You are on page 1of 20

SPE-200327-MS

Downloaded from http://onepetro.org/SPEIOR/proceedings-pdf/20IOR/1-20IOR/D011S004R001/2367695/spe-200327-ms.pdf/1 by Khalifa University of Science and Technology user on 15 October 2023
Successful Field Implementation of CO2-Foam Injection for Conformance
Enhancement in the EVGSAU Field in the Permian Basin

Amit Katiyar and Armin Hassanzadeh, The Dow Chemical Company; Pramod Patil, Rock-Oil Consulting Group;
Michael Hand, Alejandro Perozo, Doug Pecore, and Hosein Kalaei, ConocoPhillips; Quoc Nguyen, The University
of Texas at Austin

Copyright 2020, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Improved Oil Recovery Conference originally scheduled to be held in Tulsa, OK, USA, 18 – 22 April 2020. Due
to COVID-19 the physical event was postponed until 31 August – 4 September 2020 and was changed to a virtual event. The official proceedings were published
online on 30 August 2020.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may
not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
This paper presents the performance of a CO2 foam injection pilot implemented in the East Vacuum
Grayburg San Andres Unit (EVGSAU) by ConocoPhillips in cooperation with The Dow Chemical
Company. The pilot project focuses on a single CO2 injection pattern, consisting of one injector and eight
producers, selected due to signs of early gas breakthrough and poor overall sweep efficiency.
To solve these conformance issues and increase overall pattern production performance, a new foaming
surfactant with low adsorption and high gas partitioning characteristics was developed and experimentally
tested at simulated reservoir conditions. A "water alternating surfactant-in-gas" injection strategy was
created utilizing a history matched reservoir simulation model and an empirical foam model. This reservoir
model was also utilized to better understand the dependency of surfactant concentration on foam generation
and fluid diversion. Injection profile logs (IPLs) were also run, in both water and CO2 phases, prior to pilot
implementation to establish baseline injection performance.
This paper will detail several performance indicators that illustrate sustained improvement in pattern
injection and production after more than 15 cycles of alternating water, CO2+surfactant, and CO2-only
injection. During each cycle, gas injectivity trends were calculated and compared to the baseline to monitor
foam strength and performance. Four additional IPLs were run, which indicated continuous improvement
in vertical sweep efficiency and ultimately resulted in uniform injection distribution between the upper and
lower sections of the producing zone. Finally, the most significant result of the trial was the uplift in pattern
oil production. It has averaged ~20% above the baseline production forecast throughout the entire pilot
period and peaked within the first six months at ~60% above the baseline.
The success of this pilot illustrates the benefits of using a low adsorbing and CO2 soluble foaming
surfactant to address reservoir conformance issues for CO2 floods. Further optimization of the pilot based
on the simulation forecast is planned to improve long-term pilot economics.
2 SPE-200327-MS

Introduction
In situ injection of miscible supercritical CO2 (scCO2) to reduce residual oil saturation and increase oil
production has been widely applied in the United States since the early 1970s. According to the Department
of Energy (DOE), in 2014 an average of 3.1 billion cubic feet per day (Bcfd) of CO2 was utilized in CO2
enhanced oil recovery (EOR) applications (DOE/NETL-2014/1648), which contributed to about 6% of the

Downloaded from http://onepetro.org/SPEIOR/proceedings-pdf/20IOR/1-20IOR/D011S004R001/2367695/spe-200327-ms.pdf/1 by Khalifa University of Science and Technology user on 15 October 2023
US crude oil production.
Despite the overall success of CO2 EOR, the characteristically high mobility of CO2 compared to that
of the in-situ reservoir fluids can lead to poor areal and vertical conformance and reduce overall sweep
efficiency. Poor injection conformance can result in early gas breakthrough, high gas-oil ratio (GOR), sub-
optimal gas utilization (GUR), and ultimately lower oil recoveries. Water-Alternate-Gas (WAG) injection
has been mostly applied in gas-EOR implementation to limit such conformance issues, however, WAG
injection alone has limited effectiveness (Harpole and Hallenbeck 1996). Other conformance solutions such
as polymer gel injection or well-bore cementing (Sydansk, 1998) are effective if the root cause and location
of the conformance problem is well understood, but these methods are potentially irreversible and are limited
to near well-bore (Seright, 1988). In comparison, CO2-foam EOR (Al Ayesh et al. 2016, Blaker et al. 2002,
Heller 1994, Patil et al. 2018) is a robust technique to control CO2 mobility in regions near the well-bore
as well as deep into the reservoir.
CO2 foam EOR utilizes CO2-in-brine emulsion (foam) generated in situ by injecting a chemical agent
(surfactant) via either CO2 or brine. Figure 1 illustrates an example of how CO2-foam injection improves
CO2 sweep efficiency in a reservoir. The foam forms in the existing high permeability/preferential flow
channels and gas injection is diverted to the un-swept/bypassed oil zones (Heller 1994, Hirasaki and Lawson
1985, Abbas et al. 2012, Abbaszadeh et al. 2014, Blaker et al. 2002, Patil et al. 2018).

Figure 1—Illustration of CO2-foam EOR technology

Previous pilots where CO2-foam technology was implemented in the field have shown success in
achieving mobility control and some conformance improvement (Heller et al. 1985, Holm and Garrison
1988, Jonas et al. 1990, Chou et al. 1992, Stephenson et al. 1993, Hoefner and Evans 1995, Henry et
al. 1996), but despite the well documented technical success, the conformance correction did not always
prove to be economically viable. Some of the factors leading to poor economics in previous trials including
inefficient surfactant foaming due to gas-water segregation and the use of water-soluble surfactants, higher
surfactant adsorption to reservoir rock, and implementation of sub-optimal injection strategies. This paper
SPE-200327-MS 3

presents the performance analysis of a pilot aimed at addressing those key issues, which ultimately led to
a more economic project.

EVGSAU Field Overview


The primary production from EVGSAU field, located in Southeast New Mexico, started in the 1940s. The

Downloaded from http://onepetro.org/SPEIOR/proceedings-pdf/20IOR/1-20IOR/D011S004R001/2367695/spe-200327-ms.pdf/1 by Khalifa University of Science and Technology user on 15 October 2023
Grayburg/San Andres reservoir was then unitized in 1978 with the aim of flooding the unitized intervals
with water and gas. The unit includes both the Grayburg and San Andres formations, but the main zone
of interest is the San Andres, and the majority of the injection and producing wells are perforated in this
formation (King 1972). Although surrounding units began water injection at an earlier date, water flooding
in EVGSAU did not begin until 1980s. The goal of water flooding was to displace oil and increase reservoir
pressure above minimum miscibility pressure (MMP) in preparation for CO2 flooding, which began in
1985 with a WAG ratio of 2:1. CO2 flooding has continued since that time and has recovered an additional
12-13% of the original oil-in-place (OOIP) in the unit. Details regarding the implementation of the WAG
flood in EVGSAU have been previously published (Harpole and Hallenbeck 1996). Some of the difficulties
encountered during the EVGSAU CO2 flood include low reservoir pressure in parts of the unit and low
sweep efficiency of the CO2. Even with these difficulties, unit oil recovery has exceeded 150 MMSTBO
and the overall oil recovery factor in the CO2 flooded areas is above 55 % OOIP.

Table 1—EVGSAU reservoir properties

Across the northern and central portion of the field, the San Andres formation is divided into the upper
and lower intervals by the Lovington Sandstone (Figure 2). In contrast to the Upper San Andres (U-SA),
the Lower San Andres (L-SA) is characterized by thicker, higher permeability flow units that display at
least some degree of vertical communication. However, in general, the EVGSAU field is characterized with
poor vertical communication, and injectivity in the L-SA has historically been higher than that in the U-SA
(Harpole and Hallenbeck 1996). As noted in Table 1, the geology does have anhydrite minerals which can
be dissolved over time in a WAG flood, causing further conformance problems. In a previous study, it was
noted that within 10 years of the WAG flood in EVGSAU, flow channels or thief zones with up to 200 mD
permeability were created, causing severe vertical conformance issues. These severe conformance issues
led to a Department of Energy (DOE) funded CO2-foam treatment conducted in the mid 1990's (Martin
1995) that was deemed a technical success but was ultimately unsuccessful due to poor economics driven
by high surfactant adsorption.
4 SPE-200327-MS

Downloaded from http://onepetro.org/SPEIOR/proceedings-pdf/20IOR/1-20IOR/D011S004R001/2367695/spe-200327-ms.pdf/1 by Khalifa University of Science and Technology user on 15 October 2023
Figure 2—3229-W007 gas injection profile log illustrating the flow
units that makeup the primary producing formation in EVGSAU

Product Characteristics and Its Experimental Evaluation


After analyzing the limitations of previous surfactants used in scCO2 foam injection projects, Dow
successfully developed a new surfactant chemistry with the following characteristics.

Highly stable at formation salinity and temperature


To ensure that the new surfactant was going to be suitable for use in a field pilot, Dow performed aqueous
stability experiments in a sample of synthetic brine with ~12.5% total dissolved solids (TDS), designed
to closely mimic the field brine at EVGSAU. The cloud point temperature of the selected surfactant was
found to be ~ 20 °F above the reservoir temperature (101°F) at EVGSAU. This experiment indicated that
the surfactant would remain stable under real reservoir conditions.

Preferential partitioning and solubility in scCO2


Conformance issues can vary slightly within a single wellbore depending on the type of fluid being injected,
either water or scCO2, due to the differences in fluid mobility. Although it is important to achieve as efficient
injection as possible for both fluids, scCO2 is the main fluid of interest because of its EOR characteristics.
Dow designed the new surfactant chemistry to have a higher partition coefficient in scCO2, allowing it
to be pumped downhole in the scCO2 cycle instead of the water cycle, which is a key improvement over
surfactants used in previous pilots. The surfactant injection in scCO2 limits the adverse effects of gas-water
segregation on in-situ foam generation and promotes deeper foam transport (Ren et al. 2013).
SPE-200327-MS 5

The lab experiments used a PVT cell to measure the surfactant partition coefficient between brine and
scCO2 at simulated reservoir conditions of 101°F and ~2000 psi. First, the cell was loaded with brine and a
known concentration of surfactant solution at reservoir pressure and temperature. Next, the same volume of
CO2 was bubbled through the cell containing the brine/surfactant solution, and the three fluids were kept in
the cell for 24 hrs to allow for equilibration. Finally, small volumes of the solution were aliquoted out of the

Downloaded from http://onepetro.org/SPEIOR/proceedings-pdf/20IOR/1-20IOR/D011S004R001/2367695/spe-200327-ms.pdf/1 by Khalifa University of Science and Technology user on 15 October 2023
cell in steps, while maintaining constant pressure and temperature, to quantify the surfactant concentration in
brine at equilibrium. The surfactant partition coefficient at reservoir conditions was calculated as the ratio of
the mass fraction of surfactant in scCO2 to the mass fraction of surfactant in brine. The results indicated that
the mass fraction of surfactant in the scCO2 was 22 times greater than the mass fraction in the brine, meaning
the surfactant achieved the desired outcome of preferential partitioning in scCO2. Tests also indicated that
surfactant solubility in scCO2 improves with increasing pressure, which indicates that it can potentially be
transported deeper into the reservoir and not just near wellbore for conformance improvement.

Low adsorption to reservoir rock


One of the most important characteristics of the new surfactant chemistry is the low adsorption rate to
reservoir rock. This low adsorption rate creates a stronger foam capable of propagating deep into the
reservoir and improving overall project economics because a higher volume of surfactant is consumed
generating foam rather than adhering to the rock surface.
The dynamic adsorption of the surfactant at reservoir conditions and proposed injection concentration
was measured to be ~0.05 mg/g of rock. This adsorption corresponds to ~350 lbs./Acre-ft. of surfactant loss
in comparison to ~3500 lb/Acre-ft measured for the surfactant deployed in the previous EVGSAU foam
trial (Tsau and Heller 1994).

Stronger foam generation at lower surfactant concentration


Strong foam generation between brine and scCO2 in varying reservoir conditions is another important
requirement for surfactant selection. In order to test the strength and quality of foam generation, surfactant
flood experiments were performed in core samples at reservoir pressure and temperature. Foam was
generated by co-injecting surfactant, brine, and scCO2 through the core samples while adjusting, one at
a time, the surfactant concentration, injection gas fraction, and total injection rate. Secondly, the steady
state pressure drop across the core was recorded to calculate apparent foam viscosity. Finally, crude oil
was introduced into the core sample experiments to study the effects on foam strength. The following foam
behavior was observed:

• As surfactant concentration increases, apparent foam viscosity measures 2-3 orders of magnitude
greater than the viscosity of brine or scCO2 alone.
• As total injection rate increases, shear thickening rheology was observed up to an optimal injection
rate, but shear thinning rheology was observed as injection rates increased to above optimum rate.
• There is an optimal injection gas fraction to achieve the highest quality foam. As the injection gas
fraction increases beyond this optimal point, the foam begins to dry out and foam strength begins
to deteriorate.
• The final core flood experiment involved the introduction of crude oil into the system. These tests
indicated that the foam destabilizes when it encounters a sufficient volume of mobile oil (~10-15%
mobile oil fraction). This is a favorable characteristic because in a field implementation, the foam
will destabilize as the CO2 is diverted into previously bypassed areas of the reservoir with higher
oil saturations.
6 SPE-200327-MS

Pilot Pattern Selection


Figure 3 shows a map of the EVGSAU field as well as the location and layout of the pattern selected for the
foam injection pilot. EVGSAU is divided into several injector-centered WAG patterns in the center of the
field and water-only injectors around the periphery of the lease that act as a "water barrier" to maintain CO2
injection within the lease boundary. The pilot pattern is made up of a single WAG injector (3229-W007)

Downloaded from http://onepetro.org/SPEIOR/proceedings-pdf/20IOR/1-20IOR/D011S004R001/2367695/spe-200327-ms.pdf/1 by Khalifa University of Science and Technology user on 15 October 2023
with six surrounding vertical wells, which is a similar pattern alignment to other patterns in the field. There
are also two horizontal wells that have at least a portion of their horizontal sections inside of the pattern, and
for that reason they were included in the analysis; however, the team expected most of the foam response
to be observed in the vertical wells.

Figure 3—EVGSAU Field Map showing location and layout of the foam pilot pattern

Vertical conformance characterization


One of the key factors that went into the selection of this pattern was the 3229-W007 injector's history of
poor vertical conformance. As shown in Table 2, multiple injection profile logs (IPL) have been collected
in the water and CO2 cycles for the 3229-W007 injector since the early 1980s. Table 2 also shows the
breakdown of the various flow units that make up the Upper and Lower producing formations. The U-SA
is made up of layers A through E while layers G through I make up the L-SA; layer F is the Lovington
Sand formation which typically acts as a flow barrier, and prevents vertical communication between the U-
SA and L-SA formations. Based on the well logs, Layer A was characterized as having minimal reservoir
potential, which is why there are no perforations in that layer. Out of layers B, C, D and E in the U-SA
formation, layer C has historically been characterized as a major flow unit (Harpole et al. 1994b). In the
L-SA, there are only perforations in layer G and the upper portion of layer H, so any injection below the
deepest perforation in layer H is characterized as out of zone (OOZ) injection loss.
The IPLs collected after 1990 show that a significant percentage of water/gas started to be injected OOZ
(Table 2), which was likely caused by a casing integrity issue near the bottom of the wellbore (i.e. hole or
split in casing). Prior to starting the surfactant injection, additional IPLs were collected in both the water and
gas injection cycles to establish a new baseline injection distribution. The IPL collected during the water
cycle showed that 77% of the water was being injected into the L-SA and the majority of that was being
SPE-200327-MS 7

injected out OOZ (Table 2). The IPL run during a CO2 cycle indicated the same conformance issue with
75% of the gas being injected into the L-SA and 46% being injected OOZ (Table 2).

Table 2—Summary of 3229-W007 IPL data

Downloaded from http://onepetro.org/SPEIOR/proceedings-pdf/20IOR/1-20IOR/D011S004R001/2367695/spe-200327-ms.pdf/1 by Khalifa University of Science and Technology user on 15 October 2023
The complete IPL history was then combined with hydrocarbon pore volume (HCPV) data previously
calculated from well logs to estimate the cumulative hydrocarbon pore volume injected (HCPVI) in both the
U-SA and L-SA formations. Figure 4 clearly shows that because of the poor vertical conformance, the U-
SA has had <30% HCPV of CO2 injected while the L-SA has received most of the historical injection with
>60% HCPV of CO2 injected. This large variance in injection illustrated that the U-SA remained relatively
unprocessed from an EOR standpoint, and if the foam surfactant was able to re-direct CO2 from the L-SA
to the U-SA, there was an opportunity to significantly improve oil recovery in the pattern.
8 SPE-200327-MS

Downloaded from http://onepetro.org/SPEIOR/proceedings-pdf/20IOR/1-20IOR/D011S004R001/2367695/spe-200327-ms.pdf/1 by Khalifa University of Science and Technology user on 15 October 2023
Figure 4—Percent HCPV injected between the upper and the lower San Andres

Areal Conformance Characterization


Areal conformance and gas breakthrough were other issues that led to the selection of this pattern. The
team compared the gas-oil-ratios (GOR) and oil cuts during the baseline period (2014-2017) for all of
the producing wells in the pattern, and found that there were wells such as the 3236-009 that showed
a higher than normal GOR (>10 mcf/bbl) and lower than normal Oil cut (<10%) compared to the rest
of the pattern. This variance from the pattern average indicated the potential that some producers had
experienced significant CO2 breakthrough and that the pattern most likely suffered from sub-optimal areal
conformance. Gas breakthrough not only leads to poor sweep efficiency throughout the pattern, but also
lowers runtime in the producers due to gas interference in the Beam Pump and ESP artificial lift systems.
This secondary conformance issue provides a further opportunity to improve pattern performance through
the implementation of foam injection.

Pilot Injection Strategy Development


Objectives and Success Criteria
The foam pilot injection strategy was developed with the main objective being to improve overall
sweep efficiency across the U-SA and L-SA formations by correcting the vertical and areal conformance
issues mentioned above. The pilot would only be deemed successful if the surfactant injection achieved
sufficient incremental oil production to make the project economic, however, other factors such as vertical
conformance correction from IPLs, improved well runtime, and injectivity changes were key performance
metrics to indicate technical success.

Injection Strategy
The CO2 flood at EVGSAU operates as a constant pressure-controlled injection system and the available
injection pressures for water and CO2 are ~1,300 and 1,800 psi respectively. Once surfactant injection starts
and foam begins to form downhole, the gas injection rate within the pattern should begin to drop due to
the reduced gas mobility. Although the change in mobility may improve the volumetric sweep, the lower
gas injection rate may adversely impact production due to the reduced volume of CO2 contacting oil in the
reservoir. To best address this issue, the team developed a surfactant injection strategy with the following
WAG constraints:
SPE-200327-MS 9

• The total fluid injection in a year would remain the same as the baseline period to maintain reservoir
pressure and CO2 processing rate; this was achieved by increasing the GWR (gas-water ratio) in
each WAG cycle compared to the baseline. Relatively higher GWR not only limits the loss in gas
processing rate but also helps achieve a higher gas-water mixing ratio that generates stronger foam.
• WAG cycle duration was reduced to the lowest operationally feasible time to mimic the same co-

Downloaded from http://onepetro.org/SPEIOR/proceedings-pdf/20IOR/1-20IOR/D011S004R001/2367695/spe-200327-ms.pdf/1 by Khalifa University of Science and Technology user on 15 October 2023
injection environment observed in the core flood experiments. These shorter cycles maximize the
impact of surfactant in generating stronger foam.
• Each foam WAG cycle would follow the same injection pattern. The cycle starts with a fixed
volume slug of water, followed by surfactant injected with CO2, and finally surfactant injection is
stopped, and pure CO2 is injected to meet the GWR target. The initial target volume of CO2 with
surfactant (i.e. surfactant concentration) was about 20% of the total gas volume in a single cycle.
• In order to definitively attribute any oil production uplift to the foam injection at 3229-W007 and
not changes from any other outside factors, the four offset injection patterns were maintained at
their baseline WAG schedules throughout the entirety of the pilot.

Pilot Performance Analysis


Surfactant injection began in January 2018 and more than 12 cycles of alternating water, CO2+surfactant,
and CO2-only injection have been completed. Throughout the pilot period, overall performance was
monitored by analyzing the following four key data sets:

• Gas injectivity – foam generation and mobility control near wellbore.

• Gas IPLs – vertical conformance correction.

• Well runtime – gas interference in high GOR wells

• Well tests – incremental oil recovery, GOR, and oil cut.

Gas Injectivity
The baseline gas injectivity was calculated by taking the arithmetic average of daily gas injectivity for the
six latest gas cycles in the three-year baseline period (2015-2017). The two figures below show that the
baseline gas injectivity at the beginning of a gas cycle is ~0.3 bbl/d/psi and increases to ~1 bbl/d/psi over
a period of ~35 days before eventually plateauing at ~1.0-1.2 bbl/d/psi. This injectivity behavior is typical
of all the WAG patterns in EVGSAU when switching from water to CO2, and it is caused by the injection
gas gradually displacing the column of water near wellbore.
Figure 5 shows the calculated gas injectivity for each of the first six cycles of the pilot. For each of the
cycles there was an immediate drop in injectivity (>50%) compared to the baseline, which indicates strong
foam generation downhole. However, in most of the cycles, gas injectivity began to return to baseline values
after a week of surfactant injection (Cycle 3, 5, and 6 injectivity remained lower due to factors outside of
the pattern operation i.e. lower field injection pressure). The team believed that this was due to the presence
of larger void space conduits that existed further out into the reservoir. As the foam slug propagated deeper
into the reservoir, there was not enough foam in the early stages of the pilot to fill the larger void spaces;
therefore, the CO2 was able to regain mobility and the injectivity returned to baseline values.
10 SPE-200327-MS

Downloaded from http://onepetro.org/SPEIOR/proceedings-pdf/20IOR/1-20IOR/D011S004R001/2367695/spe-200327-ms.pdf/1 by Khalifa University of Science and Technology user on 15 October 2023
Figure 5—Cycle 1-6 gas injectivity vs. baseline

During Cycle 7, the 3202-012 producer was taken offline to perform a workover to optimize the downhole
production equipment. While working on the well it was discovered that the well suffered from a collapsed
casing that could not be repaired, and the well eventually had to be plugged and abandoned (P&A). The
overall production impact of the 3202-012 will be discussed in detail in a later section of the paper, but at
a high level it is important to note that this well improved the most after surfactant injection started, so the
team believed that there was direct communication between the 3202-012 and the injector through large
void space conduits. As shown in Figure 6, the cycles following the P&A of 3202-012 maintained reduced
injectivity for the entirety of the cycle, which is a distinct change in behavior compared to cycles 1-6. The
team hypothesized that after the 3202-012 was P&A'ed and the localized reservoir pressure around that
wellbore began to increase, there was a period of areal injection re-distribution that occurred, and gas was no
longer flowing through the conduit that existed between 3202-012 and the 3229-W007 injector. This change
in injectivity behavior was also observed in the production response which will be covered in a later section.

Figure 6—Cycle 8-14 gas injectivity vs. baseline


SPE-200327-MS 11

Injection Profile Logs – Conformance Improvement


Four IPLs were collected during the foam pilot to evaluate the extent of the vertical conformance
improvement within the 3229-W007 injection well (Table 3). The first two IPLs were collected during
the 3rd and 5th cycles immediately after shutting off the surfactant injection and transitioning into the CO2
only portion of the cycle. The timing of the first two IPLs was designed to determine the injection profile

Downloaded from http://onepetro.org/SPEIOR/proceedings-pdf/20IOR/1-20IOR/D011S004R001/2367695/spe-200327-ms.pdf/1 by Khalifa University of Science and Technology user on 15 October 2023
immediately after generating the foam, when it would theoretically be at maximum foam strength. The
3rd IPL was also collected during the 5th foam cycle, however, it was collected at the end of the CO2 only
portion of the cycle in order to evaluate how well the foam strength was maintained after stopping surfactant
injection. The 4th IPL was taken during the 11th foam cycle, more than 1.5 years after the previous IPL, and
was a clear illustration of how the foam maintains conformance improvement over long periods of time
with continued application.

Table 3—Injection profile log data by formation zone – baseline vs. foam pilot

The IPL data illustrated the following key improvements:

• OOZ injection losses were eliminated - The first IPL indicated that there was no more gas was
being injected OOZ below layer H, and this remained true throughout all the subsequent IPLs.
• More uniform injection distribution between U-SA and L-SA - Due to the elimination of the OOZ
injection, 21% of the gas that had previously been lost was re-distributed into both the U-SA and
the L-SA
• Conformance improvement was maintained over time – The 4th IPL was a positive indication that
foam strength and conformance improvement can be maintained for long periods of time with the
continued application of the surfactant-based conformance solution.

Improved Well Operation


As mentioned previously, all the producers within the pattern use some form of artificial lift, either an ESP
or a BPU. In both systems, excess gas in the pump from CO2 breakthrough results in a decrease in pump
efficiency and runtime, which can ultimately lead to higher bottom hole pressures and lower production
rates.
One of the first benefits of foam injection was observed through improved well runtime in wells that
historically had higher average GOR values, such as 3236-009, which most likely had issues with gas
breakthrough from the injector. Figure 7 shows that 3236-009 had an average runtime of ~15% during the
baseline period, but after surfactant injection began and direct gas breakthrough was mitigated, the runtime
increased to an average of 60% within the first six months of surfactant injection. This increase in runtime
allowed the well to run more efficiently, draw down producing bottom hole pressure, and realize immediate
incremental oil production.
12 SPE-200327-MS

Downloaded from http://onepetro.org/SPEIOR/proceedings-pdf/20IOR/1-20IOR/D011S004R001/2367695/spe-200327-ms.pdf/1 by Khalifa University of Science and Technology user on 15 October 2023
Figure 7—Monthly average percent run time for 3236-009

Pattern Production Analysis


Figure 8 shows that the combined impact of the pattern injectivity, vertical conformance, and well operation
changes that were discussed in the previous section resulted in significant pattern uplift. The uplift has been
sustained throughout most of the pilot period and was enough to make this project both a technical and an
economic success. The following points provide a summary of the pattern production improvements:

• Pattern production peaked at 60% above baseline within 6 months.

• Average pattern uplift throughout the pilot has been 20% above baseline.

• Baseline decline was adjusted after the 6th cycle due to the 3202-012 shut-in and subsequent P&A.

• Gas Utilization Ratio improved by 16% indicating a more efficient use of injected gas.

• Overall pattern recovery to date has been 31 MBO above the baseline.
SPE-200327-MS 13

Downloaded from http://onepetro.org/SPEIOR/proceedings-pdf/20IOR/1-20IOR/D011S004R001/2367695/spe-200327-ms.pdf/1 by Khalifa University of Science and Technology user on 15 October 2023
Figure 8—Pattern Production and GOR

Individual well oil production response


When analyzing the well level data, the surfactant injection did not influence all the producers in the same
way in terms of timing and magnitude of the production uplift. Some of the wells saw significant uplift
immediately after starting surfactant injection, while others did not see uplift until several months into the
pilot, and a small portion of wells saw almost no change at all. This variation in impact is due to several
factors including location relative to the injector, the presence of fractures or other void space conduits in the
reservoir, and injection strategy. Figure 9 as well as the points below summarize some of the observations
from the individual well production analysis:

• Wells such as the 3236-009 and the 3202-012, which were identified as high GOR "problem" wells
prior to starting the pilot, were the first to see uplift after starting surfactant injection. Both wells are
offset to the 3229-W007 injector in an E-W direction, which has been identified across this portion
of EVGSAU as the dominant permeability direction; therefore, it makes sense that the surfactant
injection would address these high permeability channels first.
• Wells such as the 3202-019 that lie N-S relative to the injector (perpendicular to dominant
permeability direction) did not see production uplift until several months after injection. This
difference in uplift timing is potentially an indication that the foam works its way through the
reservoir over time addressing the highest permeability channels first and is capable of eventually
achieving improved areal sweep efficiency.
• The two horizontal wells, 3236-001 and 3229-013, did not seem to exhibit any change in production
related to the surfactant injection. Only a small section of these wellbores lie within the pattern
boundary, so the team included them in the analysis but did not expect to see any significant
impacts.
14 SPE-200327-MS

Downloaded from http://onepetro.org/SPEIOR/proceedings-pdf/20IOR/1-20IOR/D011S004R001/2367695/spe-200327-ms.pdf/1 by Khalifa University of Science and Technology user on 15 October 2023
Figure 9—Examples of individual well responses to the foam injection

Simulation of Foam Pilot


Reservoir modeling and simulations were utilized to better understand surfactant injection effects on foam
generation and fluid diversion around the subject injector in the pattern. The reservoir model was first history
matched to the historical production data and pressure measurements. A lab-scale empirical foam model
was developed from core flood foaming experiments. The foam model parameters were then regressed by
history matching zonal gas flow distribution from post-foam IPLs (Table 3). In this process, key foam model
parameters were kept the same as obtained from lab work. Lastly, the developed model was used to forecast
the effect of surfactant injection concentration on the pilot outcome.

Geological Model
The original geological model was generated based on available field geological data. This model was
carved out of a wider geological model to mainly focus on foam pilot pattern while including at least one
offset injector on each side (Figure 10). The model consists of uniform grid spacing of 150 ft × 150 ft in
the areal direction and 57 layers with variable depth in the vertical direction. The layers were set based on
permeability to porosity ratio. The average areal permeability of the area covered by the model is ~10 md
and the vertical permeability were set at 5% of the areal permeability for all the grid cells.
SPE-200327-MS 15

Downloaded from http://onepetro.org/SPEIOR/proceedings-pdf/20IOR/1-20IOR/D011S004R001/2367695/spe-200327-ms.pdf/1 by Khalifa University of Science and Technology user on 15 October 2023
Figure 10—Geological reservoir model

Pre-Foam History Matching


History matching of the reservoir model was performed to honor production data during water flooding and
CO2 flooding for oil, water, and gas for eight producers of the pilot pattern. The average reservoir pressure
was first matched by adjusting producer well constraints in offset patterns throughout the model. It is crucial
to account for uneven injection distribution among different reservoir layers in the model and correctly
model the conformance issue recorded in IPL data. Absolute permeability of reservoir was adjusted by
applying permeability multipliers on a layer by layer basis. The magnitude of the multiplier varied between
1 and 14 to satisfy flow distribution among different layers. The above modification to the model resulted
in a reasonable match of the model prediction to IPL data (Table 4).

Table 4—Model prediction vs. IPL – before and after surfactant injection

The results of history matching the oil, water and gas production are shown in Figure 11. The cumulative
result is shown for all eight producer wells in the pilot. Gas production in the horizontal well (3236-001)
caused the model to overpredict gas production as shown in the figure.
16 SPE-200327-MS

Downloaded from http://onepetro.org/SPEIOR/proceedings-pdf/20IOR/1-20IOR/D011S004R001/2367695/spe-200327-ms.pdf/1 by Khalifa University of Science and Technology user on 15 October 2023
Figure 11—History match of cumulative Production (oil, water and gas) of 8 pattern producers

Foam Pilot History Matching


Foam model parameters for dependency of foam stability on surfactant concentration, water saturation and
fluid shear were obtained from the lab experiments. These parameters were directly used in the reservoir
model. The effect of oil saturation on destabilizing the foam is another key parameter in the model, and
it is important to understand that oil composition plays an important role on this parameter. For instance,
it was observed in the model that the oleic phase around the injector mainly contains CO2 (>80 mole %).
This is due to a long exposure of the oleic phase to scCO2 that resulted in altering composition of the oleic
phase. It is expected that foam stability would not be impacted by exposure to such oil compositions. Based
on this observation, the oil saturation dependency function was set as one of the adjustable parameters of
the foam model.
Moving from lab to field, the developed parameters of the foam model should be calibrated by introducing
a tunable foam mobility reduction factor. This factor, along with the oil saturation dependency parameter,
was used to obtain the best match to the IPL data from the field. The goal was to match gas distribution
between the Upper and Lower San Andres for all the IPL data points shown in Table 4.
All the parameters described above contributed in reduced mobility of the gas phase, mainly in grid blocks
around the pattern injector 3229-W007. The gas mobility multiplier varied between 0.04 and 1 with the
lowest value associated to the OOZ layers (thief layers) at the very bottom of San Andres pay zone (Figure
12). Due to a relatively higher gas flow throughout the OOZ layers, the effect of surfactant concentration
and fluid shear on reducing gas mobility are the highest in these layers, while the effect of oil saturation
in destabilizing the foam is low. This combination results in the mobility multiplier to reach relatively low
values in this region.
SPE-200327-MS 17

Downloaded from http://onepetro.org/SPEIOR/proceedings-pdf/20IOR/1-20IOR/D011S004R001/2367695/spe-200327-ms.pdf/1 by Khalifa University of Science and Technology user on 15 October 2023
Figure 12—Gas mobility multiplier after surfactant injection (Sept 2019) around
the pattern injector (Note: The multiplier in the grid blocks with no color is one).

The results shown in Figure 12 and Table 4 confirm the capability of the empirical foam model to history
match the vertical conformance correction as surfactant was injected in the reservoir. This model predicts a
significant change in areal distribution of gas flow around the pilot injector (areal conformance correction).
As shown in Figure 13, a strong gas flow connectivity is predicted by the model between the pilot injector
and the producers located on the west side of the pattern. The length of the arrows shown in this figure
represents the intensity of gas flux in each grid bock. Comparing the graphs from before and after surfactant
injection, there is a clear reduction in gas flux towards the west side of the pattern while the gas saturation
is about the same. This result agrees with the pump runtime data for producer well 3236-009 (Figure 7).

Figure 13—Aerial view of gas flux (arrows) and gas saturation (color) in an OOZ layer, before and after surfactant injection

Model Forecast
The developed model was utilized to forecast surfactant performance under four different injection
concentrations (Table 5). Gas flow distribution between the Upper and Lower San Andres was used as the
main indicator of how foam behaves at each concentration. As surfactant concentration reduced from the
18 SPE-200327-MS

current level to zero, gas flow to the Upper San Andres reduced from 45% to 27%. This change in the
flow distribution was predicted to occur in only a few weeks after altering the concentration, showing that
according to the model the foam stability is very sensitive to the surfactant concentration. This is mainly
due to relatively high volumetric flow of gas going into the OOZ layers resulting in a fast drop in surfactant
concentration in these layers. Since foam strength is a direct function of surfactant concentration, any change

Downloaded from http://onepetro.org/SPEIOR/proceedings-pdf/20IOR/1-20IOR/D011S004R001/2367695/spe-200327-ms.pdf/1 by Khalifa University of Science and Technology user on 15 October 2023
in injection concentration has a significant effect on the foam performance.

Table 5—Effect of Surfactant Concentration on Gas Flow Distribution

The model was also utilized to forecast the effect of long-term interruption in surfactant injection
(injection cease) on production. This effect was more pronounced on producer 3236-009 located on the
west side of the pattern, which was due to the strong connectivity between the pattern injector and this
producer (per production analysis). In Figure 14, the model forecast was initiated from the beginning of year
2020 by setting all the producer wells (including 3236-009) on constant bottom-hole-pressure constraints.
Two injection schemes were tested, 1) surfactant injection was continued at the current concentration (solid
line), and 2) surfactant injection was ceased in mid-2020 by keeping the rest of the variables (such as WAG
ratio) the same (dash lines). The model results showed about 15-20% jump in GOR right after surfactant
injection is interrupted. This quick response was mainly due to gas finding its way to this producer via
OOZ layers (thief zone) after surfactant injection ceases. In the field, such an increase in gas rate will result
in reduction in pumping efficiency at the producer and eventually lower the total liquid production. The
difference between the GOR curves in the figure gets more pronounced long-term, reaching ~35% in few
years due to a gradual reduction in oil production rate after the surfactant injection ceases. As shown in
Table, ceasing of surfactant injection results in a significant change in gas distribution between the Upper
and Lower SA, which adversely affects oil sweep efficiency from the Upper SA. In addition, a gradually
increasing GOR in simulation prediction suggests the need for dynamic optimization of WAG injection
strategy if it is realized in the field.
SPE-200327-MS 19

Downloaded from http://onepetro.org/SPEIOR/proceedings-pdf/20IOR/1-20IOR/D011S004R001/2367695/spe-200327-ms.pdf/1 by Khalifa University of Science and Technology user on 15 October 2023
Figure 14—Model forecast for producer 3236-009 showing effect of surfactant cease in mid-2020

Conclusions
ConocoPhillips and Dow successfully implemented a surfactant enabled CO2 foam injection pilot in the
EVGSAU field utilizing an improved foaming surfactant chemistry. A novel injection strategy was executed
that made the pilot both a technical and economic success and proved that CO2 foam injection is a viable
solution for correcting conformance issues. The key conclusions are summarized below:
1. The scCO2-soluble surfactant used in this pilot can provide desired mobility control with low
surfactant concentrations and minimum adsorption to reservoir rock, which results in faster foam
propagation and conformance correction.
2. Complete elimination of OOZ injection and redistribution of CO2 into previously un-swept flow
intervals.
3. Elimination of direct gas breakthrough resulted in increased runtime in high GOR wells.
4. Improved volumetric sweep efficiency resulted in significant pattern uplift; uplift peaked at 60%
above baseline and has averaged ~20% above baseline throughout the entire pilot.
5. Pilot uplift resulted in incremental oil recovery of 31 MBO compared to the baseline forecast.
6. The simulation forecast suggests that surfactant injection should be continued to maintain uniform
flow distribution between the Upper and Lower San Andres. However, the project economics will
ultimately determine how long the surfactant will continue to be injected in the pattern.

Acknowledgements
The authors acknowledge the lab technicians, engineers, and leadership team at The Dow Chemical
Company and ConocoPhillips for their support during this project.

References
Abbas, Sayeed, Sanders, Aaron, Falcone, Susan et al. 2012. Experimental Evaluation of ELEVATETm technology for the
Anadarko field trial: Foam sensitivity, studies - part 1, CRI # 2012013749. Report No. 978-1-61399-213-5, (2012/1/1/).
Abbaszadeh, Maghsood, Kazemi Nia Korrani, Aboulghasem, Lopez-Salinas, Jose Luis et al. 2014. Experimentally-
Based Empirical Foam Modeling, SPE-169888-MS. Presented at the SPE Improved Oil Recovery Symposium, Tulsa,
Oklahoma, USA. 2014/4/12/.
20 SPE-200327-MS

Al Ayesh, A. H., Salazar, R., Farajzadeh, R. et al. 2016. Foam Diversion in Heterogeneous Reservoirs: Effect of
Permeability and Injection Method. Presented at the SPE Improved Oil Recovery Conference, Tulsa, Oklahoma, USA.
2016/4/11/.
Blaker, Tore, Aarra, Morten G., Skauge, Arne et al. 2002. Foam for Gas Mobility Control in the Snorre Field: The FAWAG
Project, SPE-78824-PA. SPE Reservoir Evaluation & Engineering 5 (04): 317–323.
Brownlee, M. H. and Sugg, L. A. SPE-16721-MS. East Vacuum Grayburg-San Andres Unit CO2 Injection Project:

Downloaded from http://onepetro.org/SPEIOR/proceedings-pdf/20IOR/1-20IOR/D011S004R001/2367695/spe-200327-ms.pdf/1 by Khalifa University of Science and Technology user on 15 October 2023
Development and Results to Date, SPE-16721-MS. Presented at the SPE Annual Technical Conference and Exhibition,
Dallas, TX. 1987/1/1/.
Chou, S. I., Vasicek, S. L., Pisio, D. L. et al. 1992. CO2 Foam Field Trial at North Ward-Estes. Presented at the SPE
Annual Technical Conference and Exhibition, Washington, D.C. 1992/1/1/. 10.2118/24643-MS.
Harpole, K. J. and Hallenbeck, L. D. 1996. East Vacuum Grayburg San Andres Unit CO2 Flood Ten Year Performance
Review: Evolution of a Reservoir Management Strategy and Results of WAG Optimization. Presented at the SPE
Annual Technical Conference and Exhibition, Denver, Colorado. 1996/1/1/. 10.2118/36710-MS.
Harpole, K. J., Siemers, W. T., and Gerard, M. G. 1994a. CO2 Foam Field Verification Pilot Test at EVGSAU: Phase
IIIC--Reservoir Characterization and Response to Foam Injection. Presented at the SPE/DOE Improved Oil Recovery
Symposium, Tulsa, Oklahoma. 1994/1/1/. 10.2118/27798-MS.
Harpole, K. J., Siemers, W. T., and Gerard, M. G. 1994b. CO2 Foam Field Verification Pilot Test at EVGSAU: Phase
IIIC-Reservoir Characterization and Response to Foam Injection. Presented at the SPE/DOE Improved Oil Recovery
Symposium, Tulsa, Oklahoma. 1994/1/1/.
Heller, J. P., Boone, D. A., and Watts, R. J. 1985. Testing CO2-Foam for Mobility Control at Rock Creek. Presented at
the SPE Eastern Regional Meeting, Morgantown, West Virginia. 1985/1/1/. 10.2118/14519-MS.
Heller, John P.. 1994. CO2 Foams in Enhanced Oil Recovery. In Foams: Fundamentals and Applications in the Petroleum
Industry, ed. Laurier L. Schramm, Vol. 242, Chap. 5, 201–234. Washington, DC: Advances in Chemistry, American
Chemical Society.
Henry, Richard L., Fisher, D. Ramsey, Pennell, Stephen P. et al. 1996. Field Test of Foam to Reduce CO2 Cycling.
Presented at the SPE/DOE Improved Oil Recovery Symposium, Tulsa, Oklahoma. 1996/1/1/. 10.2118/35402-MS.
Hirasaki, G. J. and Lawson, J. B. 1985. Mechanisms of Foam Flow in Porous Media: Apparent Viscosity in Smooth
Capillaries, SPE-12129-PA.
Hoefner, M. L. and Evans, E. M. 1995. CO2 Foam: Results From Four Developmental Field Trials. SPE Reservoir
Engineering 10 (04): 273–281. 10.2118/27787-PA.
Holm, L. W. and Garrison, William H. 1988. CO2 Diversion With Foam in an Immiscible CO2 Field Project. SPE
Reservoir Engineering 3 (01): 112–118. 10.2118/14963-PA.
Jonas, T. M., Chou, S. I., and Vasicek, S. L. 1990. Evaluation of a CO2 Foam Field Trial: Rangely Weber Sand Unit.
Presented at the SPE Annual Technical Conference and Exhibition, New Orleans, Louisiana.
King, Robert E. 1972. Stratigraphic Oil and Gas FieldsClassification, Exploration Methods, and Case Histories: American
Association of Petroleum Geologists.
Ma, Kun, Lopez-Salinas, Jose L., Puerto, Maura C. et al. 2013. Estimation of Parameters for the Simulation of Foam Flow
through Porous Media. Part 1: The Dry-Out Effect. Energy & Fuels 27 (5): 2363–2375. http://dx.doi.org/10.1021/
ef302036s.
Martin, F. D., Stevens, J. E., and Harpole, K. J. 1995. CO2-Foam Field Test at the East Vacuum Grayburg/San Andres
Unit. SPE Reservoir Engineering 10 (04): 266–272. 10.2118/27786-PA.
Patil, P. D., Knight, T., Katiyar, A. et al. 2018. CO2 Foam Field Pilot Test in Sandstone Reservoir: Complete Analysis of
Foam Pilot Response. Presented at the SPE Improved OilRecoveryConference, Tulsa, Oklahoma, USA. 2018/4/14/.
10.2118/190312-MS.
Ren, Guangwei, Zhang, Hang, and Nguyen, Quoc. 2013. Effect of Surfactant Partitioning on Mobility Control During
Carbon-Dioxide Flooding. SPE Journal 18 (04): 752–765. 10.2118/145102-PA.
Stephenson, Derril J., Graham, Andrew G., and Luhning, Richard W. 1993. Mobility Control Experience in the Joffre
Viking Miscible CO2 Flood. SPE Reservoir Engineering 8 (03): 183–188. 10.2118/23598-PA.
Tsau, Jyun-Syung and Heller, J. P. 1994. CO2 Foam Field Verification Pilot Test at EVGSAU: Phase IIIA-Surfactant
Performance Characterization and Quality Assurance. Presented at the SPE/DOE Improved Oil Recovery Symposium,
Tulsa, Oklahoma. 1994/1/1/. 10.2118/27785-MS.

You might also like