A Novel Upscaling Approach to Enhance Polymer Flooding Modelling and Capture the Complex Flow Dynamics

Abstract Polymer flooding is a widely utilized enhanced oil recovery (EOR) technique that enhances sweep efficiency by injecting polymer solutions to modify fluid mobility. The effectiveness of polymer flooding is governed by key mechanisms, including mobility control, polymer-rock interactions, gravity segregation, viscous crossflow, and reservoir heterogeneity. These elements combined are critical in fluid displacement mechanisms. Gravitational forces induce gravity segregation, which affects the distribution of fluids within the reservoir and can significantly influence displacement efficiency and sweep patterns. Viscous crossflow mechanisms play an important role in determining flow behaviour and displacement efficiency in porous media, consequently impacting displacement patterns and recovery rates. Heterogeneous reservoirs display significant variations in permeability, adding complexity to fluid dynamics due to varying permeability levels affecting the overall efficiency of polymer flooding. Therefore, it is crucial to accurately capture these effects in simulation modelling. While high-resolution grid models are essential for accurately capturing these complex flow dynamics, their computational demands necessitate the implementation of upscaling techniques to achieve a balance between accuracy and efficiency. This study introduces a systematic upscaling methodology comprising four stages: the first stage involves employing single-phase upscaling to compute average permeability and porosity. Then, relative permeability is upscaled using a pore-volume-weighted-average method, accounting for changes in fluid viscosity due to polymer concentrations. The next stage focuses on regressing the sigmoid polymer viscosity multiplier factor to achieve a more precise polymer viscosity distribution in the reservoir, and the last stage involved tuning the Todd-Longstaff mixing parameter coefficient. The approach was evaluated using 1D and 2D case models within the E100 Black Oil simulator, incorporating homogeneous, layered, and heterogeneous reservoirs. The developed methodology was further extended to 3D models with randomly correlated permeability fields, allowing for an investigation of reservoir heterogeneity and anisotropy on polymer flooding performance. By varying the dimensionless correlation length, the study assessed the impact of these factors on fluid flow dynamics. The significantly improved alignment. In summary, proposed upscaling framework has successfully improved the matching of key parameters with fine-grid simulation results, particularly in key parameters such as oil recovery factor, water cut profiles, reservoir pressure, and polymer production trends.