Systematic Modelling of Unstable Displacement

Modelling unstable displacement is a challenge which may lead to large errors in reservoir simulations. Field scale coarse grid simulations therefore need to be anchored to more reliable fine grid models which capture fluid displacement instabilities in a physically correct manner. In this paper, a recently developed approach for accurately modelling viscous fingering has been applied to various types of unstable displacement. The method involves estimation of dispersivity of the porous medium and length scale of the model to determine the required size of the simulation grid cell. Fractional flow theory is then applied to obtain the correct saturation of the injected phase in the unstable fingers formed due to the adverse mobility ratio. Unstable displacement experiments have been history matched using 2D-imaging of in-situ saturation as a calibration of our method, before carrying out sensitivity calculations on the effect of fluid viscosity, and rock heterogeneity. Our modelling approach allows us to carry out simulations using a conventional numerical simulator using elementary numerical methods (e.g. single-point upstreaming). The methods used to model instability (Sorbie et al, 2020) was originally developed for immiscible water/oil systems. The current paper now presents new results applying this approach to unstable gas displacements, where adverse viscosity ratios may be even higher than in water/oil systems. The displacement with injected gas is shown to be influenced by mass exchanges between the gas and oil as the alternating fluids (water and gas) are injected in WAG processes. Swelling of fingers delay the gas front and WAG processes divert the injected gas and improve sweep efficiency. We have also modelled water-oil displacement at adverse mobility and shown the benefit which is obtained by reducing the instability by adding polymers to viscosify the injected water. The impact of rock heterogeneity has different effect depending on buoyancy forces and the degree of crossflow into the high permeable zones. This paper extends our novel approach to modelling the fine scale distribution of the injected fluids in adverse mobility ratio displacements. This approach has now been applied to both, gas/oil and water/oil systems where viscous fingering is present, either at extremely adverse mobility ratios and/or for reservoirs where the permeability field is very heterogeneous.