Abstract
<jats:title>Abstract</jats:title> <jats:p>Polymer flooding in viscous oil reservoirs has traditionally been explained through improved mobility control. However, recent research has highlighted that this classical "extended Buckley–Leverett" interpretation fails to capture the full physics of immiscible displacement at unfavorable viscosity ratios. Instead, the emergence of viscous crossflow at the finger scale has been identified as the dominant recovery mechanism in polymer flooding of viscous oils. This mechanism enables rapid oil banking and incremental recoveries far beyond those expected from mobility ratio improvement alone.</jats:p> <jats:p>In this work, this new simulation methodology for viscous fingering — based on fractional flow construction, maximum mobility relative permeabilities, and fine-scale correlated permeability fields — has been applied for the first time to a field polymer flood case outside the North Sea: Hilcorp's Milne Point J-Pad project on Alaska's North Slope. The study was carried out in two stages. First, a conceptual model of a horizontal injector–producer pair was analyzed, demonstrating how the viscous crossflow mechanism can explain the observed early oil bank formation, the reduction in water cut and the early appearance of polymer in breakthrough water at the producer. Second, the methodology was extended to the full J-Pad geomodel, enabling direct comparison of simulated viscous fingering patterns and crossflow dynamics with published production trends.</jats:p> <jats:p>The results show that the viscous crossflow framework, validated in laboratory slab floods up to 7,000 mPa·s oil, can be successfully upscaled to real field conditions in Alaska. The simulations reproduce the distinctive recovery features of the Milne Point polymer flood without requiring ad-hoc transmissibility multipliers or evolving fracture models. This confirms that polymer flooding in viscous oils is not merely a matter of mobility control but is instead governed by the interaction of immiscible fingering and viscous crossflow.</jats:p> <jats:p>This work provides a mechanistic reinterpretation of the Milne Point polymer flood. It demonstrates that the same physical principles underpinning slab-scale experiments are directly applicable to full-scale field projects. The implication is that many ongoing and planned polymer floods can be better optimized by explicitly incorporating viscous fingering and crossflow into simulation workflows, rather than relying solely on conventional relative permeability functions.</jats:p>