A numerical framework for predicting oil bank formation during foam flooding

Water-alternating-gas (WAG) flooding is widely used for enhanced oil recovery (EOR), yet significant amounts of oil often remain trapped after extensive WAG operations. Foam flooding improves sweep efficiency by reducing gas mobility, and oil bank formation is a key displacement mechanism in foam-assisted EOR. However, predicting the reservoir initial and injection conditions that promote oil bank formation remains challenging. Previous studies have either relied on highly simplified foam models to enable analytical classification of oil bank formation scenarios or, when adopting more realistic foam models, restricted the analysis to a limited number of representative cases. This work presents a numerical framework to predict oil bank formation during foam flooding, using the industry-standard foam model implemented in CMG-STARS. The framework systematically evaluates a wide range of reservoir initial and injection conditions, incorporates sensitivity and uncertainty analyses of foam model parameters, and is validated against both synthetic and experimental datasets. The results demonstrate that oil bank formation is not a universal outcome of foam flooding, but depends on specific combinations of reservoir initial conditions, injection conditions, and foam properties. Moreover, stable oil banks delay gas breakthrough and substantially reduce the gas-oil ratio while maintaining comparable oil recovery under the investigated conditions, highlighting their importance for sweep efficiency and foam-assisted EOR performance. Sensitivity analyses further show that uncertainties in foam model parameters strongly influence the occurrence and topology of oil banks, while permeability heterogeneity and gravity affect their stability. The proposed framework provides a practical methodology for identifying operating conditions that favor oil bank formation and supports the design of foam-assisted EOR processes.