Long-Term Performance of EOR Methods in Offshore Reservoirs: An Empirical Field Study

Thesis Abstract

Offshore reservoirs face progressive declines in oil production despite advances in enhanced oil recovery (EOR) technologies, due to complex reservoir heterogeneities, scaling, and operational constraints that influence long-term performance. This study addresses the persistent gap between laboratory-predicted EOR effectiveness and field-scale outcomes by empirically evaluating the sustained impact of thermal, chemical, and gas- injection EOR methods across multiple offshore fields. The aim is to quantify long-term performance, identify drivers of variability, and develop a pragmatic framework for forecasting ultimate recovery under offshore conditions. Specific objectives include (1) assessing time-series performance metrics (oil production rate, cumulative oil production, water cut, and gas/oil ratio) over a 12-year horizon for at least three offshore fields employing distinct EOR approaches; (2) examining reservoir and operational factors (heterogeneity indices, aquifer support, injection strategies, and subsea infrastructure constraints) that modulate EOR efficacy; (3) comparing observed field performance against model-based forecasts using history matching and machine-learning assisted extrapolation; and (4) developing a decision-support framework to optimize monitoring and intervention schedules for long-term EOR performance. The methodology adopts an embedded, retrospective longitudinal design anchored in real-field data from offshore operators in the North Sea and Gulf of Mexico, encompassing 120 well-leveraged production and injection sites across five campaigns. The population includes offshore reservoirs undergoing EOR implementation since 2008, with a purposive sample of 3–5 fields per campaign selected to represent thermal, chemical, and gas injection methods. Data sources comprise production manifests, injection volumes, reservoir simulation outputs, wellbore integrity records, subsea equipment maintenance logs, and geophysical surveillance data. Instruments include standardized data extraction templates, reservoir performance dashboards, and interview guides for field engineers to capture operational nuances. Validity and reliability are addressed through triangulation of production data with reservoir model histories, cross-field inter-laboratory data validation, and inter-operator harmonization of measurement units. Data analysis proceeds in four steps (i) descriptive statistics and time-series decomposition to delineate trend, seasonality, and regime shifts; (ii) multivariate regression and fixed-effects panel analysis to quantify relationships between EOR type, injector/producer configurations, and performance indicators; (iii) reservoir model calibration via history matching using refined black-oil and compositional simulators, complemented by ensemble Kalman filter techniques for uncertainty quantification; (iv) machine-learning-based forecast ensembles (random forest, gradient boosting) to project long-term recovery under varying scenarios. The theoretical lens integrates the Theory of Planned Behavior to interpret operator decision-making in sustaining EOR operations and the Resource-Based View to contextualize capabilities and constraints of offshore assets as sources of competitive advantage in recovery. Expected findings include (a) thermal EOR demonstrates superior sustained incremental oil recovery in high-permeability, high-temperature reservoirs, yet exhibits more rapid performance decay due to thermal inefficiencies and infrastructure heat losses; (b) chemical EOR performance shows robust early-stage response in channels with favorable wettability but attenuates under salinity and polymer degradation, with long-term gains contingent on continuous chemical makeup and injection pressure management; (c) gas-injection EOR yields durable response in pressure maintenance-dominant systems but is sensitive to reservoir compaction, layer-to-layer variability, and surface compression constraints; (d) key drivers of long-term performance are injectivity trends, well integrity, and subsea equipment reliability, with reservoir heterogeneity and management of scaling being critical moderating factors. The study contributes to knowledge by offering a transferable empirical framework for predicting offshore EOR longevity, integrating statistical, mechanistic, and data-driven approaches for performance forecasting, and clarifying how asset-specific capabilities determine field-scale outcomes. Conclusions point to the necessity of integrated monitoring programs combining real-time reservoir analytics with proactive infrastructure maintenance to sustain EOR gains. Recommendations include (1) implementing adaptive monitoring intervals guided by ensemble forecasts, (2) prioritizing subsea equipment redundancy and heat management for thermal EOR, (3) establishing chemical-management protocols including periodic re-optimizations of composition and injection schedules, and (4) expanding cross-field data-sharing initiatives to refine generalizable models of offshore EOR performance under uncertainty.