Analytical and numerical solutions of chemical flooding in a layered reservoir with a focus on low salinity water flooding

Chemical flooding has been implemented and studied intensively as an EOR method. One such process, Low Salinity WaterFlooding (LSWF) has become increasingly applied. Simulations can be performed for these processes to predict behaviour and make field management decisions. These are costly and incur numerical errors. Analytical solutions to flow behaviour have been developed previously for waterflooding in reservoirs that consist of non-communicating layers. We extend that analysis here for chemical flooding, and in particular for LSWF. We also extend the analysis that we developed previously to include dispersion effects. We then compare the analytical predictions to the more realistic case of flow across communicating layers to assess crossflow effects. We derive a mathematical form of fractional flow theory for a set of non-communicating layers that can be used to predict fluid flow and solute transport including the location of waterfronts. This model corrects for the effects of numerical and physical dispersion. We examine the validity of this analytical model by comparing it to simulations of fluid flow behaviour in non-communicating layers first and then in communicating layers. We use dimensionless numbers that can be used to deduce the inter-layer relationships of the various fronts that form as a function of viscous crossflow. We examined models with different degrees of heterogeneity under various mobility ratios. The analytical method worked very well compared to numerical simulations in the absence of cross-flow. Our results show that for virtually homogenous reservoirs, the crossflow has negligible effect on oil recovery. For moderately heterogeneous reservoirs, the crossflow has a negative effect reducing the recovery factor. Cross flow resulted in varying effects ranging from a reduced ultimate recovery of 2% or increased it by 9%, relative to the original oil in place. The former occurred for models with a mobility ratio at the leading formation waterfront that was less than one combined with low heterogeneity while the latter occurred for highly heterogeneous cases.