Developing and evaluating a model for Surfactant-Foam Flooding

Many reservoirs are considered unsuitable candidates for Enhanced Oil Recovery. Surfactant-Foam flooding, also referred to as Low-Tension-Gas flooding, could be a solution to enhance the oil recovery in tight, saline or high temperature reservoirs. SF-flooding combines reduction of the oil-water interfacial tension by surfactant, with mobility control provided by foam. While coreflood experiments have been performed, no Surfactant-Foam model has been developed yet. Therefore, the objective of this research is to develop and evaluate a conceptual model for Surfactant-Foam flooding. Several available foam- and surfactant flooding models are analyzed. By comparing the various approaches, two models are selected based on the identified modelling objectives, model complexity and available experimental data. In this study a Surfactant-Foam model is developed that combines an implicit-texture foam model with a two-phase effective surfactant model. The model assumes foam is in local-equilibrium and correlates salinity and a minimum surfactant concentration with a lowered water-oil interfacial tension. The model does not explicitly model the micro-emulsion phase. Furthermore, the model includes surfactant adsorption and salt and surfactant dispersion. An attempt to model foam diversion with surfactant dispersion has been made, but results are limited due to numerical instability. For every coreflood the effective dispersion coefficient is determined by fitting the experimental effluent salinity data with a solution to the 1D advection-diffusion equation. The performance of the combined model is evaluated with a reservoir simulator. According to the simulation study the surfactant model affects the gas relative permeability through a correlation with the connate water saturation. The defined chemical connate water saturation always affects the gas mobility, either by reducing the gas relative permeability, or by minimizing the impact of the limiting capillary pressure. The results show that an experimentally applied salinity gradient, correlated with a decrease and increase in water-oil interfacial tension, cannot be modelled with capillary number dependent relative permeability curves. In this research it is assumed that the water-oil interfacial cannot increase, after it achieved an ultra-low value. According to the current desaturation approach, oil prefers to flow in the presence of water in stead of gas at ultra-low interfacial tension. More research is required to investigate if this can be related with physics, or if it is purely a modelling artifact. Furthermore, the simulation study shows that the foam matching parameters of the implicit-texture model require optimization for cores with a mutual difference in connate water saturation, due to variation of the impact of the limiting capillary pressure. Simulations are performed to history match experimental data and are aimed to match the measured pressure drop, oil production and effluent salinity profiles. As Surfactant-Foam flooding is taking its first steps in the laboratory, the accuracy and amount of available data is limited. To improve the validity of the model parameters, coreflood experiments in which the foam and surfactant processes are decoupled were interpreted. The surfactant concentration and water saturation are identified as the main drivers affecting the foam texture. Waterflooding pressure data was used to determine the non-chemical relative permeability parameters. The analyzed Surfactant-Foam corefloods are conducted in tight Indiana Limestone cores. Carbonates often have a complex pore structure due to their dual porosity. The experimental data suggests the presence of heterogeneities such as high permeable zones. Therefore, a one-dimensional model is expanded to a two-dimensional model, represented by two layers. Both the one and two-dimensional Surfactant-Foam model achieve a reasonable history match with the cumulative oil recovery and pressure gradient. The two-dimensional model, with a small difference in flow capacity between the layers, successfully matches the size of the oil cut. With the two-dimensional model an improved match with the effluent salinity profile can be achieved at the expense of the oil cut and recovery match. The contribution of measurement errors and heterogeneities to the effluent salinity profile requires more research. As the experimental data are subject to a high level of uncertainty, more corefloods should be performed to identify the most suitable geological representation of the Indiana Limestone.

Subject

surfactant-foam flooding
EOR
reservoir engineering
simulation study
modelling

To reference this document use:

http://resolver.tudelft.nl/uuid:1637a7c0-02cb-41ad-9be8-6db976fd63b6

Embargo date

2017-09-18

Part of collection

Student theses

Document type

master thesis

Rights

Files

PDF

Thesis_M.I.C.Knol.pdf

46.82 MB

Close viewer