Simulations of Alkali-Polymer Experiments: Modeling of In-Situ Emulsion Generation and Transport of Oil-In-Water Emulsion in Porous Media

The injection of alkali in acidic viscous oils is known to promote the in-situ formation of emulsions during chemical oil recovery. Naphthenic acid components react with the alkali to form in-situ surfactants, which support oil emulsification at the water-oil interface. It is believed that emulsification and transport of the dispersed oil in the presence of polymer can significantly improve oil recovery. In earlier work, we proposed a new mechanistic non-equilibrium model to simulate alkali-polymer processes for different oil viscosities (2000 – 3500 cP at 50°C) with an acid number of around 4 mg KOH/g. The model considers emulsion generation kinetics, polymer, and emulsion non-Newtonian viscosity through a straightforward modelling strategy. The emulsified oil was treated as a dispersed component in water phase (O/W emulsion), while the water phase mobility considered the apparent aqueous phase viscosity containing dispersed oil and polymer. In the above referenced work, seven alkali-polymer corefloods performed with different alkali types and slug sizes were history matched. We showed that the model is capable of appropriately matching the experiments. Kinetics obtained by history match show that emulsion formation under the conditions here studied is alkali type dependent. In the current work, we applied our alkali-polymer model in two displacement tests (Hele Shaw cell) with two different oil viscosities (2000 – 200 cP at 50°C). These new experiments included secondary water flood, tertiary polymer flood and quaternary alkali-polymer flood. The initial conditions of alkali-polymer (AP) flood were obtained after properly modelling the unstable immiscible floods and polymer floods. For modelling the polymer floods (2D slabs), three models were evaluated: 1) extension of relative permeability curves applied to water flood, 2) Killough method (hysteresis for the water phase) and relative permeability power-law extensions and 3) two relative permeability curves with polymer concentration dependency. Our alkali-polymer model was employed for simultaneously history matching 1D and 2D experiments performed with 5 g/L of Na2CO3 and polymer. When comparing alkali-polymer results, a good agreement was found for the complete set of experiments. In addition, fitting parameters (kinetics and emulsion viscosity) were close to the parameters reported in the earlier study. Finally, fitted alkali-polymer parameters were employed for predicting alkali-polymer outputs in the second slab (with similar alkali-polymer concentration but lower oil viscosity). Even if experimental observations are relatively well represented, a lower value of incremental oil recovery (<3 % OOIP) was obtained. We believe that the use of a less viscous oil (diluted oil) in the experiments may influence the generation and transport of formed emulsions.