Optimizing Field Development of Greenfield Polymer Floods using Experimental Design

Abstract

Waterflooding viscous oil suffers from poor displacement efficiency, requiring excessive volumes throughput for a given recovery. Heterogeneity reduces recovery further by reducing volumetric sweep efficiency. These can be mitigated by modifying injection viscosity with polymer. An improved mobility ratio reduces the throughput to achieve a given recovery, increases volumetric sweep efficiency, and reduces water production and associated handling costs. These three factors make polymer flooding an attractive alternative to waterflooding. From a volumetric sweep perspective, there is no practical limit to the benefit of increased injection viscosity. However, increased fluid viscosity decreases injectivity, through damage (residual resistance factor) or increased resistance to flow with injection of a more viscous fluid. Injectivity issues can be overcome by decreasing well spacing, but at a capital cost. This introduces a de facto limit on polymer concentration; by reducing the concentration we can increase injectivity and increase well spacing, but at the expense of reduced macroscopic displacement efficiency and increased throughput. There are two decision variables (polymer concentration and well spacing) that have three consequences (injectivity, processing rate, and total volume throughput). We have designed a workflow that maximizes NPV for given reservoir and flood properties by optimizing well spacing and polymer concentration. For given polymer concentration and spacing we analytically solve for injection rate vs. time and use that in fractional flow calculations to determine oil production vs. time. The workflow is automated, so a wide range of decision variables can be easily explored via Experimental Design. Surface response methods are then used to determine well spacing and polymer concentration that maximize net present value. The methods developed for the optimization are presented, and the workflow steps are outlined. The method was applied to a field study; for the reservoir properties and under tertiary polymer flooding, the optimized set of conditions was 1900 ppm polymer and an injector-producer spacing of 700 ft. Injecting polymer prior to any water injection reduces the spacing to 380 feet and increases the polymer concentration to 2250 ppm, in large part because the low oil bank mobility dominates injectivity decline. This suggests an important design consideration of pre-conditioning the reservoir prior to instigating a polymer flood.