Speaker
Description
Understanding emulsion formation and transport in porous media is critical for improving oil recovery and predicting flow behavior during water-based enhanced oil recovery (EOR). This study investigates nanoparticle-assisted emulsion generation, stability, and flow behavior through an integrated experimental and data-driven approach.
Laboratory screening experiments were first conducted to evaluate the stabilizing performance of metal-oxide nanoparticles (NiO, Al₂O₃, TiO₂) in surfactant-assisted oil–water emulsions under varying salinity and acidic conditions. Nickel oxide nanoparticles exhibited superior emulsion stability and monodisperse droplet size distributions, maintaining stability even at low pH. These findings guided subsequent coreflooding experiments performed on Berea sandstone cores under capillary-dominated flow conditions.
Coreflooding tests were conducted on cores with an absolute permeability of 174 mD and connate water saturation of 17%. Secondary recovery using chemical flooding resulted in an oil recovery factor of 61.3%. Subsequent emulsion generation and low-salinity water (LSW) flooding increased the total recovery factor to 68.6%, demonstrating a clear incremental recovery due to emulsion-assisted mechanisms. In-situ generated emulsions were observed to be stable and monodisperse, as confirmed by microscopic analysis of produced fluids.
During both chemical flooding and emulsion injection stages, a significant increase in pressure drop was observed compared to conventional waterflooding. The elevated differential pressure indicates increased flow resistance associated with emulsion formation and transport within the porous medium. This behavior suggests effective mobility control, where the higher apparent viscosity of emulsions reduces the mobility ratio, promotes flow diversion, and improves sweep efficiency.
To complement the experimental observations, a machine learning framework was developed to predict the apparent viscosity of natural water-in-oil emulsions across a wide range of shear rates and physicochemical conditions. Trained on over 1000 experimental data points, gradient boosting models achieved high predictive accuracy (R² ≈ 0.97), successfully capturing the non-Newtonian rheology of emulsions.
Overall, the combined experimental–computational approach provides quantitative insight into emulsion-mediated flow mechanisms in porous media and highlights the potential of nanoparticle-assisted emulsions for enhanced oil recovery.
| Country | Kazakhstan |
|---|---|
| Acceptance of the Terms & Conditions | Click here to agree |
