhanced oil recovery (EOR) processes such as water-alternating-gas (WAG) and tertiary gas injections. The main objective of this paper is to study flow behavior of two-phase relative permeability using different fluid pairs in an unconsolidated sandstone reservoir.
A representative sample was selected from the main rock type in the reservoir to measure steady-state relative permeability at full reservoir conditions using live fluids and in-situ saturation monitoring. The following sequence of measurements were followed in the same sample without changing the experimental setup to obtain relative permeability data using different fluid systems:
Primary drainage gas-water relative permeability down to irreducible water saturation (Swi)
Replacement of gas by oil injection at Swi
Wettability restoration and effective oil permeability (Ko(Swi))
Secondary drainage gas-oil relative permeability down to residual oil saturation (Sorg)
Replacement of gas by oil injection and Ko(Swi)
Imbibition water-oil relative permeability down to residual oil saturation (Sorw)
Flow-through cleaning and 100% water saturation by miscible flooding
Primary drainage oil-water relative permeability down to Swi and Ko(Swi)
The relative permeability data were thoroughly evaluated, and quality checked from pressure data, volumetric production and in-situ saturation monitoring. The effective oil permeability (Ko(Swi)) was reproduced at the start of the secondary drainage and imbibition floods, as well as the end of the primary drainage, which gave more confidence in the data quality. Both the gas injection experiments (i.e., drainage gas-water and gas-oil Kr) gave very similar behavior, which could help predict one from the other. The primary drainage oil-water Kr gave higher oil permeability (in water) than the gas permeability in both the water and the oil phases in the gas injection experiments. The oil mobility in the gas flood was significantly higher than the oil mobility in the water flood even though both floods gave the same oil recovery. The drainage and imbibition oil-water Kr curves showed large hysteresis in the oil phase while no hysteresis was measured in the water phase.
The relative permeability experiments were numerically simulated by direct pore-scale modeling in digital rock models, which were obtained on rock samples representing lower and higher permeability ranges in comparison with the physically tested sample. The digital rock modeling allowed to obtain data for the most permeable rock sample, which was not possible to test in the physical SCAL due to its highly unconsolidated nature. The digital Kr curves compared reasonably well with the physical data and provided in-depth interpretation of the different flow characteristics for the different fluid systems. It was possible to reach relatively low residual oil saturation (i.e., ~15%) from the gas-oil relative permeability test despite the high oil/gas viscosity ratio. The results in this paper emphasized the importance of relative permeability measurements in different fluid systems, which can be incorporated in three-phase relative permeability models for improved understanding of reservoir performance in different EOR schemes.
