Revisiting polymer selection workflows for chemical enhanced oil recovery

Abstract

The workflow used to select and characterize polymers for chemical EOR has remained essentially unchanged since the first field projects in the 1960s — viscosity curves, filtration tests, and then coreflood experiments from which parameters are extracted for history matching. This paper argues that this standard workflow is poorly suited to predicting real field behavior, and that it has systematically prevented engineers from correctly modelling polymer injection and forecasting injectivity.

Three sources of misrepresentation are examined. First, laboratory polymer solutions are pristine: field preparation through pumps, valves, and perforations alters molecular weight distribution and in-situ rheology in ways that laboratory tests do not replicate. Second, standard filtration protocols (filter ratio test, pre-filtration before coreflood) are inconsistent across the industry and do not reflect field conditions. Third, nearly all corefloods are run at fixed injection rate — mimicking an extrusion process rather than the constant-pressure-drop displacement that governs most of the reservoir. A revised, field-first workflow is proposed, along with practical recommendations for more representative polymer characterization.

Key Takeaways

  • A slight pre-shearing of the polymer solution (causing only ~10% viscosity loss) can reduce the resistance factor by 30% or more and halve retention — meaning laboratory results from unsheared, pristine solutions can be highly misleading for field predictions.
  • There is no industry standard for polymer solution preparation, pre-filtration, or filter ratio testing. The wide variability in protocols makes cross-study comparisons unreliable and contributes to injectivity forecasting errors.
  • Corefloods run at fixed injection rate artificially amplify shear-thickening effects that are unlikely to occur in the field, where polymer enters the reservoir through perforations, microfractures, or pre-existing water channels — not directly into tight matrix at high velocity.
  • The shear rate in the reservoir drops rapidly with distance from the wellbore. Before designing any laboratory experiment, engineers should calculate the shear rate profile from surface to reservoir depth and design tests to represent the dominant flow regime (usually the low-shear Newtonian plateau deep in the reservoir).
  • Rheological data from viscometers fed directly into simulators systematically overestimate the injectivity loss predicted for polymer floods — the resistance factor measured in cores (not bulk viscosity) is the more relevant and reliable input.
  • A “reverse” workflow — characterize the reservoir and near-wellbore conditions first, then design lab tests to reflect those conditions — would produce more representative results and avoid dismissing viable projects based on overly pessimistic simulations.
Read Article

Related Articles