Abstract
Successful design and evaluation of polymer flooding field applications require good injectivity among other factors. For a proper polymer solution the injectivity in a porous medium is crucial, where considering Newtonian and nonnewtonian polymer behaviour plays a major role in order to avoid nullifying the whole economic performance of a project. Former polymer pilots in Germany and other countries considered using shear devices to improve the polymer injectivity, without consider analysing the changes in polymer viscoelasticity. Current interpretations also tend to describe viscoelastic behaviour either through first normal stress difference (N1) or oscillatory shear (relaxation) solely, without considering the probable relation of both. This paper suggests a different approach to assess polymer injectivity. It is based on the evaluation of induced dynamic degradation and polymer viscoelastic analysis comparing N1 and relaxation behavior.
Optimum polymer solutions for two different unconsolidated reservoirs with 7066 mg/l and 4000 mg/l in TDS and oil viscosity of 6.5 mPas and 3 mPas respectively, were defined and analysed. Three high molecular weight (MW) partially hydrolysed polyacrylamides (HPAM) were used. Firstly, the behavior of mother and stock solutions were defined for each polymer by shearing the polymer molecules after dissolution with a dynamic shear device, operated with a constant pressure drop and turbulent flow conditions, aiming to degrade mechanically the longest macromolecules without significantly reducing the mean value of MW. Secondly diluted polymer solutions were prepared between 600 ppm to 2000 ppm using a quasi-standardized procedure and rheologically characterized defining the required concentration for a viscosity at 3 s−1 and reservoir temperature to match the measured live oil from PVT analyses.
The defined concentrations were characterized as function of temperature for modelling purposes. The elastic properties were characterized by N1 and oscillatory measurements. Injectivity was checked by pressure controlled filtration test and sand pack experiments. Viscosity losses due to the shearing were found only about 10% in a typical shear rate for flow conditions (lower than 10 s−1). The benefits of shearing became evident once the solutions were checked through the filtration, showing values for the filtration ratio between 1 to 1.1 for sheared solutions compared to 1.6 to 2.1 for unsheared solutions. Preliminary sand pack experiments showed a good agreement with filtration tests. Polymer viscoelasticity in relaxation measurements clearly depicted a weak elastic behavior with the prevalence of viscous modulus against elastic modulus. An important finding was a gel like behavior in solutions prepared at 7066 mg/l TDS, where the equilibrium of gel response showed a plateau at low frequencies, making the response independent of the frequency applied. Thus cross-over points were not found, hence no relaxation times were identified. N1 measurements provided a good elastic response reaching values between 20 to 40 Pa, which is in agreement with some literature values.
These results enable to understand why pre-treatment polymer methods, such as dynamic shearing, help to improve injectivity, which combined with a fully viscoelastic characterization generate guidelines for an accurate polymer selection in accordance with increasing evidence that viscoelastic characteristics of polymers help improving polymer flood efficiency.