Assessment of Enhanced Oil Recovery Techniques in Mature Fields Using Reservoir Performance Data

Thesis Abstract

The depletion of primary reservoir energy in mature oil fields necessitates the application of effective enhanced oil recovery (EOR) techniques to maximize hydrocarbon extraction and prolong field life. This study aims to systematically assess the performance and effectiveness of selected EOR methods—namely waterflooding, gas injection, and chemical EOR—within mature reservoirs by analyzing reservoir performance data. The specific objectives include quantifying the incremental recovery attributed to each EOR technique, identifying the reservoir characteristics influencing EOR success, and developing predictive models to optimize sequestration and recovery strategies. Employing a mixed-methods research design, the study integrates quantitative data analysis with qualitative insights to holistically evaluate EOR performance across three mature fields with proven reservoir data spanning the past 15 years. The population comprises approximately 150 production wells with comprehensive historical data, from which a stratified random sampling technique selected 50 wells representing different reservoir zones, EOR techniques, and production histories. Data collection involved extracting detailed reservoir performance data—including pressure readings, injection rates, production volumes, water cut, and other relevant parameters—from company databases complemented by field logs and prior exploration reports. Data validity and reliability were ensured through calibration with field measurements and cross-verification with previous studies. The analytical framework centers on multiple regression analysis to identify the key factors influencing recovery efficiency, supported by Analysis of Variance (ANOVA) tests to compare the impacts of different EOR methods across reservoir zones. Additionally, time-series analysis was employed to evaluate trends in reservoir pressure and production rates pre- and post-EOR implementation. To develop predictive models, the study utilized machine learning techniques—specifically, random forest algorithms—integrated within the reservoir performance data to forecast long-term recovery trajectories under varying operational scenarios. The theoretical underpinning drew upon the theories of reservoir engineering—particularly Darcy’s Law for fluid flow and the Theory of Recovery Efficiency—and the Resource-Based View (RBV) to contextualize how reservoir characteristics influence technological deployment and success. Expected findings of this research are that specific reservoir attributes—such as porosity, permeability, and residual oil saturation—significantly correlate with the effectiveness of different EOR methods; for instance, waterflooding performs optimally in high-permeability zones, while chemical EOR shows higher incremental recovery in reservoirs with finer pore structures. The analysis is anticipated to reveal that gas injection schemes contribute substantially to pressure maintenance, thereby improving overall recovery factors by an average of 18%. Furthermore, the predictive models are expected to demonstrate high accuracy (above 85%) in simulating reservoir responses, enabling more targeted and economically viable EOR application strategies. This study contributes valuable empirical evidence to the body of knowledge on EOR optimization, providing a comparative assessment of techniques tailored to reservoir heterogeneity. Its innovative integration of advanced statistical and machine learning approaches offers a significant methodological contribution for reservoir management. Additionally, the development of a decision-support framework based on reservoir parameters and performance history bridges the gap between theoretical models and practical operations. The main conclusion underscores the importance of reservoir-specific EOR selection based on robust data analytics; chemical EOR and polymer flooding emerge as promising options for reservoirs with moderate porosity but low permeability, whereas waterflooding remains effective in highly permeable zones. Recommendations include adopting a data-driven EOR planning framework, enhancing reservoir monitoring systems, and conducting further research into hybrid EOR methods that combine chemical and gas techniques. The study advocates for integrating reservoir performance data analytics into routine reservoir management to improve recovery outcomes, reduce costs, and extend the productive lifespan of mature oil fields. Future investigations should explore real-time data integration and the application of advanced AI techniques for dynamic reservoir simulation, thereby advancing innovative and sustainable extraction practices in mature fields.