Adapted two-phase capacitance-resistance modeling of polymer flooding in synthetic and benchmark reservoirs with varying degree of heterogeneity.

Polymer flooding is a widely applied enhanced oil recovery method that improves sweep efficiency by increasing injected water viscosity and achieving a more favorable mobility ratio. Field-scale performance evaluation remains challenging when detailed reservoir data are unavailable, motivating the use of data-driven approaches such as Capacitance-Resistance Model (CRM). However, conventional single-phase CRM formulations neglect two-phase flow effects, limiting their applicability to polymer flooding. This study presents a two-phase CRM that explicitly incorporates polymer-induced viscosity variations and saturation-dependent mobility. The proposed approach was validated using two case studies, including a benchmark reservoir model (PUNQ-S3) and a synthetic reservoir using CMG STARS simulator. Results demonstrate that the two-phase CRM consistently improves predictive accuracy in synthetic reservoir cases, reducing forecast errors from 10 to 22% with the conventional CRM to consistently below 4% across both homogeneous and heterogeneous cases, demonstrating nearly a tenfold reduction in prediction error. Similarly, in the benchmark PUNQ-S3 case, forecast error decreased from 18% with the conventional CRM to about 9% with two-phase CRM. The two-phase CRM successfully captured polymer-induced mobility control, displacement dynamics, and interwell connectivity, whereas the conventional CRM failed under strong heterogeneity. The proposed two-phase CRM significantly outperforms the conventional formulation, offering a physics-informed, computationally efficient, and field-relevant framework offering strong potential for polymer flood evaluation under realistic reservoir conditions.