Luận văn thạc sĩ Kỹ thuật môi trường: Study on enhanced oil recovery by CO2 microbubbles injection

INTRODUCTION

Research Overview

1.1.1 Enhanced Oil Recovery Using Carbon Dioxide (CO 2 -EOR)

Global energy consumption is expected to grow continuously with an average annual rate change of 1.3 %, from 601.5 quadrillions Btu in 2020 to 886.3 quadrillion Btu in 2050 (Nalley et al., 2021) Hydrocarbon resources (including conventional oil, unconventional oil, and natural gas) have played a critical role in addressing the world’s energy needs until at least the middle of this century (Figure 1.1)

Figure 1.1 Long-term world energy consumption with projection to 2050(adapted from

Figure 1.2 Overview of oil recovery stages

Generally, the production time along the life of conventional petroleum fields has been subdivided into three phases: primary, secondary, and tertiary recovery During the primary stage, hydrocarbon fluids are produced by natural mechanisms, including dissolved-gas drive, gas-cap drive, aquifer drive, gravity drainage, and combined drive (Willhite, 2018) Besides, artificial techniques could be applied to lift the oil in the production tubing once the reservoir pressure reduces Typically, only up to 30% of the original oil in place (OOIP) is produced after this stage In the secondary stage, waterflooding is commonly utilized to force the remaining oil out of the pore structure and maintain the reservoir pressure This approach could recover about 30-50% of OOIP and leave a large amount of oil trapped in the formation The tertiary recovery is considered to

With the increasing impact of global warming over the last decades, more carbon dioxide (CO2) needs to be diminished to a sustainable level as a long-term consequence Hence, carbon mitigation techniques have gained global attention as a potential method to reduce CO2 emissions At present, most research focuses on the topics of carbon capture and storage (CCS) and carbon capture, utilization and storage (CCUS) CCS comprises capturing CO2 generated by burning fossil fuels for energy and sequestering it underground permanently (Dejam et al., 2018a, 2018b; Vo Thanh et al., 2019, 2021; Vo Thanh, Sugai, Nguele, et al., 2020)

While in the CCUS, captured CO2 is effectively utilized in industrial processes (Han et al., 2019) One of the primary techniques of CCUS is the utilization of CO2 for enhancing oil recovery (CO2-EOR)(Mohagheghian et al., 2019; Vo Thanh, Sugai, & Sasaki, 2020) Additional oil could be extracted by utilizing large amounts of CO2 Moreover, depleted oil and gas reservoirs can be potential formations for CO2 storage (Bachu, 2016) From an industrial ecology perspective (Meylan et al., 2015), CO2-EOR is reasonable and sustainable to control global CO2 emissions

The advantages of employing CO2 as a displacement agent during EOR have received significant attention in the petroleum industry An additional benefit is that this method would provide an economical approach to the geological storage of CO2 to reduce atmospheric CO2 concentrations (Azzolina et al., 2015; Mohagheghian et al., 2019; Vo Thanh et al., 2019; Vo Thanh, Sugai, & Sasaki, 2020; Vo Thanh, Sugai, Nguele, et al., 2020) However, CO2 flooding can result in low recovery efficiencies because of the high mobility of CO2 in flooding areas Specifically, the injected CO2 tends to flow through highly permeable layers or fractures, which leads to poor sweeping efficiency in the low permeability zones (Lake, 1989) For this, CO2 foam is often considered to improve mobility control during CO2-EOR operation by increasing gas viscosity and redirecting fluid to low permeable areas (Du et al., 2018, 2020; J Yang et al., 2019) However, the foam becomes unstable under harsh environments Therefore, the stability of foam is a challenge for this application in EOR (Razavi et al., 2020)

1.1.2 CO 2 microbubbles – Colloidal gas aphrons

Recently, microbubbles (defined as having sizes of 10 to 100 μm) have become of interest as a means of removing contaminants from aqueous solutions (Hashim et al., 2012; Molaei et al., 2015), as components of oil well-drilling fluids (Alizadeh et al., 2019; Pasdar et al., 2019; Tabzar et al., 2015a; Zhu et al., 2020) and also with regard to EOR (Natawijaya et al., 2020; Shenglong Shi, Yefei Wang, Shixun Bai, Mingchen Ding, 2017) The use of microbubble-based fluids is growing rapidly in the oil and gas industry One advantage of microbubbles is that they have a unique structure differing from conventional foams that maintain their stability for longer periods under severe conditions

The microbubbles were first reported as colloidal gas aphrons (CGAs) by (Sebba, 1987) CO2 microbubbles comprise a spherical core made of gaseous CO2 with a multilayer covering comprising surfactant molecules and a viscous liquid This multilayer structure, made of an inner layer (between the gaseous core and the liquid layer) and an the bubbles comprise a spherical gas core with a surfactant layer (Telmadarreie et al., 2016)

Wang et al., (2001) applied CGAs to separate the heavy metal elements from the aqueous solution in mineral processing They found that CGAs could have an outstanding performance in CuO flotation at specific operating conditions Waters et al., (2008) evaluated the efficiency of the CGAs flotation system in separating CuO from SiO2 The results revealed that CGAs utilization increased CuO recovery significantly compared with previous methods

Hashim et al., (1998) recommend using CGAs in recovering cellulosic pulp from contaminated effluent Nancy Bjorndalen et al., (2009) pointed out that CGAs fluid can successfully block the micromodel The oil-based CGAs drilling fluid was examined by (Shivhare et al., 2014) They reported that the aphrons exhibit a good plugging performance in porous media and restrict formation damage due to fluid invasion H N Bjorndalen et al., (2014)conducted flow tests using CGAs fluid prepared by polymer and surfactant

It was inferred that CGAs fluid could effectively block the water-wet porous media Pasdar et al., (2019) studied the fluid invasion control ability of CGAs-based fluid using a micromodel system They found that fluid invasion through fracture can be decreased significantly by injecting CGAs fluid Several studies have indicated that microbubbles can seal highly permeable layers in heterogeneous porous media during the EOR process and so improve sweeping efficiency and oil recovery One group applied a microbubble foam to shallow reservoirs and concluded that these microbubbles blocked porous media via the Jamin effect (E Yang et al., 2020) As a microbubble flows through a pore, it will experience a capillary force if its diameter is larger than the pore throat (Wright, 1933) (Shi et al., 2016)conducted double sandpack experiments and determined that microbubbles blocked the high permeability sandpack while increasing the swept volume in the sandpack with lower permeability (Shenglong Shi, Yefei Wang, Shixun Bai, Mingchen Ding, 2017) attempted a micromodel test of plugging performance and showed that microbubbles were capable of temporarily plugging the highly permeable regions such that subsequent flow was forced into the low permeability areas

The CGA flow properties in porous media were also investigated using a modeling approach Alizadeh et al., (2015) developed a mathematical model to predict the stability of microbubbles in drilling fluid in operational conditions of a gas well Alizadeh and Alizadeh et al., (2017) also presented a mathematical model to analyze the transportation of microbubbles in porous media They thought that the invasion of microbubble fluid in porous media is influenced by the ratio of bubble diameter to grain size

Several studies found remarkable stability of the microbubbles compared with conventional foams(N Bjorndalen et al., 2008; Growcock et al., 2004; Pasdar et al., 2018a, 2018c) Ivan et al., (2001) examined the effect of elevated pressure on CGAs and found that these foams remained stable up to a pressure of 10.3 MPa, while Fred Growcock, (2004) demonstrated that CGAs could survive for a significant time span under pressurization as high as 27.6 MPa N Bjorndalen et al., (2008) visually assessed the and reported that these materials were stable up to 13.7 MPa, while (N Bjorndalen et al., 2008) showed that CGAs became unstable at temperatures ranging from 50 to 75 °C The blocking performance of microbubbles is greatly affected by their stability and size distribution (Longe, 1989) and (Jauregi et al., 1997) evaluated the effects of the amount of surfactant on the stability of CGAs, and both concluded that increasing the surfactant concentration improved the CGA stability (Pasdar et al., 2020) showed that increased viscosity also enhanced the stability of CGAs (Tabzar et al., 2015b) performed static drainage tests and observed that the amount of a xanthan gum (XG) polymer in the CGA dispersion played an essential role in conferring stability Overall, the stability of microbubbles appears to be greatly affected by the concentrations of both polymers and surfactants in the foam

Both static liquid drainage (Yan et al., 2005) and bubble size distribution (Pasdar et al., 2018c) can be used to assess the stability of microbubbles The static liquid drainage methods measure the liquid phase volume drained from the microbubble system as a function of time, and several researchers have used this technique to study the stability of CGAs (Yan et al., 2005) proposed an empirical model to characterize the liquid drainage from CGA dispersions, while (Sadeghialiabadi et al., 2015) investigated the effects of geometric and operating variables on CGA stability using the drainage curve method (Tabzar et al., 2020) also studied the stability of nano-enhanced CGAs by monitoring drainage rates In contrast, the bubble size distribution technique evaluates increases in bubble size over time as a measure of stability Several methods have been developed to ascertain bubble size distribution, including visual, electro-resistivity and acoustic techniques (Chen et al., 2017)

Visual methods (including microscopy, photography, and video microscopy) are most frequently used to measure particle and bubble size distributions (Maaref et al., 2018; Moradi et al., 2011) Optical microscopy in particular has been widely employed to ascertain the size and stability of CGAs As an example, Zhu et al (Zhu et al., 2020) determined the bubble size distribution and examined the effect of attapulgite on CGA drilling fluid stability using optical microscopy in conjunction with a Gaussian statistical distribution Parmar et al (Rajeev Parmar, 2015) generated a microbubble suspension by transferring a mixture of gas and liquid to a pressure chamber and found a Weibull distribution of bubble sizes based on image analysis It should be noted that neither of the above two studies employed a goodness of fit test to determine which mathematical distribution function best represented the experimental data Raquibul (Alam et al., 2017) proposed that the bubbles produced in a laboratory-scale electroflotation cell had a log- normal diameter distribution based on high goodness of fit Nevertheless, few reports to date have examined the size distributions of CO2 microbubbles intended for EOR

In addition, there is still disagreement concerning the effects of the surfactant on the microbubble diameter distribution Xu et al (Xu et al., 2009a) reported that increases in the surfactant concentration decreased the bubble diameter, in contrast to the statement concentrations lower than the critical micelle concentration (CMC), increments in the amount of surfactant decreased the CGA bubble size (N Bjorndalen et al., 2008)

Research Objectives

This study primarily aims to investigate CO2 microbubble generation considering the stability, bubble size distribution, rheological property, and EOR efficiency of microbubble solution by conducting a series of laboratory experiments The detail of each objective is presented as follows:

• To investigate the stability and bubble size distribution with the impact of chemical components

• To understand the plugging mechanism of microbubbles in porous media through sandpack flooding experiments

• To evaluate the EOR ability of microbubbles in both homogeneous and heterogeneous formation

Thesis Outline

This dissertation consists of five chapters

Chapter 1 starts with an introduction to the importance of CO2-EOR and emphasizes the advantages of CGAs-microbubbles in the petroleum industry This chapter also discusses the CO2 microbubbles and their characteristics based on previous studies

Chapter 2 presents a method to generate CO microbubbles using a high-speed bubble size distribution was determined by optical microscopy and fitted to theoretical distribution functions

Chapter 3 focuses on the plugging ability of CO2 microbubbles in porous media Firstly, the measurements and analyses have been carried out to examine the characteristics of CO2 microbubbles fluid Then, it describes the flooding experiments under various conditions The effect of gas: liquid ratio, formation permeability, injection flow rate and heterogeneity have been evaluated

Chapter 4 discusses the performance of CO2 microbubbles in enhanced oil recovery The effectiveness of the proposed approach was evaluated by a series of sandpack flooding experiments The results from homogeneous and heterogeneous sandpacks demonstrate its advantages over in oil production improvement

Chapter 5: summarizes the findings from the present research Finally, recommendations are made for further studies.

EXPERIMENTAL DESIGN AND CHARACTERIZATION OF CO 2

Introduction

Previously, there was still disagreement concerning the effects of the surfactant on the microbubble diameter distribution Xu et al (2009b) reported that increases in the surfactant concentration decreased the bubble diameter, in contrast to the statement that the size of CGA microbubbles increased with increasing surfactant concentrations (Arabloo et al., 2014; Pasdar et al., 2018c) N Bjorndalen et al (2008) also showed that, at surfactant concentrations lower than the critical micelle concentration (CMC), increments in the amount of surfactant decreased the CGA bubble size The contradiction in these results shows the necessity of performing additional work to study the effects of surfactant concentration on microbubble size There is also a need for an efficient means of reducing experimental uncertainty when investigating the size distributions of CO2 microbubbles

This chapter examined the effects of the polymer, surfactant and salt concentrations on the stability of CO2 microbubbles using drainage tests This work also employed microscopic imaging together with statistical interpretation to determine the effects of the above parameters on the microbubble size distribution A further objective of this chapter was to obtain a better understanding of the variations in CO microbubble diameter

Materials

The chemicals used in this study include biopolymer Xanthan Gum (XG) and anionic surfactant sodium dodecyl sulfate (SDS, purity ≥ 99.8) CO2 was supplied by a domestic company with a purity of 99.9% Deionized water was used to make the base solutions with specific chemicals concentrations These components are successfully applied in previous studies and provide good generating microbubbles performance (Arabloo et al., 2014; Tabzar et al., 2015a) Sodium chloride (NaCl) was added to examine the effect of salinity on the CO2 microbubble fluids and to prepare the synthetic formation brine All chemicals were supplied by Junsei Chemical (Japan) and deionized (DI) water was used to prepare all aqueous solutions.

Experimental methods

A series of saline solutions was prepared by dissolving specific amounts of NaCl in 300 mL DI water (in Table 2.1) The base solutions were then obtained by adding varying amounts of the SDS and XG polymer to these saline solutions, followed by stirring for 2 h using a magnetic stirrer (MS-H280-Pro Model, Dlab Scientific Inc.) at 1000 rpm to achieve complete dissolution

Table 2.1 CO2 microbubbles with various compositions Sample

2.3.2 Preparation of CO 2 microbubble fluids

Figure 2.1 presents a diagram showing the apparatus used to generate CO2 microbubbles In this process, 200 mL of a base solution was transferred into a 300 mL container, after which CO2 gas (99.9% pure) was injected from the bottom of the container through a diffuser at a certain flow rate using a flow controller The dispersion was subsequently homogenized by stirring at a rate of 8000 rpm for 4 min using an overhead mixer (HG-200 Hsiangtai Model, As One Corporation) The gaseous CO2 diffused into the base solution eventually broke down into microbubbles with micron-scale diameters (Figure 2.2) All experiments were performed at ambient temperature and pressure

Figure 2.1 Schematic diagram of the preparation of CO2 microbubbles: (1)

Homogenizer, (2) Polymer and surfactant solution, (3) Porous stone (gas diffuser), (4)

Gas flow meter, (5) Pressure regulator, (6) CO2 gas tank

Figure 2.2 Apparatus of CO2 microbubbles generation

In preparation for stability tests, a quantity of each CO2 microbubble dispersion was transferred into a 300-mL graduated cylinder and allowed to stand As time passed, the aqueous solution drained from the microbubbles and the volume of this solution was recorded over time The maximum volume of drained liquid (200 mL) was obtained at the point at which the CO2 microbubbles had entirely collapsed (in Figure 2.3) A kinetic model was used to quantify the base solution drainage from each CO2 microbubble dispersion over time This model was previously proposed by Yan et al (Yan et al., 2005) and is based on Equation 2.1:

𝑡 𝑛 +𝑇 1/2 𝑛 , (2.1) where V t (mL) and V F (mL) are the volume of drained solution at time t (min) and the final volume of drained solution, respectively, T 1/2 (min) is the half-life (the time required for the drained liquid to equal 50% of V F ), and n is an exponent that defines the sigmoid character of the model curve When assessing the stability of CO2 microbubbles, a specific drainage rate constant (K) can be obtained by differentiating Equation 2.2 as

Figure 2.3 Schematic process of drainage test

2.3.4 Determination of CO2 microbubble size

The CO2 microbubbles were visualized and the bubble size distributions were evaluated by taking small aliquots of each dispersion from the test containers immediately after preparation of the dispersion and 60 min after preparation Each sample was transferred to a glass microscope slide A transmitted-light microscope with a charge- couple device camera connecting to a desktop computer was used to capture digital images of the CO2 microbubbles

Figure 2.4 Set up for visualization of CO2 microbubbles: (1) Microscope, (2) charge- coupled device (CCD) camera, (3) computer, and (4) glass-slide

Several images were acquired from each specimen for statistical analysis and the average diameters of the CO2 microbubbles in these images as well as the D10, D50 and D90 values were determined Here, D10, D50 and D90 represent the diameters for which 10%, 50% and 90%, respectively, of the microbubbles were smaller in size Figure 2.4 presents a diagram of the microscopy imaging system used to evaluate the CO2 microbubbles

The captured images were processed using the ImageJ software package after being bubbles from each sample to ensure a representative size distribution Figure 2.5 summarizes the enhancement procedure for a typical image

Figure 2.5 Analyzing procedure of a CO2 microbubbles sample (a) Raw image, (b) 8-bit image enhanced by contrast, (c) Image after thresholding, (d) Bubbles counting and analyzing

The output data from the ImageJ software were analyzed in the MATLAB program to obtain each bubble size distribution (BSD) and were also subjected to additional statistical analysis It was essential to determine the exact distributions and so the optimal probability distribution function (PDF) was applied to the experimentally measured size data

Around 300 microbubbles were examined in each sample to estimate the average diameter (D avg ), given by Equation 2.3:

𝑁 (2.3) where D i (m) is the diameter of the observed microbubbles and N is the total number of microbubbles in each sample

Three pdfs were applied to the distributions: normal, log-normal and Weibull A normal PDF is one that conforms to the equation (Pinho et al., 2018):

𝜎√2𝜋 , (2.4) where  is the mean, σ is the standard deviation, and x represents the diameter of a bubble The log-normal PDF is given by (Moradi et al., 2011):

2𝜎2 , (2.5) where σ is the theoretical standard deviation, à is the theoretical mean of ln(x), and x is the diameter of a bubble The Weibull PDF can be written as (Moradi et al., 2011):

𝑓(𝑥) = 𝑏𝑎 −𝑏 𝑥 𝑏−1 𝑒 −𝑥𝑏 𝑎 , (2.6) where x is the diameter of a bubble and a and b are shape and scale parameters, respectively given dataset Higher P-values also indicate better agreement between the data and the theoretical distribution, while a P-value less than a significance level of 5% demonstrates that the experimental data do not conform to a particular theoretical distribution (Alam et al., 2017).

Results and Discussions

Figure 2.6 presents a diagram of an aphron microbubble based on the structure proposed by Sebba along with an optical microscopy image of the present CO2 microbubbles As noted, CO2 microbubbles can provide a CGA in which the microbubbles have a gaseous CO2 core surrounding by a thin aqueous film This thin film is made of surfactant molecules and has three layers (Sebba, 1987) The addition of XG polymer increases the viscosity of the outer film and so strengthens the aphron structure such that the foam can endure harsh conditions such as high pressure and temperature (Pasdar et al., 2018b) Compared with microbubbles, a typical bubble is made by one single covering of surfactant molecules

Figure 2.6 The difference between CO2 microbubbles and conventional foam, introduced in their structure

Figure 2.7 shows the experimental CO2 microbubble drainage process and demonstrates that the dispersion separated into two phases over time Figure 2.8 plots the drainage data for solutions having varying SDS concentrations without the XG polymer as functions of time It can be seen that all the microbubbles collapsed entirely within approximately 20 min in each case Each plot is quite similar, which confirms that (in the absence of the XG polymer) the SDS concentration had only a minimal effect on the stability of the CO2 microbubbles Figure 2.8 also plots K as a function of the SDS more stable dispersions, and the enhanced stability observed at higher SDS concentrations can be attributed to the presence of a greater number of surfactant molecules at the bubble surfaces, which in turn strengthened the microbubble shells and provided good surface elasticity (Yan et al., 2005) These effects delay the liquid drainage and reduced bubble coalescence as a result of greater electrostatic repulsion between the microbubbles (Jauregi et al., 1997)

Figure 2.7 Experimental photograph of CO2 microbubbles drainage with SDS ( 3g/L) and XG (0 g/L)

Figure 2.8 Effect of SDS concentration on the stability of CO2 microbubbles (with 0 g/L

XG) Figure 2.9 plots the drainage data over time for dispersions prepared using various

XG polymer concentrations with an SDS concentration of 3 g/L It is evident from these plots that the XG polymer concentration had a significant effect and that higher XG polymer concentrations improved stability, which means that the drainage rate constant was inversely proportional to the XG polymer concentration Specifically, the K value

−3 −5 −1 −1 dispersions The polymer would be expected to raise the viscosity of the base solution while inhibiting gas diffusion from the core to the bulk liquid, which would consequently stabilize the microbubbles (Hosseini-Kaldozakh et al., 2019) These results are consistent with earlier studies (Pasdar et al., 2018c; Tabzar et al., 2020) examining the effects of polymers on CGA stability, which demonstrated improvements in stability at higher polymer concentrations

Figure 2.9 Effect of XG concentration on the stability of CO2 microbubbles (with 3 g/L

Figure 2.10 Effect of NaCl concentration on the stability of CO2 microbubbles (with 3 g/L SDS and 5 g/L XG) Figure 2.10 plots the drainage data as functions of time for different NaCl concentrations with SDS and XG polymer levels of 3 and 5 g/L, respectively These results confirm that the microbubble stability was slightly increased by increasing the NaCl concentration up to 10 g/L The addition of NaCl likely formed a condensed layer around the bubbles by reducing the electrostatic repulsion between adjacent sulfate ions, which

The high drainage rate associated with reduced stability at greater NaCl concentrations can be ascribed to an increased gravitational effect (Tabzar et al., 2020)

Figure 2.11 shows seven optical microscopy images of CO2 microbubble dispersions

Figure 2.11 Micrographs of CO2 microbubbles: (a) S1 sample, (b) S2 sample, (c) S3 sample, (d) S4 sample, (e) S5 sample, (f) S6 sample, and (g) S7 sample Scale bar: 100

m Figure 2.12 Figure 2.12 Bubble size distribution (BSD curves predicted by Normal, Log-normal and Weibull model for (a) S1 sample, (b) S2 sample, (c) S3 sample, (d) S4 sample, (e) S5 sample, (f) S6 sample, and (g) S7 sample.shows the fitting results obtained using the MATLAB package when applying the normal, log-normal and Weibull distribution functions to the experimental data obtained from the CO2 microbubble dispersions at t = 0 min It is clear from these data that the normal and Weibull density functions provided good fits to the experimental distributions Quantile-quantile (Q-Q) plots are also presented to demonstrate the fitting of the theoretical functions to the experimental data A Q-Q plot is a scatter diagram produced by plotting the experimental data against expected values obtained from the fitting to the distribution Each Q-Q plot includes a straight line at 45° If the distribution function is identical to the experimental data, then the plot should roughly agree with this reference line Figure 2.13 provides such plots for the three theoretical distributions as applied to the seven datasets These Q-Q plots indicate a greater departure from the reference line for all the datasets when using the normal and log-normal functions In contrast, the Weibull distribution approximates the reference line Table 2.2 presents the AD values for the three mathematical distributions as applied to the seven CO2 microbubble samples in this study

For each individual dataset in Table 2.2, a Weibull distribution provided a statistically significant fitting and gave the lowest AD value In addition, the P-values obtained from the AD calculations were consistently higher than 0.05 with Weibull distributions for all seven data sets It can therefore be concluded that Weibull distributions best described the BSDs In contrast, the AD values corresponding to the normal and log- normal distributions were high with very low P-values Therefore, the BSDs of CO2 microbubbles were not well predicted by either type of distribution

Figure 2.12 Bubble size distribution (BSD curves predicted by Normal, Log-normal and

Weibull model for (a) S1 sample, (b) S2 sample, (c) S3 sample, (d) S4 sample, (e) S5 sample, (f) S6 sample, and (g) S7 sample

Figure 2.13 Q-Q plots for (a) S1 sample, (b) S2 sample, (c) S3 sample, (d) S4 sample, (e)

Table 2.2 AD values and corresponding P-values for different bubble size distributions

2.4.4 Factors affecting the BSD of CO 2 microbubbles

Effect of SDS surfactant concentration

Figure 2.14 Influence of SDS concentration (1, 2, 3 g/L) upon bubble size (b) BSD at three SDS concentrations, experimental and fitted results are represented using icons and solid lines, respectively Experiments were conducted with samples having SDS concentrations of 1, 2 or

3 g/L along with a constant XG polymer concentration of 5 g/L The images in Figure 2.14 demonstrate that the bubble sizes in sample S1 were significantly larger than those in samples S2 and S3 Increasing the SDS concentration was also found to reduce the D50 and

D90 values Specifically, the average bubble diameter decreased significantly, from 63.75 to 47.37 m, as the SDS concentration was increased from 1 to 3 g/L As seen in Figure 2.14 a, the proportion of fine bubbles increased remarkably as the surfactant concentration increased and the bubble size distribution was shifted to smaller values and became narrower

Similar trends have been observed in some previous studies (Ahmadi et al., 2015; Chaphalkar et al., 1993; Xu et al., 2009a) and this phenomenon can be attributed to the behavior of the surfactant at the liquid-gas interface Chaphalkar et al (Chaphalkar et al., 1993) reported that increasing the surfactant concentration reduces the interfacial tension between the gas and bulk liquid, which results in a higher probability of breakup and a decrease in bubble size

Effect of XG polymer concentration

Figure 2.15 shows the effect of varying the XG polymer concentrations with a fixed SDS concentration of 3 g/L on the size of the CO2 microbubbles As can be seen in Figure 2.15a, D10 remained relatively constant while D50 and D90 decreased slightly as the polymer concentration was increased Consequently, the average diameter was reduced from 55 to 47.37 m as the XG polymer concentration went from 1 to 5 g/L Figure 2.15b demonstrates that the bubble size distribution shifted slightly toward the lower diameter direction as the XG polymer concentration was raised It is also apparent that the concentration such that the migration of CO2 was inhibited and smaller bubbles were obtained

Figure 2.15 Influence of XG concentration (1,3,5 g/L) upon bubble size (b) BSD at three

XG concentrations, experimental and fitted results are represented using icons and solid lines, respectively

The effect of NaCl concentration on the size of the CO2 microbubbles at an XG polymer concentration of 5 g/L and an SDS concentration of 3 g/L is shown in Figure 2.16a These data confirm that increasing the NaCl concentration did not significantly affect the bubble size up to the addition of 10 g/L NaCl, with a decrease in the average diameter of only 47.38 to 46.11 m However, with a further increase in the NaCl concentration from 10 to 20 g/L, D10, D50 and D90 all increased, while the average diameter increased to 51.90 m The variations in bubble size are also demonstrated in Figure 2.16b The bubble size distribution became much broader at the highest NaCl concentration of 20 g/L and shifted to higher diameters Interestingly, the bubble size distributions were similar at both 0 and 10 g/L NaCl It is evident from Figure 2.11g that there was bubble coalescence at 20 g/L NaCl The observed decrease in the average diameter of the CO2 microbubbles at low NaCl concentrations can be attributed to the double-layer compression resulting from the presence of an electrolyte in the base solution (Jauregi et al., 1997; Xu et al., 2009a) However, because higher NaCl concentrations decreased the viscosity of the solution (Keshavarzi et al., 2020), the average diameter was increased

Figure 2.16 (a) Influence of NaCl concentration (0, 10, 20 g/L) upon bubble size (b)

BSD at three NaCl concentrations, experimental and fitted results are represented using icons and solid lines, respectively

Changes in CO 2 microbubble size over time nearby bubbles The Laplace equation suggests that the pressure difference, P, between bubbles can be expressed as (Xu et al., 2009a):

𝑟 2), (6) where P is the pressure difference between two bubbles, P 1 and P 2 are the internal pressures of bubbles with radii r 1 and r 2, respectively, and  is the interfacial tension A smaller bubble will have a higher internal pressure than a larger one, and so gas will diffuse from the smaller to the larger via the bulk solution As a result, the average bubble diameter will increase as a consequence of the growth of larger bubbles and the loss of bubbles with smaller diameters (Pasdar et al., 2018c; Xu et al., 2009a)

Figure 2.17 presents microscopy images of the CO2 microbubbles in samples S1, S2 and S3 as acquired 60 min after homogenizing The bubble size distributions are also depicted using the fitted Weibull functions Figure 2.17d demonstrates that sample S3 had a narrower size distribution (with 3 g/L SDS) than samples S1 (1 g/L) and S2 (2 g/L) This observation indicates that the sample prepared with a higher surfactant concentration was more stable

Summary

This chapter evaluated the effects of varying the concentrations of a Xanthan gum (XG) polymer, a surfactant (sodium dodecyl sulfate: SDS) and sodium chloride (NaCl) on both the stability and bubble size distribution (BSD) of CO2 microbubbles CO2 microbubble dispersions were prepared using a high-speed homogenizer in conjunction with the diffusion of gaseous CO2 through aqueous solutions The stability of each dispersion was ascertained using a drainage test, while the BSD was determined by optical microscopy and fitted to either normal, log-normal or Weibull functions The results showed that a Weibull distribution gave the most accurate fit for all experimental data Increases in either the SDS or XG polymer concentration were found to decrease the microbubble size However, these same changes increased the microbubble stability as a consequence of structural enhancement The addition of NaCl up to a concentration of 10 g/L (10g/1000g) decreased the average bubble size by approximately 2.7% Stability was also reduced as the NaCl concentration was increased because of the gravitational effect and coalescence.

FLOW PERFORMANCE OF CO 2 MICROBUBBLES IN POROUS

Introduction

This chapter aims to understand better flow restriction caused by CO2 microbubbles in porous media Bubble size distribution, stability, and rheological behavior of CO2 microbubble fluids were thoroughly investigated Besides, we conducted the sandpack flooding experiment to examine the plugging characteristic of CO2 microbubbles The pressure drop along the single sandpack model was recorded to evaluate the blockage efficiency for different gas: liquid ratio of CO2 microbubble fluids, sand pack permeabilities and injection flow rates Furthermore, the fractional volumes produced from dual-core sandpack models (simulated heterogeneous formations) were collected to analyze the swept improvement of CO2 microbubbles.

Experimental section

Silica sands with various meshes of 40-60 (250-400 m), 60-70 (212-250 m), and 100-170 (90-150 m) were used for preparing the different permeability sandpacks

The base fluid was prepared by dissolving an amount of XG polymer (5000 mg/L), SDS surfactant (3000 mg/L) in 100 mL DI water using a magnetic stirrer (1000 rpm for

120 min) The base solution was transferred into a 200 ml container, and CO2 was passed into the base fluid through a diffuser at a 20 ml/min gas flow rate using a flow regulator The dispersion was homogenized at 8000 rpm for 4 min and gas: liquid ratio () of CO2 microbubble fluid ranging from 10% to 40% by adjusting the gas injection time More details were presented in Chapter 3 For this action, the gaseous CO2 entrained into the base fluid subsequently breaks into microscopic bubbles that are stabilized by polymer and surfactant molecules

The stability of CO2 microbubbles was determined by visual monitoring The fresh

CO2 microbubbles were transferred immediately to 10mL measuring cylinders for monitoring Due to all sandpack flooding experiments being carried out within 2 hours,

CO2 microbubbles fluids must maintain their stabilities for more than 3 hours without phase separation

The rheological properties of CO2 microbubble fluids were studied using a viscometer (Brookfield DV-I Prime) at room temperature Fluid samples were placed in the gap between a sample chamber and a rotating cylindrical spindle The change of viscosity and shear stress versus various rotational speeds (from 0.5 to 100 rpm) were recorded Four rheological models, such as Bingham plastic, Herschel-Bulkley, Power- law, and Casson, were presented to estimate the relationship between shear rate and shear stress to investigate the rheological characteristics of CO2 microbubble fluids (Tabzar et al., 2015a)

Bingham plastic model is presented by Equation 4.1:

𝜏 = 𝜏 0 + 𝜂𝛾, 𝜏 0 ≥ 0, 𝜂 > 0 (4.1) where  the shear stress, Pa;  the shear rate, s -1-  the yield stress, Pa;  plastic viscosity, Pa.s

Herschel-Bulkley model is presented by Equation 4.2:

𝜏 = 𝜏 0 + 𝐾𝛾 𝑛 , 𝜏 0 ≥ 0, 𝐾 > 0,0 < 𝑛 < 1 (4.2) where  the shear stress, Pa;  the shear rate, s -1 ;  the yield stress, Pa;   the fluid consistency, n: the flow behavior index

Power-law model is presented by Equation 4.3:

𝜏 = 𝐾𝛾 𝑛 , 𝐾 > 0, 0 < 𝑛 < 1 (4.3) where  the shear stress, Pa;  the shear rate, s -1 ;   the fluid consistency, n: the flow behavior index

Casson model is presented by Equation 4.4:

𝜏 0.5 = 𝜏 0 + 𝐾𝛾 0.5 , 𝜏 0 ≥ 0, 𝐾 > 0 (4.4) where  the shear stress, Pa;  the shear rate, s -1 ;  the yield stress, Pa;   the fluid consistency

In addition, to represent the quality of fit, the coefficient of determination (R 2 ) and root mean square error (RMSE) were used

Coefficient of determination is expressed by Equation 4.5:

𝑖 −𝑦̅) 2 (4.5) where 𝑦 𝑖 : the experimental value, 𝑦̂ 𝑖 : the predicted value, 𝑦̅: the average experimental value

Root Mean Square Error is expressed by Equation 4.6:

𝑁 𝑖=1 (4.6) where 𝑦 𝑖 : the experimental value, 𝑦̂ 𝑖 : the predicted value, 𝑁: the number of data points

The silica sands were cleaned and dried to remove impurities before use Coarse sands (250-400 m) were used to make high permeability sandpacks (2.0 darcy), whereas sands with 212-250 m and 90-150 m were used for medium permeability (1.0 darcy) and low permeability (0.5 darcy), respectively Acrylic tubes were used as sandpack holders in this study Each sandpack was shielded by rubber caps with contributor flowlines Every single sandpack model was tightly packed with fresh sands mixing DI water Then it was thoroughly saturated with DI water by continuous injecting at 1 mL/min to measure permeability and porosity The porosity was then calculated by the mass balance The permeability was obtained based on Darcy’s law measurement

3.2.6 CO 2 microbubble fluid flow tests

Figure 3.1 shows the sketch of sandpack flooding apparatus The description of experimental procedures is highlighted below

Figure 2 Microscopic image and schematic view of a CO2 microbubble

Figure 3.1 Diagram of the experimental setup for CO2 microbubble fluid flow in sandpacks for holding DI water, and pre-generated CO2 microbubbles fluid The pressure change over the sandpack is measured using a pressure transducer (GC31-364, Nagano Keiki Co Ltd., Japan) connected to the computer The sandpacks have a diameter of 25 mm, and a length of 65 mm The effluents are collected by the graduated glass tubes

The following fluid flow experiments were performed to investigate the plugging ability of CO2 microbubbles in porous media (see Figure 3.2) A saturated single-tube sandpack was installed instead of the dual-core sandpack presented in Figure 3.1

Firstly, the DI water was injected into the sandpack at an adjusted flow rate to attain a steady injecting pressure Then, 5 pore volume (PV) of CO2 microbubbles was injected with the same flow rate The pressure drop during the CO2 microbubbles injection phase was recorded By utilizing CO2 microbubble fluid with various gas-liquid ratios, sandpack permeabilities, and injection flow rates, several single-core sandpack flow tests could be performed

Dual-core sandpack test: Four heterogeneous dual-core sandpack models were conducted First, two saturated sandpacks with different permeabilities were arranged parallel and connected to the flooding setup (see Figure 3.1) Next, DI water was injected into the sandpack models (1mL/min) until pressure drop along the sandpacks became stable Then CO2 microbubble fluid was injected at the same flow rate for 4 PV

The effluents were collected by replacing the graduated measuring tube at regular intervals The flows of effluents from the high and low permeability sandpack were collected to measure during injection processes The pressure drop at these phases was also recorded.

Figure 3.2 Flowchart of the experimental procedure.

Results and Dicussions

Figure 3.3 presents the micrographs of CO2 microbubbles with different gas:liquid

Figure 3.3 Micrographs of CO2 microbubbles with different gas:liquid ratios

As illustrated, the CO2 microbubble fluids with various gas-liquid ratios have a similar diameter distribution The bubble size distribution and cumulative bubble size distribution of the four CO2 microbubble fluids are illustrated in Figure 3.4

Bubble size ranges for the prepared fluids with gas-liquid ratios of 10 %, 20 %, 30

%, and 40 % are 11.1 m – 87.6 m, 12.82 m - 106.8 m, 9.3 m - 93.1 m, 8.7 m - 99.9 m, the median diameters (D 50 ) are 41.2 m, 42.2 m, 41.3 m, and 41.2 m, and the average diameters (D avg ) are 41.5 m, 42.3 m, 41.8 m, and 42.8 m, respectively These results reveal that the prepared CO2 microbubble fluids with different gas: liquid ratios had a comparatively similar bubble size distribution

Figure 3.4 Bubble size distribution of CO2 microbubbles with different gas: liquid ratios

Figure 3.5 Stability of CO2 microbubble fluids at different gas: liquid ratios

As one can see, CO2 microbubble samples prepared with different gas:liquid ratios are all stable without phase separation after setting for three hours

Table 3.1 presents the fitting results of four models for the experimental data

Table 3.1 Rheological model’s parameters of CO2 microbubble fluids

Figure 3.6 presents the corresponding statistical svalues (R 2 and RMSE) of these rheology models

Figure 3.6 Values of (a) R 2 and (b) RMSE of fitting models with CO2 microbubble fluids at different gas: liquid ratios

Among four rheology models, the Herschel-Bulkley model shows the highest R 2 values and the lowest RMSE values (Figure 3.6); however, the yield point values ( 0 )are negative, as seen in Table 3.1 Therefore, they are unreasonable results when fitting the nonlinear regression of the shear rate and shear stress Also, the Bingham-plastic model ranks as the worse performance (lowest R 2 values and highest RMSE values) compared

As a result, the Power-law is the robust model for describing the rheology of CO2 microbubble fluids since its good prediction (high quantity of R 2 , low values of RMSE) and practical model parameters

Figure 3.7 presents the experimental shear stress data with fitting curves using the Power-law model versus the shear rate of CO2 microbubble fluids at different gas: liquid ratios

Figure 3.7 Fitting cures of Power-law model for CO2 microbubble fluids at different gas: liquid ratios

The rheological parameters of the Power-law model are illustrated in Table 3.1 The consistency value, K, tends to increase with increasing  The flow index, n, is smaller than 1, which demonstrates that the CO2 microbubble fluids in this study behave as a shear- thining fluid The results also indicate that n remains the same regardless of the gas: liquid ratio (about 0.2) Figure 3.8 illustrates the variation of viscosities with rotational speeds of

CO2 microbubble fluids using the viscometer

Figure 3.8 The plot of viscosity vs shear rate for CO2 microbubble fluids at different gas: liquid ratios

As the rotational speed increases, the viscosities of CO2 microbubble fluids decrease rapidly The fluid with a higher ratio of gas: liquid had a higher viscosity The apparent viscosity (@100 rpm) of CO2 microbubble fluids with gas: liquid ratios of 10, 20,

30, and 40 % are 222.6, 241.7, 243.5, and 250.1 cP, respectively

3.3.2 CO 2 Microbubble Fluid Flow in Homogeneous Porous Media

Figure 3.9 (a) and (b) show pictures of injected and produced CO2 microbubble samples in one typical sanpack flow test

Figure 3.9 (a) Injection CO2 microbubble fluid; (b) Produced fluid from sandpack; (c)

Microscopic image of CO2 microbubbles in porous media

As one can see, the number of CO2 microbubbles in the effluent sample reduces considerably compared with the influent sample Figure 3.9 (c) also illustrates the microscopic image of CO2 microbubbles distributed in porous media It clearly demonstrates the pore throat blocking performance of CO2 microbubbles to restrict the liquid flow in sandpack Figure 3.10 presents a drawing of one CO2 microbubble entering a pore constriction

Figure 3.10 Schematic representation of blockage mechanism as CO2 microbubble enters the pore throat

It will be under the action of capillary resistance force because of the “Jamin effect” (Wright, 1933), which is defined as the Equation 4.7

𝑅 2) (4.7) where P c is the differential capillary pressure caused by the Jamin effect, P 1 and

P 2 are pressure at the front and back of the microbubble R 1 and R 2 are the front and back curvature radius when microbubble is deformed, respectively  is the interfacial tension

Effect of gas: liquid ratio

Figure 3.11 presents the pressure drop by injecting CO2 microbubble fluids with different  of 10%, 20%, 30%, and 40% into 1.0 darcy permeability sandpacks (injection rate of 1 mL/min)

Figure 3.11 Pressure drop changes during CO2 microbubble fluid injecting as a function of gas: liquid ratios

Generally, the pressure drop across the sandpacks significantly increased when CO2 microbubble fluids with  of 10%-40% were injected This is because when CO2 microbubbles go through the small pore throats, they need extra energy to overwhelm the differential pressure caused by the “Jamin effect” (E Yang et al., 2020) As one can see, Figure 3.11 illustrates that CO2 microbubble fluids reveal two types of flow regimes

During the first stage, CO2 microbubbles needed to travel through the whole system and temporarily block the large channel, which caused a sharp growth in the displacement pressure drop After that, CO2 microbubbles were flushed out of sandpack, so the injection pressure increased gradually and approached a stable state By increasing , the concentration of microbubbles in the fluid was increased, resulting in a gradual increase in injection pressure

When more microbubbles accumulate at the inlet of the sandpack, they cause fewer microbubbles to infiltrate into the system Therefore, the fluids have higher microbubbles penetrate sandpack, resulting in more significant fluid flow restrictions because of the cumulative “Jamin effect” For the case of 10%, the pressure drop rose steeply in the initial stage and then reached a plateau of 38 KPa after injecting about 1.3 PV

This indicated that the fluid with a low concentration of microbubbles could breakthrough faster than the other case and reduce the pore throat plugging efficiency Preparing CO2 microbubble fluid with a suitable  value based on a correlation of bubble size, , and permeability could have an insightful meaning to the CO2 microbubble application in EOR operation By considering the injecting and plugging performance, the optimum gas: liquid ratio value is 20%

Figure 3.12 illustrates the impact of permeability on the CO2 microbubble fluids injection  was kept at 20%, and the injection flow rate was set to 1 mL/min

Figure 3.12 Pressure drop changes during CO2 microbubble fluid injecting into sandpack with different permeabilities

As can be noticed, displacing pressure drops increase considerably as sanpack permeability decreases The steady pressure drop increases from roughly 35 Kpa to 40 Kpa by reducing the permeability of the sandpack from 2.0 to 1.0 darcy When the permeability of sandpack decreased to 0.5 darcy, the pressure drop increased quickly and reached a steady stage of 70 Kpa, almost twice larger than sandpack of 0.5 darcy

These results indicate that CO2 microbubbles are more easily transported in high permeability sandpacks This result can be attributed to the fact that the pore throats in low permeability sandpacks have smaller diameters, causing higher capillary pressure As a result, the CO2 microbubbles could penetrate deeper into the high permeability sandpack Therefore, in the case of heterogeneous porous media, CO2 microbubbles could block large channels in high permeable zones and effectively divert the following flow into low permeable zones This evidence is very suitable for fluvial channel reservoirs, whereas the depositional environment made the reservoir more complicated in terms of EOR application Therefore, the plugging effect of CO2 microbubbles could play an essential role in CO2-EOR project in fluvial reservoirs

Effect of the injection flow rate

Figure 3.13 shows the pressure drop results for CO2 microbubble fluid with  20% in 1.0 darcy sandpack at different injection rates (0.5, 1, and 2 mL/min)

Figure 3.13 Pressure drop changes during CO2 microbubble fluid injecting with different flow rates

As can be observed in Figure 3.13, the pressure drop is more significant with higher sandpack Based on this mechanism, the CO2 microbubbles can be transported through the pore throat to entrain into the following pores when the displacing pressure exceeds the resistance force resulting from the “Jamin effect”

Therefore, the increased injection pressure gradient due to the high injection rate would result in more and more CO2 microbubbles overcoming their capillary resistance forces This led to require more CO2 microbubbles to be trapped in porous media To consider CO2 microbubbles for real field study, the injection rate is one of the critical operational control of the EOR project The CO2 microbubbles could replace the conventional CO2 in the CCUS projects Due to the favorable higher injection rate for the CCUS project, the higher injection rate injected inside the reservoirs that drove the higher amount of CO2 could be stored in the hydrocarbon reservoirs In addition, the CO2 microbubbles-EOR has a dual function in the cleaner environment and improving the oil production in the EOR projects

3.3.3 CO 2 Microbubble Fluid Flow in Heterogeneous Porous Media

Introduction

Carbon dioxide - Enhanced Oil Recovery (CO2-EOR) has long been applied in the oil and gas industry This is a practical approach to utilize and reduce the CO2 emissions eliminated from the industrial operation However, injection CO2 gas usually suffers from formation heterogeneity, which results in low oil recovery Therefore, numerous researches have been conducted in order to improve the performance of CO2 gas injection in the heterogeneous reservoir By considering the advantage of CO2 microbubbles, this chapter aims to evaluate the efficiency of CO2 microbubbles flooding on the residual oil recovery within both homogenous and heterogeneous formations.

Experimental section

In this study, a base solution was prepared by mixing Xanthan Gum (XG) polymer (5000 mg/L) and Sodium Dodecyl Sulfate (SDS) surfactant (3000 mg/L) into 100 mL distilled water for 1 hour at room conditions To generate CO2 microbubble fluid, CO2 gas was diffused into the base solution during the mixture was stirred at 8000 rpm for 4 minutes using high-speed homogenizer equipment A similar procedure was presented in previous chapters Then, several microscopic images of CO2 microbubbles were captured using a digital microscope and analyzed by ImageJ software to determine the bubble size distribution Their size varies mostly from 10 to 100 m, and the average size is 42.3 m (in Figure 4.2)

Figure 4.2 Size distribution of CO2 microbubbles

The crude oil sample from a Japanese oil field was used in the experiments The candidate oil had a density of 867.6 kg/m 3 and a viscosity of 9.54 cP A synthetic brine with a salinity of 20000 mg/L was prepared from Sodium Chloride, NaCl and distilled water All used chemicals were supplied by Junsei Chemical Co., Ltd Japan, and the purity of CO2 gas is 99.99%

Figure 4.3 shows the experimental setup of sandpack flooding tests The setup includes a displacement pump, Jacketed cells for holding fluids, a pressure transducer, sandpack holders, a controlled thermostatic bath, and fraction collectors All flooding tests were carried out at 45 o C

The silica sand was tightly packed with brine water in sandpack holders of 6.5 cm in length and 2.5 cm inner diameter The porosity and pore volume were determined by weight method Before packing, fresh sand was cleaned and then dried Silica sands with various grand sizes of 250-400 m, 177-212 m, and 90-150 m were used to prepare the different permeability sandpacks The permeability of each sandpack was measured using Darcy’s Law at a single water flow rate of 0.5 ml/min

Where Q is the flow rate (m 3 s -1 ), k is the permeability (D),  is the water viscosity

First, the sandpacks with different permeabilities were connected and placed in a thermostatic water bath at 45 o C for 1 hour Next, the sandpacks were saturated with the crude oil separately at a 0.25 mL/min injecting rate until oil saturation reached 98% before aging for 24 hours at the same temperature After that, synthetic brine was injected into the model for a certain PV at 0.5 mL/min (for dual sandpack model) or 0.25 mL/min (for single sandpack model) flow rate until oil production was neglected Then, CO2 microbubbles fluid was injected at the same flow rate value Finally, a certain volume of brine was subsequently injected The effluents were collected by using fractionators The volumes of oil and water production were recorded over the period time of flooding experiments

Figure 4.3 Schematic of sandpack flooding test

Table 4.1 Summary of flooding experiments

Chase water flooding Salinity (mg/L

Results and Discussion

4.3.1 Oil recovery in single sandpack

Figure 4.4 Oil recovery performance from single sandpack flooding test

The water injection recovered about 61.4% of the original oil in place (OOIP) in this experiment However, during this flooding stage, there was a dramatic increase of 54

% in oil recovery factor after injection of 1 PV, and then only 7.4 % OOIP was produced between 1-3 PV of water injection Because of capillary force, the injected water preferably flows and displaces the oil in large pore spaces, which causes the flow channeling (see Figure 4.5 a) Besides, the water cut rapidly increased after the water breakthrough and reached 98% after injecting 3 PV of water A low and stable response pressure drop of around 1.7 KPa can aid in further understanding of the result This indicates that significant residual oil is still trapped in smaller pore spaces with higher capillary restrictions

During CO2 microbubble injection (0.5 PV), about 3.3% of OOIP was improved while the water cut declined gradually Furthermore, we observed a sharp increase in pressure drop to around 23 KPa, which indicated the movement and replacement of CO2 microbubbles in pore spaces It can be seen from Figure 4.5 b, that CO2 microbubbles are captured in pore space, resulting in blocking there The maximum pressure drop was reached after CO2 microbubble flooding

At the following stage, chase water flooding was performed The CO2 microbubbles are trapped in pore spaces and divert the major flow, which in terms of increases the sweep efficiency in smaller pore spaces (Figure 4.5 c) Therefore, the chase water flow could sweep and displace the remaining oil out of the sandpack Thus, the curve of oil recovery depicts an increment in oil production of about 23.6% The water cut decreased significantly at the beginning of the chase water flooding stage and then increased to 99%

On the other hand, the pressure drop declines speedily until it reaches 2.1 KPa This states that the chase water was diverted to the small pore spaces region at first due to the plugging of CO2 microbubbles However, the subsequent water flows back to the low capillary force region and bypasses the oil bank (Figure 4.5 c)

Figure 4.5 Oil displacement in micrometers scale and corresponding effluents at the flooding stages ( major flow)

4.3.2 Oil recovery in parallel sandpack

Figure 4.6 shows the fractional flow of the parallel sandpacks when injected base solution with a permeability ratio of 1:4 After injection of a base solution, the fractional flow ratio approaches 70:30 and then bounces back to 90:10, indicating a slight effect on flow diversion The plateau pressure drop in Figure 4.6 shows an insufficient plugging in the high permeability sandpack In Figure 4.7, with CO2 microbubbles injection, the fractional flow ratio is decreased to 40:60 Based on the results, CO2 microbubble fluid demonstrates an adjustable ability to divert the flow and improve oil recovery in heterogeneous sandpacks The black line in Figure 4.7 shows that the pressure drop increases continuously until the later CO2 microbubbles injection, indicating an excellent restriction flow ability For parallel sandpack with a permeability ratio of 1:2, the diversion rates change significantly in both high and low permeability sandpack(in Figure 4.8) This means the fractional flow in low permeability sandpack increased, and the low permeability/ high permeability diversion flow ratio is about 50:50 As shown in Figure 4.8, a favorable displacement was achieved during CO2 microbubble injection in both high permeability sandpack and low permeability sandpack

Figure 4.6 Fractional flow in parallel sandpack with base solution injection (Permeability ratio of 1:4)

Figure 4.7 Fractional flow in parallel sandpack with CO2 microbubbles injection

Figure 4.8 Fractional flow in parallel sandpack with CO2 microbubbles injection

(Permeability ratio of 1:2) The above results demonstrate that CO2 microbubbles have a superior plugging ability to base solution However, CO2 microbubbles do not always ensure excellent flow restriction ability in heterogeneous formation The reason for this is as the permeability ratio increases, more CO2 microbubbles will be needed to flow into the high permeability region to occupy the pore spaces before raising the fractional flow alteration This was correspondingly indicated by the highest pressure growth in the parallel sandpacks with a lower permeability ratio (Figure 4.8)

Table 4.2 presents oil recovery results of parallel sandpack flooding tests Because of the difference in permeability of sandpacks, most produced oil comes from high base solution can restrict the flow in the sandpack with the high viscosity effect (236.9 cp at shear rate of 28 s-1).The injection of base solution can enhance the oil recovery in both high permeability and low permeability sandpack with an increment of 52.81% and 52.64%, respectively Also, it was more interesting to see a 38.48 % and 75.62 % increment in oil production in the high and low permeability sandpack with the injection of CO2 microbubbles, respectively For a permeability ratio of 1:2, up to 30.43% and 48.48% of OOIP were achieved after CO2 microbubble flooding in high permeability and low permeability sandpack, respectively A considerable increment in oil recovery was witnessed after injecting CO2 microbubble (39.27 % and 58.29% for the permeability differences of 1:2 and 1:4, respectively) Notably, the total oil recovery improvement was proportional to the heterogeneity, and most of the produced oil came from the sandpack with lower permeability Since the heterogeneity in the reservoir, the displacing fluid primarily flows in the high permeability zone during the water flooding stage and abandons a large amount of trapped oil in the low permeability zone Thus, it caused the low cumulative oil recovery The volume of unswept oil increase with an increase in permeability ratio Consequently, after CO2 microbubble injecting, the recovered oil in the higher heterogeneity reservoir was more significant

Figure 4.9 illustrates the displacement efficiency during EOR stage of three experiments With the same permeability ratio of 1:4, the total displacement efficiency is higher with CO2 microbubbles injection It should be noted that the base solution injection increases the displacement efficiency in the high permeability sandpack; meanwhile, CO2 microbubbles injecting improves the displacement efficiency in the low permeability sandpack This was because the CO2 microbubbles could fill the pore space, accumulate and effectively bridge the flow channels in the high permeability sandpack Therefore, the subsequent flow is diverted to the low permeability sandpack On the other hand, the base solution can restrict the flow in the sandpack with the high viscosity effect This mechanism may lead to a weak plugging ability in heterogeneous porous media

When comparing the displacement efficiency in the high permeability ratio with that in the low permeability ratio model, the total displacement efficiency is higher in the low model This result indicates that the pore size of the porous media is essential to the mobility control ability of CO2 microbubbles The average pore throat radius of the sandpacks could be determined using Equation 4.2 (Alvarado et al., 1979)

𝜑 (4.2) where r pore represents the average radius of pore throat (m), K is sandpack permeability (m 2 ), and φ is sandpack porosity (dimensionless) The ratio of bubble radius to pore throat radius (R) is calculated using Equation 4.3

The results are presented in Table 4.3 For the parallel sandpacks model (permeability ratio of 1:2), R is high in both the high permeability sandpack and the low permeability sandpack, indicating the plugging performance was dominant It also there is a considerable difference in R-value among the sandpacks ( permeability ratio of 1:4) CO2 microbubbles may penetrate easily in the high permeability sandpack, which causes a decrease in plugging performance As a result, the displacement efficiency becomes lower in the higher permeability ratio

Table 4.2 Flooding results of the parallel sandpack flooding tests

Displacement efficiency * (%) Water flooding After EOR

*Displacement efficiency = 100*(1-Sor/Soi); Soi : the oil saturation at the beginning of the EOR stage; Sor: the residual oil saturation

Figure 4.9 Displacement efficiency for different conditions

Table 4.3 The average pore size and the ratio of bubble size to pore size for sanpacks

Code Permeability ratio Sandpack rpore (m) R(-)

Figure 4.10 Top view captures of parallel sandpack (permeability ratio of 1:2) during

CONCLUSIONS AND RECOMMENDATION

Major findings of the research

The major findings of the present thesis are summarized in the following paragraphs

(1) The present work demonstrated a new system for generating CO2 microbubbles in conjunction with various polymer, surfactant, and salt concentrations

• The majority of the CO2 microbubbles had sizes in the range of 10‒100 m and the size data were well fit using a Weibull distribution

• Surfactant concentration had a considerable effect on bubble size, such that

CO2 microbubbles with smaller diameters were obtained at higher surfactant concentrations Increasing the XG polymer concentration decreased the bubble diameters but narrowed the bubble size distributions

• A stability analysis of the CO2 microbubble samples revealed that increasing the XG polymer and SDS concentrations slowed liquid drainage from the microbubbles The XG polymer concentration had the strongest effect on stability Although the results indicated that the CO2 microbubbles were most stable at an optimal salinity, the highest NaCl salt concentration gave the least stable sample because of the gravitational effect and coalescence

• This work addressed substantial aspects of CO2 microbubbles application in the EOR process, particularly the importance of stability and BSD of the pertinent materials In addition, this study also illustrated the significance of evaluating the goodness-of-fit values for BSD models before assessing the related parameter Such considerations have not been addressed in the previous studies Therefore, the results obtained in this study would be beneficial to assist the development of microbubbles design in oil and gas technology

(2) Several sandpack flooding experiments were systematically designed to investigate plugging characteristics of CO2 microbubbles in porous media Although unconsolidated formations were employed in this work, the dual-core sandpack model with different permeability gives an insight into the movement behavior of CO2 microbubbles in heterogeneous porous media The obtained results can be helpful in the upscaling process from the lab scale to the field scale Based on the experimental results, the following conclusions could be drawn:

• The CO2 microbubble fluids behave as shear-thinning fluid regardless of  value The rheological behavior of CO2 microbubble fluid is described with the Power-law model due to its sufficient accuracy Furthermore, the apparent viscosity of CO2 microbubble fluids increases as the gas:liquid ratio increases

• The gas:liquid ratio significantly affects the plugging ability of CO2 microbubbles in porous media The pressure drops first increased and then decreased with  (range from 10 to 40%), peaking at  = 20% Because of

“Jamin effect”, CO2 microbubbles can temporarily block pore constriction and resist flow in porous media Therefore, the permeability of sandpack significantly influences the plugging performance of CO2 microbubbles Besides, the CO2 microbubble fluid injection with a higher flow rate had a higher pressure drop over the sandpack

• The dual-core sandpack flow tests indicate that CO2 microbubbles have good properties for diverting flow and improving swept volume in the low permeable region of the heterogeneous formation They can make a temporary plugging zone in the high permeability sandpack, and change the porous media, CO2 microbubbles fluid is a good candidate for enhancing hydrocarbon recovery

(3) Homogeneous single sandpack and heterogeneous parallel sandpacks flooding tests were conducted to evaluate fractional flow and oil recovery enhancement of CO2 microbubbles flooding

• The CO2 microbubbles showed an excellent ability to deform themselves and plug the pore throat in porous media At the reservoir temperature, in the homogeneous sandpack model, oil recovery efficiency is more than 26.3% of OOIP over the water flooding because of improved microscopic sweep efficiency caused by pore plugging

• In the heterogeneous model, CO2 microbubbles flooding could significantly improve the displacement efficiency in a low permeability sandpack compared to base solution flooding with the same permeability ratio The CO2 microbubble could adjust to fractional flows in the heterogeneous reservoir and displace the remaining oil

• As a result, the injection of CO2 microbubbles improved the total oil recovery up to 86.9% compared to the injection of base solution with 75.28% in total When the low/high permeability ratio of the parallel sandpack is reduced to 1:2, injecting CO2 microbubbles enhanced the oil recovery to 93.28 % in total The displacement efficiency increases with the decrease of sandpack heterogeneity The results suggest that

CO2 microbubble is favorable to enhanced oil recovery in heterogeneous reservoirs.

S Energy Information Administration, October 2021)