Laboratory Evaluation of Polymer Retention in a Heavy Oil Sand for a Polymer Flooding Application on Alaska's North Slope

Abstract

SummaryFor a polymer flooding field trial in a heavy oil reservoir on Alaska's North Slope, polymer retention is a key parameter. Because of the economic impact of retention, this parameter was extensively studied using field core material and conditions. In this paper, multiple types of laboratory measurements were used to assess hydrolyzed polyacrylamides (HPAM) polymer retention, including a brine tracer, effluent viscosity, total effluent organic carbon, and effluent chemiluminescent nitrogen. Retention tests were conducted in different Milne Point Schrader Bluff sands, with extensive permeability, grain size distribution, X-ray-diffraction (XRD), and X-ray fluorescence (XRF) characterizations. Several important findings were noted. Polymer retention based on effluent viscosity measurements can be overestimated unless the correct (nonlinear) relation between polymer concentration and viscosity is used. Polymer degradation (either mechanical or oxidative) can also lead viscosity-based measurements to overestimate retention. Inaccessible pore volume (PV) (IAPV) can be overestimated if insufficient brine is flushed through the sand between polymer banks. Around 100 PVs of brine may be needed to displace mobile polymer to approach a true residual resistance factor and properly measure IAPV. Even for a sandpack with kwsor = 20 md, IAPV was zero for HPAM with a molecular weight (Mw) of 18 MM g/mol. Fine-grained particles (<20 µm) strongly impacted polymer retention values. Native NB#1 sand with a significant component of particles <20 µm exhibited 290 µg/g, while the same sand exhibited 28 µg/g after these small particles were removed. Polymer retention did not necessarily correlate with mineral composition. The NB#1, NB#3, and OA sands had similar elemental and clay compositions, but the NB#1 sand exhibited ∼10 times higher retention than the NB#3 sand. Polymer retention did not necessarily correlate with permeability. NB#1 sand exhibited much higher retention than OA sand, even though NB#1 sand was twice as permeable as OA sand. No evidence of chromatographic separation of HPAM molecular weights was found in our experiments. Although retention tended to be greater without a residual oil saturation (than at Sor), the effect was not strong. Aging a core (with high oil saturation) at 60°C reduced HPAM retention by a factor of two. Under similar conditions, polymer retention was greater for a higher Mw HPAM (18 MM g/mol) than for a lower Mw HPAM (10 to 12 MM g/mol). In many cases with high polymer retention values (e.g., 240 µg/g), polymer arrival at the end of the core was relatively quick, but achieving the injected concentration occurred gradually over many PVs. This effect was not caused by chromatographic separation of polymer molecular weights. Results from modeling of this behavior were consistent with concentration-dependent polymer retention. The form assumed for the retention function in a simulator can have an important impact on the timing and magnitude of the oil response from a polymer flood. Field-based observations can underestimate polymer retention, depending on when the tracer and polymer concentrations were measured and the assumptions made about reservoir heterogeneity.