AI-assisted optimization of enhanced oil recovery using nanofluid flooding in carbonate reservoirs

Project Abstract

This study presents an integrated AI-driven framework for optimizing enhanced oil recovery (EOR) in carbonate reservoirs through nanofluid flooding, addressing the critical challenges of heterogeneity, low oil mobility, and high-capillary force resistances that limit conventional EOR methods. A hybrid modeling approach combines machine learning (ML) surrogates with physics-based reservoir simulations to enable rapid screening and robust optimization of nanofluid properties (particle size, concentration, zeta potential, and stability), injection strategies (rate, pressure, pulsing schemes), and surface chemistry modifiers that influence wettability alteration and interfacial tension. The methodology begins with a comprehensive dataset generated from high-fidelity reservoir simulations augmented by laboratory measurements of nanofluid stability and rock-fluid interactions under carbonate conditions. Feature engineering captures reservoir heterogeneity, mineralogy, brine salinity, pH, temperature, and capillary pressure curves to train ML models that predict oil saturation, relative permeability, and breakthrough times under varying nanofluid scenarios. These models are embedded in an optimization loop using multi-objective algorithms to maximize cumulative oil production and recovery factor while minimizing injection costs, energy usage, and nanoparticle loss due to adsorption and filtration. Uncertainty quantification (UQ) is integrated via Bayesian ensembles to assess the confidence bounds of EOR performance given uncertain rock properties and fluid characteristics, enabling risk-informed decision-making. The study also investigates the synergistic effects of nanofluid chemistry on wettability alteration in carbonate matrices, evaluating the balance between improving oil mobility and maintaining reservoir rock integrity. A novel coupling scheme links flow physics with data-driven predictors to allow real-time optimization in field pilots, complemented by sensitivity analyses that identify the most influential parameters driving incremental recovery. Results from synthetic and scaled-core experiments demonstrate that optimized nanofluid formulations can achieve a substantial reduction in interfacial tension, a shift toward favorable wetting states, and enhanced oil displacement efficiency, leading to higher oil recovery factors compared with baseline polymer and CO2-based EOR methods. The framework shows robustness across varying pore structures and heterogeneities, with UQ analyses indicating acceptable risk levels under typical operational uncertainties. Economic assessment highlights potential improvements in project economics through reduced chemical usage, lower injection volumes, and accelerated production timelines, while environmental considerations address nanoparticle fate, transport, and potential containment strategies. The research provides a scalable, adaptive optimization toolset for designing field-ready nanofluid EOR campaigns in carbonate reservoirs, offering actionable guidelines on (i) selecting nanoparticle types and surface modifiers compatible with carbonate rocks, (ii) tailoring injection protocols to reservoir-specific capillary and permeability distributions, and (iii) implementing real-time monitoring and adjustment protocols to sustain optimal performance. The study contributes to the broader understanding of nanofluid–rock–oil interactions under high-temperature, high-salinity carbonate environments and advances the practical deployment of AI-enabled EOR strategies that can extend the productive life of mature carbonate fields.

Project Overview

What This Project Is About

---

A straightforward exploration of using tiny, specially designed liquids (nanofluids) to push more oil out of carbonate rock formations. The project combines a practical oilfield technique (enhanced oil recovery) with advances in nanotechnology to improve how water floods drive oil toward production wells.

---

---

The Problem It Addresses

---

---

---

Objectives of the Project

---

1. Learn how nanofluids interact with carbonate rock and oil.
2. Assess how different nanoparticle types and concentrations affect oil recovery.
3. Develop a simple model to predict recovery improvements.
4. Propose a practical, safe lab experiment plan for testing nanofluid flooding.
5. Explain potential environmental and cost considerations.

---

---

What You Will Do Step by Step

---

Review basic concepts of oil recovery and nanotechnology. Design small-scale lab tests to compare standard water floods with nanofluid floods. Collect data on oil production and fluid behavior. Analyze results using simple, user-friendly tools to identify trends. Summarize findings and discuss practical implications for field applications.

---

---

Expected Outcome

A clear understanding of whether nanofluid flooding can boost oil recovery in carbonate rocks, with simple guidelines on when and how it might be used, plus considerations of cost, safety, and environmental impact.