Pore scale network modelling of residual oil saturation in mixed-wet systems

Abstract

The prediction of residual oil saturation (Sor) and relative permeabilities after waterflooding in mixed-wet systems is a very challenging task. These are important parameters which must be estimated for a full field simulation of waterflooding. The Sor also defines the target oil for any proposed EOR process after an initial waterflood. Pore-scale network modelling can be used to estimate relative permeabilities, and the amount and nature of the trapped residual oil if the correct physics of oil displacement are properly included. During the waterflooding of mixed-wet systems, oil may drain down to relatively low residual saturations. Such Sor levels can only be calculated correctly when oil layers in pore corners are included in the pore-scale modelling. van Dijke and Sorbie (J. Coll. Int. Sci. 293 (2006) 455) obtained an accurate thermodynamically derived criterion for oil layers’ existence in pores with non-uniform wettability caused by ageing, which is more restrictive than the previously used geometrical layer existence criterion. This thermodynamic criterion has been included in a newly developed two-phase pore network model to calculate realistic Sor values for mixed-wet sandstones. A new ncornered star pore shape characterization technique has also been implemented in this model since the precise description of the pore shape was found to be important. Two unstructured networks, derived from Berea sandstone have been used for a number of sensitivities of the Sor and relative permeabilities with respect to wettability conditions. It is shown that Sor is lower for the more strongly oil-wet cases, while the water relative permeability curves increase gradually with oil-wetness at the higher water saturations. It has also been shown that pore shape approximations and oil layers collapse criterion have a significant impact on the Sor and the relative permeabilities. In particular, the thermodynamic oil layer existence criterion gives higher and more realistic Sor compared to previously used geometrical criterion. The network modelling has been used to match experimental data for water-wet and mixed-wet systems. In particular, the good agreement with mixed-wet systems strongly indicates that using the correct oil layer existence criteria is a significant step forward in the reliable prediction of Sor.