Despite recent drop in the growth of global oil demand, the trend is expected to gradually pick up and continue increasing. Industrial and transportation sectors are still considered the highest consumers of oil. The petrochemicals sector’s demand for oil is increasing sharply and expected to continue in that fashion for the upcoming years. As oil fields age and mature, extraction of oil via primary and secondary techniques becomes, to an extent, inefficient. That encourages more research and development in enhance oil recovery (EOR) methods. In EOR, the aim is to either change a physical or chemical property of reservoir fluid in order to improve the oil recovery factor. The techniques can be categorized as thermal, physical, chemical or gaseous. In this experimental study, the lessons learned from gas flooding methods and surfactant flooding methods are taken into account in order to come up with a novel Foam-Assisted Chemical Flooding (FACF) procedure that aims to enhance the oil recovery factor to its maximum. In this approach, a surfactant slug solution is injected into a core at residual oil after water flooding conditions to mobilize trapped oil by capillary pressure. Then, a surfactant drive solution is co-injected with N2 for foam generation to serve as mobility buffer displacing the accumulated mobilized oil. The experimental study is performed under reservoir conditions of 90 ±1oC temperature and 20 bar of back pressure. In this study, surfactant stability is tested in synthetic formation brine. Then, phase behaviour tests are conducted to identify the capability of the surfactant to reduce o/w interfacial tension (IFT) to ultra-low values. The resulting solutions are categorized into their associated Winsor Types and classified based on salinity as under-optimum, optimum and over-optimum. A final surfactant slug solution is formulated based on these tests. Afterwards, bulk foam tests are performed in absence and presence of crude oil to test surfactant foaming ability and the resulting from stability and strength. Core-flood experiments are carried out to assess the possibility of generating foam in porous media in absence of crude oil and at residual oil to waterflooding. Full EOR FACF experiments are conducted, two at under-optimum and two at optimum salinity conditions. Two FACF experiments are performed with the assistance of medical CT scanner. In one FACF experiment, the foam is pre-generated utilizing a mixing tee and then injected into the Bentheimer sandstone. The study reported here showed that surfactants are not stable in synthetic seawater injection brine, as it tends to form complexes in presence of divalent ions, and subsequently generate precipitations. Stability was achieved by removing the divalent ions from the synthetic brine. In addition, phase behaviour study yielded that surfactant (A) is a better o/w IFT reduction agent than surfactant (B). A distinct layer of micro-emulsion was observed in excess of water and oil phases. On the other hand, surfactant (B) displayed better foaming abilities than surfactant (A) in absence of crude oil. Using surfactant (B), foam was generated in multiple qualities in a Bentheimer sandstone core-flood experiments in absence of crude oil. The critical foam gas fraction was found to be 75%. However, attempts to generate foam in porous media at residual oil to water flooding conditions were not successful. In three FACF core-flooding experiments, weak and unstable foam was generated during the surfactant drive co-injection phase. Whereas, in the last FACF experiment where a mixing tee was utilized, pressure drop and gas breakthrough data show that stable foam was generated. The CT images from two FACF experiments, one at optimum and the other at under-optimum salinity conditions displayed unstable water front in waterflooding phase, and unfavourable mobility conditions, during the surfactant slug injection. However, the effect of salinity conditions was seen in the different oil bank shapes in both experiments. The one at under-optimum salinity condition showed more unstable front. The study reported here showed that the FACF technology yields improving oil recovery of 70±5%, 77±5% and 73±5% for the FACF experiments where it was very challenging to generate foam in-situ and reached up to 80±5% of oil initially in place in the case where foam was pre-generated outside the core (55%, 61%, 59% and 46% are the oil recovery factors after waterflooding, respectively). The study revealed that drive foam strength has a bigger impact than its surfactant slug salinity.

EOR processes employing gas injection can be very efficient in recovering oil where the gas sweeps. Unfortunately, gas injection has poor sweep efficiency because or reservoir heterogeneity, viscous instability and gravity segregation of injected gas to the top of the formation. Foam can improve the sweep efficiency in gas injection in Enhanced Oil Recovery. The best injection strategy to overcome gravity override in homogeneous reservoirs is a SAG process with a large slug of surfactant followed by a large slug of gas. An additional advantage of SAG injection is increased gas injectivity. Water is displaced from the near-well region where foam weakens, this raises gas mobility and increases injectivity. However, during liquid injection, the injectivity is considered to be poor.
Liquid injectivity is not fully understood during SAG process. In this study several core-flood experiments were conducted to investigate how injectivity is affected by the injection strategy. Through these experiments, it has been found that the gas injection flow rate has no significant effect during the gas injection period, foam collapses after roughly the same number of pore volumes of gas injected regardless of the injection rate.

During the liquid injection period, liquid was injected at different flow rates following a gas injection period. The results suggested a moderately shear-thinning behaviour. Liquid injection rate was increased 10, 40 and 100 times, but the rise in pressure gradient is not proportional to the increment in injection rate.

Furthermore, the effect of the size slug injected was investigated on the subsequent liquid injection. Results show that during prolonged periods of gas injection after foam, a region near the inlet is formed in which the gas mobility is much greater and the liquid mobility is much greater than downstream during the subsequent liquid injection. The bigger the gas slug size, the better the subsequent liquid injectivity in the nearest region to the inlet.

The effect of foam quality was studied; it has been found that foam quality has no big effect during the subsequent gas and liquid injections. The results of foam injection at 0.60 and 0.95 quality followed by gas injection show that foam collapsed after roughly similar number of pore volumes injected regardless of foam quality. During the liquid injection period the trends of the pressure gradient were similar at foam initial quality injection 0.60 and 0.95.

Finally, in order to verify the capability of the radial model developed by Gong et al. \cite{Gong}, the experimental data was fitted to the linear-flow and radial-flow models. The results suggest that the simulation based on Peaceman equation underestimated the gas and liquid injectivity in SAG process, and furthermore, the conventional simulator cannot represent the effect of gas injection on the subsequent liquid injection.

Gas injection processes suffer from poor volumetric sweep efficiency due to phenomena such as viscous fingering caused by density contrast between the injected gas and the displaced fluid, channeling due to heterogeneities in the reservoir and gravity override due to the inherent lower density of gas. One possible way to overcome these negative effects is by foaming the gas. Foam can be defined as a dispersion of gas separated by lamellae in a continuous liquid phase. Lamellae stability is enhanced by adding surfactants to the aqueous phase (Lake, 2014). The underlying idea is that the pitfalls of gas injection can be overcome by trapping the gas and therefore reducing its mobility. The benefits of foam flooding have been reported by several authors (Almaqbali et al., 2017, Li et al., 2009, Chalbaud et al., 2002, Patil et al., 2018 ) however, to implement field scale projects, an accurate representation of foams in models that can be applied to reservoir simulators is of paramount importance. In this sense, two main families of foam modelling exist today. One family referred as population balance models, treat foam texture and bubble size explicitly and another family, that treats foam texture implicitly by applying a mobility reduction factor to gas mobility.
In the first part of this document, a model that treats foam texture implicitly will be used to evaluate foam performance in a heterogenous reservoir consisting of two layers with different permeabilities. The model known as STARS (Computer Modelling Group, 2010) reduces gas mobility using semiempirical functions that represent the physics governing foam texture. One of the functions used by the model is the dryout function, which represents the effects of water saturation on foam coalescence. This function is tuned by a couple parameters, namely, fmdry and epdry that control the water saturation at which foam dries out and how fast it dries out, in other words, how abrupt its collapse is. Farajzadeh et. al, (2015) found that these parameters were permeability dependent.
Inspired by these findings, in the modelling section of this project, foam performance in a two layer reservoir with different permeabilities under varying injection conditions is assessed. Each layer will have its unique set of parameters, namely, fmmob, epdry and fmdry. We are interested in seeing the effects that epdry has on foam performance and the effects of having vastly contrasting foam strengths between layers (large fmmob contrast). The results suggest that epdry can have significant impact on foam performance in the high quality regime. Interestingly, it was found that large fmmob contrast reduce vertical conformance and can reduce recovery efficiency in the high quality regime.
It is known that oil can have negative impact on foam stability, however most studies treat gas and oil as different phases. Under miscible conditions, oil and gas become one phase and the extent to how oil destabilizes foam is not clearly understood. Kahrobaei et. al (2017) found in core flood experiments a unique rheological behavior in foams with miscible mixtures of Carbon Dioxide – Decane. Their experiments showed three distinguishable apparent viscosity regimes depending on Carbon dioxide fraction. Regime 1, in which apparent viscosity increased with increasing CO2 fraction; Regime 2 in which apparent viscosity decreased with increasing CO2 fraction; and Regime 3 in which apparent viscosity is constant independently of the CO2 molar composition.
Inspired by Kahrobaei et. al, (2017) findings, the second part of the project attempts to investigate, using microfluidics, the effects in foam texture caused by compositional changes in an AOS surfactant solution and a miscible Carbon Dioxide – Decane mixture. Unfortunately, due to set-up limitations and inability to reliably mix the components, the experiments were inconclusive. Nonetheless, valuable lessons learnt are presented that will facilitate the experimental approach if someone decides to continue with this line of investigation.

The mobilization of the trapped residual oil is an important part of Enhanced Oil Recovery. The desaturation of nonwetting fluids from porous media is often described using capillary numbers which are a ratio of viscous forces over the capillary forces between the wetting and the nonwetting fluids. Twodimensional microfluidic devices (micromodels) play an important role as they allow for the visualization of the two phase flow in porous media. Even though the existing definitions for capillary numbers work fine for describing the mobilization of the trapped nonwetting phase in geological rock, problems arise when applying these definitions for capillary numbers on micromodels. The conventional definition of the capillary numbers do not allow to visualize the desaturation of nonwetting phases in different micromodels on a single trend. This research presents the derivation of a new definition of a capillary number that can be used to better analyze the mobilization of trapped nonwetting ganglions inside micromodels. The new capillary number is based on a force balance on a trapped ganglion and the corresponding mobilization criteria of the trapped nonwetting phase. The new definition of the capillary number consists of the conventional definition of the capillary number and a geometric term that accounts for the geometry of the micromodel, including the sizes of the pore bodies and throats as well as the ganglion length. The functionality of similar definitions of the capillary number for roughened fractures has previously been published by Al Quaimi and Rossen (2017). The functionally of the new capillary number has been tested by analyzing published desaturation data for various micromodels, each using different wetting and nonwetting phase combinations using both the conventional and the new definition of the capillary number. It is found that the new definition allows for better analysis of the mobilization of the trapped nonwetting fluid inside micromodels, as the desaturation curves for different models can now be visualized by a single trend. Also, it is suggested that the geometrical parameters of the new equation are more significant for describing the mobilization of the trapped nonwetting phase than the of the medium and the interfacial properties of the wetting and nonwetting phases.