A Unified Reactive Transport-Instability Framework for Enhanced Oil Recovery

Thesis Abstract

This study addresses the persistent gap in predictive understanding of how coupled reactive transport and hydrodynamic instabilities influence the efficiency of chemical-enhanced oil recovery (CEOR) processes in heterogeneous reservoir systems. The central aim is to develop a unified framework that integrates reactive transport theory with instability-driven flow phenomena to optimize CEOR performance under field-relevant conditions. Specific objectives are (1) to quantify the interplay between mineral dissolution/precipitation reactions and porosity–permeability evolution; (2) to characterize how transport instabilities (e.g., viscous fingering, channelization) interact with reaction fronts to alter sweep efficiency; (3) to formulate a reduced-order, hybrid model that couples advection–diffusion–reaction dynamics with stability analysis for multiphase flow; (4) to calibrate and validate the framework against laboratory coreflood experiments and field-scale analogs; and (5) to derive actionable insights for operational optimization of surfactant-polymer and gas-assisted CEOR strategies. Methodologically, the study adopts a multipronged design combining controlled laboratory experiments, numerical simulation, and theoretical development. The population comprises core samples from a mature, heterogeneous sandstone reservoir and representative bentonite-containing shales, with permeabilities ranging from 0.5 to 2.0 Darcy and porosities of 12–22%. A stratified sampling scheme yields 24 core plugs for experiments and 6 reservoir-representative models for numerical analysis. Data collection employs high-resolution X-ray computed tomography to monitor porosity and permeability evolution, inline impedance spectroscopy for reaction progress, and laser-induced fluorescence for multiphase distribution. Chemical systems mimic CEOR fluids with appendants a surfactant solution, a polymer viscosity modifier, and a CO2–brine gas–liquid interface. Operational conditions reflect reservoir-relevant temperatures (60–90°C) and pressures (15–30 MPa). The instruments generate time-resolved datasets of solute concentrations, reaction rate constants, permeability alterations, and pore-scale flow patterns. Analytical methods integrate established and novel approaches. Parameter estimation utilizes nonlinear least squares and Bayesian calibration to constrain reaction rate constants and porosity-permeability relationships in the reactive transport model. Stability analysis of the coupled system employs linear stability theory (LST) to identify dominant wavenumbers associated with transport instabilities under varying reaction extents. Numerical simulations implement a hybrid framework that couplesAnisotropic poroelastic transport with phase-field representations of multiphase flow and a reaction-transport module governed by Monod-type kinetics for rock-fluid interactions. Model validation uses cross-validation against laboratory coreflood results, with additional external validation against historical field data from a mid-life field in a mature carbonate–sandstone reservoir. Statistical inference employs regression analysis to relate instability metrics to recovery factors and ANOVA to test the significance of fluid composition, injection rate, and temperature on oil recovery. Sensitivity analyses quantify the influence of pore-scale heterogeneity, reaction rates, and instabilities on macroscopic sweep efficiency. Key expected findings include (i) identification of critical Peclet and Damkohler number regimes where reactive transport enhances or suppresses instability-driven channelization; (ii) demonstration that targeted reaction control can damp unfavorable fingering while promoting beneficial front stabilization, thereby improving volumetric sweep efficiency by 8–15% in synthetic scenarios; (iii) development of a reduced-order unified framework that accurately predicts oil recovery outcomes within ±5% of benchmark experiments for 60–100 coreflood simulations. The study contributes to knowledge by bridging reactive transport theory with instability phenomena in CEOR contexts and by providing a transferable modeling framework combining LST, pore-scale- to field-scale coupling, and data-driven calibration. The main conclusion posits that a unified reactive transport-instability framework enables proactive control of reaction fronts and flow instabilities to optimize CEOR performance, with practical recommendations including calibrated injection strategies, fluid compositions, and operating windows that minimize detrimental channeling while maximizing sweep efficiency. Recommendations for practice emphasize embedding the framework within reservoir simulation tools for real-time optimization and extending the approach to carbonated CEOR and polymer-surfactant–gas systems to generalize applicability across reservoir types.