Assessment of grading polymer injection by simulation at laboratory scale

Abstract

When the mobility ratio is unfavorable, or the heterogeneity in the reservoir is high, the water flooding losses its effectivity causing early water breakthrough and low recovery factors.Polymer flooding improves the unfavorable mobility ratio by increasing the water viscosity and reducing the permeability to water due to polymer adsorption. Typically, these projects inject a finite volume of the polymer due to economic aspects since the polymer cost predominantly affects operation expenditures. However, this injection scheme can present an abrupt viscosity transition at the polymer slug tailing. The chase water can form fingers, bypassing the polymer slug and destroying its integrity. Thus, grading viscous injection is proposed to mitigate that issue. Polymer slugs with different concentrations are sequentially injected, followed by chase water. According to the literature, graded polymer schemes are not usual in the field, and only a few studies are available at the laboratory scale, including simulation. This dissertation evaluates the viscous grading polymer scheme by numerical simulation at a laboratory scale through performance indicators. The specific objectives are to build a comprehensive workflow that integrates laboratory data and validates the proposed procedures on three experimental tests reported in the literature.The proposed workflow includes three parts. The input data preparation section describes how to integrate the data from rock-fluid studies and single-and two-phase polymer core floods to calculate the parameters required by the simulator. The simulation grid selection describes the usage of static and dynamic grids to reduce the numerical dispersion and speed up the simulation runs. And finally, the history matching stage presents a workflow to fit the model systematically.The results demonstrate the workflow applicability by simultaneously fitting the cumulative oil, water, and differential pressure histories of the graded and continuous flooding schemes. The integration of cumulative volume and pressure responses allowed detecting key events, and the dynamic grid amalgamation reduced the computational time compared with the static grid. The performance indicators based on fitted models helped identify the most efficient injection scheme according to the polymer mass used, oil recovery, injected/produced water, and injectivity compared to water flooding. The three-slugs strategy performed better than the two-slug one. The proposed workflow integrating laboratory data can be helpful to the researchers involved with core floods and small-scale simulation and, also, with modeling and simulations at pilot and field scales.