Vous êtes sur la page 1sur 13

SPE-169705-MS

Colloidal Dispersion Gels (CDG): Field Projects Review


Manrique, E., Reyes, S., Romero, J., Aye, N., Kiani, M., North, W., Thomas, C., Kazempour, M., Izadi, M.,
Roostapour, A., Muniz, G., Cabrera, F., Lantz, M., Norman, C., TIORCO LLC
Copyright 2014, Society of Petroleum Engineers
This paper was prepared for presentation at the SPE EOR Conference at Oil and Gas West Asia held in Muscat, Oman, 31 March2 April 2014.
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 th e 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 prohi bited. 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
Colloidal Dispersion Gels (CDGs) have been successfully tested in Argentina, China, USA, and recently in Colombia.
However, questions remain whether CDGs can be injected in large volumes and propagate deep into the formation without
reducing injectivity and also improve sweep efficiency.
This paper summarizes 31 implemented and ongoing CDG projects in Argentina, Colombia and the U.S. since 2005.
Project summary review includes main reservoir properties, operating conditions, pore volume of chemical injected, general
project performance, and especially, a detailed analysis of injection logs addressing the injectivity of CDG. Additionally, a
general approach for history matching CDG floods is described.
CDG injection volumes in projects reviewed vary from a few thousand barrels to hundreds of thousands of barrels.
Projects evaluated did not show injectivity reduction even after more than 600,000 barrels injected in one well. Polymer
concentration and polymer:crosslinker ratios ranged from 250 to 1,200 ppm and 20:1 to 80:1, respectively. Aluminum citrate
is the most common crosslinker used in field projects. However, chromium triacetate has been used in high salinity and
hardness conditions. Key variables to sustain the injection of large volumes of CDG below maximum operating pressure are
polymer:crosslinker ratios, polymer concentration, and injection rates to a lesser extent. CDG projects have evolved from
small to large treatment volumes showing a positive impact on oil recoveries. Despite large volumes of CDG injected
production of polymer in offset producers has rarely been detected. Wellhead pressure response, CDG viscosity, and
adsorption/retention (RRF) represents the most important variables needed to match CDG floods.
This study provides the status of the technology and field evidence that CDGs can be injected in large volumes and can
propagate into the reservoir without injectivity constraints. This review will also provide guidance to successfully design and
evaluate CDG pilot projects. Lessons learned from operating and modeling CDG projects will also be presented.
Introduction
The concept of reservoir permeability correction using sequential injection of polymers and multivalent metal ions
solutions (i.e., Aluminum Citrate - AlCit) was originally patented by Gall (1973). The Residual Resistance Factor (RRF)
generated by this method was claimed to be greater than the RRF produced by the injection of polymer only. In 1980 a
commercial scale application of the technology was reported in the North Burbank Unit (NBU), Oklahoma (Moffitt et al.
1993). Applications at NBU evaluated different strategies including the sequential injection of polyacrylamide solutions and
crosslinking solutions (AlCit or Chromium Propionate). Incremental cost per barrel reported in these projects ranged from
$US 12 to $US 14. Lower cost per incremental barrel was reported using chromium triacetate as a crosslinker. However, the
reduction in costs was mainly attributed to the use of produced brine rather than fresh water as injection fluid.
In the mid-1980s, the co-injection of polymers and crosslinkers was introduced to the industry and was referred as indepth colloidal dispersion gels (CDGs). The co-injection of the polymer and crosslinker solutions simplified operations and
allowed the injection of larger volumes of chemicals. Mack and Smith (1994) reported a summary of 29 CDG projects (19
successful, 3 marginally economic and 7 unsuccessful) in the Rocky Mountain area. CDG technology generated attention in

SPE 169705

China (Zhidong et al. 2011), other regions in the U.S. (Manrique and Lantz, 2011) and more recently in Argentina (Diaz et al.
2008; Muruaga et al. 2008; Menconi et al. 2013) and Colombia (Maya and Castro 2013; Castro et al. 2013). However,
despite numerous successful field results reported in the literature, laboratory-scale experiments (Al-Assi et al. 2006;
Ranganathan et al. 1998; Seright 1994 and 2013) have generated controversy regarding the ability to inject CDGs in large
volumes without reducing injectivity while also improving sweep efficiency.
Spildo et al. (2009 and 2010) demonstrated that CDG aged a few days could propagate through Berea cores and increase
oil recovery at irreducible oil saturation to water. Castro et al. (2013) recently validated that CDG (Aluminum Citrate and
HPAM) freshly made or aged for one week can propagate in Berea core plugs. Additional results of this study showed that
incremental oil recoveries were higher and differential pressures were lower when injecting freshly-made CDG compared to
one-week-aged CDG. Figure 1 depicts polymer and aluminum concentration recovery (at irreducible oil saturation to water or
Sorw) in the Berea core floods reported by Castro et al. (2013). Injected CDG was prepared in synthetic formation brine (1%
total salinity) using 600 ppm of polymer (HPAM) and a polymer:crosslinker ratio of 20:1 as reported by Spildo et al. (2010).
Elution of polymer and aluminum in the experiment injecting freshly-made CDG (Fig. 1a) shows clear differences with oneweek-aged CDG (Fig. 1b). The data show that polymer and aluminum are produced at a constant ratio after injecting
approximately 3.5 PV of freshly made CDG at different injection rates (Fig 1a). However, for the one-week-old CDG,
polymer and aluminum are produced at a constant ratio after approximately 1 PV injected (Fig. 1b). This result was expected
due to the injection of pre-formed CDG (polymer and aluminum already cross-linked before its injection). These results
validate that CDG can propagate in the porous media as reported by Spildo et al. (2010) and provides additional evidence to
support the feasibility of injecting large volumes of CDG without well injectivity constraints.

Figure 1. Concentration of polymer and aluminum as a function of produced PV during oil recovery corefloods (@ Sorw) using
CDG freshly made (a) and aged for one week (b)

Despite the existing debate regarding the effectiveness of applying CDG or weak gels for conformance or mobility
control purposes, the number of laboratory and fields studies continues to increase demonstrating the interest of this topic due
to its significance for increasing oil recovery in waterflood projects. Mumallah (1987) suggested that to achieve in-depth
permeability correction, dilute polymer solutions and crosslinker (gelling agent) should be injected in one slug or in a
sequential mode (polymer-crosslinker-polymer) allowing the weak gels to form in-situ and partially blocking the invaded
zone. Mumallah also stated that in-depth permeability correction operations are preferred over polymer flood only or nearwell gel treatments because a large portion of the reservoir can be treated at a reasonable cost and the permeability reduction
lasts for some time. The concept of in-depth profile modification to improve waterflood sweep efficiencies continues to
evolve from Deep Diverting Gels or DDG (Fletcher et al. 1992) to the concept of Thermally Active Polymers (TAP) also
known as BrightWater (Frampton et al. 2004; Pritchett et al. 2003; Salehi et al. 2012). In 2003 the application of weak gels
for in-depth profile modification and oil displacement was reported by Wang and Liu (2003). Shi et al. (2011) reported a
detailed literature review on various microgel methods including a modeling approach for CDG. Zhidong et al. (2011)
provided a comprehensive comparison of CDG and polymer flooding projects implemented in the Daqing Field in China.
More recently, a comparison between polymer flooding and in-depth profile modification using analytical and numerical
methods was presented by Seright et al. (2012).
It is important to remark that one of the main motivations of in-depth injection profile modification since its early stages
of development (Gall 1973; Frampton et al. 2004) was the potential limitation of severe injectivity reduction using polymer
flooding or conventional polymer gels (i.e., MARCITsm gels). Injectivity losses are not only associated to the potential risk of
losing oil production (poor voidage replacement) but also can be limited by injection capabilities of existing surface facilities.
Therefore, the injection of viscous polymer solutions could be limited in some scenarios. To mitigate injectivity losses the

SPE 169705

injection of polymer solutions above fracture gradients has been proposed and implemented in the field. However, injection
above parting pressure is not necessarily the best operating strategy for chemical floods and can lead to early chemical
breakthrough increasing operating expenditures (OPEX) due to increased crude oil-chemical separation and water treatment
for re-use or disposal. Additionally, high polymer production can also lead to productivity losses not commonly documented
in the literature (Singhal 2011; Choudhuri, et al. 2013) and that will impact project economics.
For the above reasons, CDG has been considered as a feasible technology for in-depth conformance and as a mobility
control strategy to improve oil recovery and reduce water production in waterflood projects. In addition to project economics,
some of the variables frequently considered for evaluation of CDG technology over straight polymer flooding or near
wellbore treatments with conventional polymer gels normally include one or more of the following conditions and possibly
others, depending on the reservoir:

Maturity of the waterflood (Evaluate evidence for presence of remaining movable oil)
Waterfloods operating under adverse mobility ratios
Low reservoir permeability
Thin reservoirs (Net pay thickness < 40 ft) injecting water with vertical wells
Potential injectivity constraints due to narrow margin between maximum injection and reservoir pressures (Assumes
injection below parting pressure)
Limited water handling capabilities
Requirement to minimize or delay polymer production

This paper will summarize 31 implemented and ongoing CDG projects in Argentina, Colombia, and the U.S. since 2005.
The project summary review includes main reservoir properties, operating conditions, pore volume of chemical injected and
general project performance. We focus on the analysis and possible interpretation of injection logs (i.e., plots of injection
parameters) as an attempt to provide evidence regarding the injectivity and propagation of CDG. Injection log interpretation
will also include the use of Hall plots comparing CDG vs. polymer injection and different conformance technologies (TAP,
CDG, and MARCITsm gels). A special case including wells with CDG re-treatments will be also presented. Finally, a general
approach for the prediction and history matching of CDG floods is also described.
CDG Project Reviews
Table 1 shows a summary of the main reservoir and operating conditions of the CDG projects reviewed. Projects include
the use of CDG technology for in-depth conformance, mobility control or both. A third group considered the injection of
polymer gels to reduce water channeling followed by CDG as a mobility control (Menconi, et al. 2013; Muruaga, et al.
2008).
Table 1Summary of reservoir and operating conditions of CDG projects reviewed
Basic Reservoir Properties
Temperature (F)
Permeability (mD)
Average Net Pay (ft)
Oil Viscosity (cP)
Pressure at Start (psi)
Basic Operating Parameters
Polymer Conc. (ppm)
Crosslinker
Polymer:Crosslinker Ratio
Injection rates (bbl/d)
Maximum Operating Pressure (psi)
Volume Injected (bbls/well)

Range
80 - 210
10 - 4,200
20 - 200
5 - 30
0 - 1,400
250 - 1,200
Aluminum Citrate (23 of 31)
Chromium Acetate (8 of 31)
20:1 to 80:1
150 - 2,000
750 - 2,200
10,000 - > 650,000

Use of CDGs as an in-depth conformance or mobility control strategy depends on multiple variables. Adverse mobility,
narrow margin of maximum injection and reservoir pressures, and high water cuts were common factors in the projects
reviewed. Additionally, some of the projects included evaluation of the injection of bulk polymer gels (MARCITsm gels)
before starting injection of CDGs. However, in some of the projects, bulk polymer gels generated marginal results due to
low injectivity (wells pressured-up too fast) leading to small volume treatments that generated results below expectations.
Fig. 2 depicts injection logs of the 31 CDG projects evaluated. CDG projects reviewed are mainly for in-depth
conformance and an ongoing mobility control project represented by the only well reporting more than 650,000 bbl injected
(Castro et al., 2013). None of the projects showed prolonged continuous increase in injection pressures.

SPE 169705

Fig. 2CDG injection logs (wellhead injection pressure vs. Cum. CDG injected) of 31 projects reviewed

Once the technical and economic feasibility of CDG is validated a pilot injection scheme is proposed including
contingency plans to address possible unexpected situations at injectors (e.g., sharp increases in pressure) or producers (e.g.,
polymer production). Before project startup, injection profiles (i.e., ILT) are usually available to indicate the location and the
water intake of different perforated intervals. To promote CDG flow through high flow-capacity zones, injection rates are
reduced if required. This will also reduce the risks of injecting higher viscosity fluids above frac gradient. If the reduction of
injection rate impacts oil production rates in offset producers, then this is also included in the evaluation to minimize
production losses during the treatment that can impact project economics. Defining maximum operating conditions also
represents a key operating variable before CDG injection is started. Maximum operating pressure could be limited by surface
facilities or formation parting pressure and will contribute to modifying the treatment by adjusting key operating variables
such as polymer concentration, polymer:crosslinker ratios, and/or injection rates.
For the purpose of this analysis CDG conformance projects were divided into two groups, (a) wells starting with positive
well head pressure, and (b) wells starting under vacuum. Typically, for the treatments starting with positive injection
pressure, pressure buildup response was observed after a few hundred to 20,000 bbls of CDG injected. Once injection
pressure is approaching maximum operating conditions, polymer concentration and/or polymer:crosslinker ratios and
injection rates (to a lesser extent), are typically adjusted accordingly to continue injecting close to maximum operating
conditions.
To further address the injectivity of CDG, a project implemented in a thin (20 to 25 ft thick) and relatively low
permeability (1.8 to 600 md) reservoir in the U.S. was selected for a more detailed discussion. CDG was considered in this
field due to fast tracer breakthrough (5 to 30 days) in four injectors with 5-acre well spacing in irregular patterns. To
minimize the risk of polymer production, all injectors were treated with approximately 1,000 bbls of MARCITsm gels before
CDG injection started. Average permeability of this reservoir was estimated to be 300 md. The oil viscosity was 8 cp at
reservoir conditions. After completing the MARCITsm gel treatment, CDG injection started with a polymer concentration of
600 ppm and a polymer:crosslinker ratio of 30:1. At project startup the reservoir pressure was 250 psi and the maximum
operating condition was defined as 750 psi. After approximately 10,000 bbls of CDG had been injected in each well, all
injectors reached an injection pressure of 600 psi. At this point polymer concentration was decreased to 450 ppm keeping the
same polymer:crosslinker ratio (Fig. 3). Pressure continued to increase and polymer concentration was reduced to 300 ppm
until the end of the treatment without modifying polymer:crosslinker ratio or injection rates (200 to 220 bbl/d/well). The
volume of CDG injected ranged from 16,000 to 18,000 bbls per well representing approximately 4% of the pilot pore
volume.
Fig. 3 shows the injection logs for the CDG injection in four wells after the injection of 1,000 bbls/each of MARCITsm
gels. Results clearly suggest that CDG could be injected without face plugging as reported in some laboratory studies (AlAssi et al. 2006; Ranganathan et al. 1998; Seright 1994). This example will also be addressed in the CDG simulation section

SPE 169705

explaining the use of the gel option to history match pressure buildup observed in the field. It can be noticed that pressure
buildup observed in all four wells shows a similar pattern and a reasonable fit can be obtained with a conventional
polynomial equation of 2nd order (Fig. 3a). To filter noise in the data gathered for each of the injectors, the data sets were
smoothed using Loess (quadratic fit) method prior to the curve fitting process. It was found that different forms of Power
functions fit the data quite well (Fig. 3b). Interestingly, similar trends were found in more than 80% of the field cases
analyzed providing useful information regarding possible gel formation and its propagation in the porous media.

Fig. 3CDG injection logs implemented in a thin and low permeable reservoir in the U.S.

Fig. 4 shows well injectivity for all CDG projects evaluated in this study. Most wells treated with CDG did not show
injectivity reduction except for those treatments starting after the injection of MARCITsm polymer gels (highlighted in Fig.
4), which were applied due to early tracer breakthrough and/or well injecting water under vacuum. One good example of
combining MARCITsm gels followed by CDG injection has been reported by Muruaga et al. (2008) and Menconi et al.
(2013). Combination of MARCITsm gel injection is not only limited to CDG; Saez et al. (2012) and Paponi et al. (2013) have
recently reported on projects combining MARCITsm gels followed with polymer injection to solve severe water channeling in
a viscous oil reservoir in Argentina.

Fig. 4Well injectivity for CDG projects evaluated

The Hall plot represents another tool to monitor project performance of conformance treatment and EOR floods. The Hall
plot was originally proposed to evaluate the performance of waterfloods and estimate skin effects in water injection wells
(Hall 1963). Buell et al. (1990) proposed a method to use Hall plots for both water and polymer floods (Non-Newtonian
fluids). Honarpour and Tomutsa (1990) also proposed the use of Hall plots for monitoring and reservoir characterization
purposes in Bell Creek water and micellar-polymer flood. Hall plots of in-depth conformance technologies (i.e., CDGs and

SPE 169705

TAPs) generally show a gradual increase of positive skin compared to near wellbore treatments such as MARCITsm gels,
which show a sharper increase of positive skin. CDG treatments can also show a sharp increase in Hall plots when injected at
high polymer concentrations with polymer:crosslinker ratios between 20:1 to 40:1. However, this behavior should be
considered well specific and cannot be generalized as a rule.
In an attempt to compare Hall plot differences between conformance technologies, projects were selected that had similar
injection rates ( 1,000 bbl/d) injecting (1) MARCITsm gels, (2) CDG, and (3) TAP technologies (Fig. 5). The Hall plot
comparisons in Fig 5 are for the first few thousands of barrels injected for each treatment. MARCITsm gel was injected at a
concentration of 4,500 ppm (Black dashed line) and all TAP (BrightWater) projects were pumped using a concentration of
5,000 ppm. Hall plots for the CDG projects shown in Fig. 5 fell in between TAP and MARCITsm gels. CDG treatment 1
(Green dashed line) was injected using a polymer concentration of 600 ppm and polymer:crosslinker ratio of 20:1. CDG
treatment 2 (Red dashed line) was implemented using a polymer concentration of 300 ppm and a polymer:crosslinker ratio of
40:1. Although comparing Hall plots of different projects and technologies in various fields is challenging, this example
presents a general idea of the injection performance of different conformance technologies. As expected, TAP Hall plots do
not show major changes during TAP injection due to the low initial viscosity at these concentrations for this polymer system
and lower temperatures usually observed in water injection wells. Once TAP is activated the Hall plot will show a gradual
increase of positive skin (Choudhary et al. 2014). CDG projects considered for comparison purposes (Fig. 5) show a rapid
positive skin due to the viscosity of the systems injected as a conformance strategy and low reservoir permeabilities.

Fig. 5Example of Hall plots for two CDG projects evaluated compared with different conformance technologies

El Tordillo Field (Argentina) represents one of the fields with the largest number of CDG treatments (11 wells). Reported
incremental oil recoveries are between 3 and 3.5% of the OOIP (Menconi et al. 2013). However, CDG injection was
combined with MARCITsm gel treatments to control severe water channeling through high permeability channels present in
the reservoir. Again, the combination of MARCITsm gels before CDG in these treatments did not generate injectivity
constraints or face plugging as reported in laboratory studies (Al-Assi et al. 2006; Ranganathan et al. 1998; Seright 1994).
The operator is continuing to use the technology and improving its implementation based on lessons learned and detailed
reservoir management strategies to develop the field.
Daqing field in China (Chang et al. 2004; Zhidong et al. 2011) and more recently Dina Cretceos field in Colombia
(Castro et al. 2013) are the projects with the largest CDG injected volumes per well documented in the literature. Daqing
field experiences have been widely documented in the literature but field information and detailed performance are not
available. However, recent comparison of CDG vs. polymer flooding was summarized by Zhidong et al. (2011). In the case
of Dina Cretceos field, the pilot started in 2011. Over 650,000 bbls of CDG had been injected in the first pilot injector
without major injectivity problems. Based on CDG oil production response and water cut reduction the pilot has been
expanded by increasing the number of injectors as reported by Castro et al. (2013).
Considering the fact that there are not many fields (other than Daqing) reporting CDG and polymer flooding in the same
reservoir, this paper compared a few ongoing polymer and CDG floods implemented in low permeability reservoirs (5 to 300
md) where the data is available. Hall plots were again used to demonstrate the injectivity of large volumes of polymer and
CDG in reservoirs with reasonable similarities (Fig. 6). The polymer projects are injecting above the frac gradient (1,500 to

SPE 169705

3,000 bbl/d) using polymer concentrations between 500 and 800 ppm while CDG projects are injecting below the frac
gradient (1,000 to 2,000 bbl/d) with polymer concentration between 300 and 600 ppm and variable polymer:crosslinker ratios
(40:1 to 80:1). Hall plots of polymer injection (dashed lines in Fig. 6) clearly show the effects (periods of skin decrease)
generated by injection rate changes to keep injecting above frac gradient in a low permeability reservoir. The Hall plot of
CDG injection (solid lines in Fig. 6) are comparable to polymer floods considering that injection rates are below the fracture
gradient and the main operating variable to keep injecting high volumes below or close to the maximum operating conditions
is the variation of polymer:crosslinker ratios and polymer concentration or injection rates to a lesser extent.

Fig 6Example of Hall plots comparing CDG vs. Polymer injection in reservoirs with similar characteristics

Polymer and CDG projects shown in Fig. 6 are performing above expectations using different operating strategies. As
stated earlier, comparing Hall plots of polymer based technologies in different fields is challenging but it is clear that CDG
can be injected in large volumes without reducing injectivity and propagating through the reservoir despite some laboratory
studies concluding the opposite. The Hall plot presented in Fig. 6 cannot explain the kinetics of formation or flow
characteristics of CDG in the reservoir. However, the CDG projects are showing incremental oil, decrease in water
production without polymer production at a cost below $US 5 per incremental barrel. Therefore, there is a discrepancy
between laboratory and field evidence that needs to be re-addressed because evidence from field projects show low cost per
incremental barrel, decrease in water production and no evidences of polymer production or injectivity/productivity losses.
The following sections will continue presenting additional evidence that CDG can be formed and propagated in the reservoir
without causing well plugging.
Special Case: Wells with CDG Re-treatments
Well retreatment for in-depth profile modification has been proposed as a possible strategy to improve sweep efficiency
in water injection projects (Choudhary et al. 2014). Fundamentally, the objective of a second (or additional) treatments (i.e.,
re-treatments) is to prevent or delay water channeling of secondary thief zones to continue improving volumetric sweep
efficiency and hence increasing oil recovery. The first treatment will partially block higher flow-capacity intervals and
subsequent treatments may target less permeable (lower flow capacity) intervals and/or expand transmissibility reduction
generated by the first treatment. This concept was recently evaluated using a numerical simulation approach by Seright et al.
(2012). Well re-treatments with CDG have been also reported by Diaz et al. (2008) and Castro et al. (2013).
Fig. 7 depicts two wells reporting CDG re-treatments. Subsequent CDG slugs were injected after several months of water
injection at the same injection rates. The water injection data are not shown in Fig. 7. From the injection logs it can be
concluded that the re-treatments with CDG did not show injectivity constraints, which supports the conclusion that face
plugging of the injectors was not occurring. CDG Re-Treatment 1 (Fig. 7) consisted of the co-injection of a polymer solution
of 400 ppm and variable polymer:crosslinker ratios using AlCit as a crosslinker. CDG Re-Treatment 2 (Fig. 7) co-injected
polymer and Chromium Triacetate as a crosslinker. CDG Re-Treatment 2 case considered different polymer concentrations
(300 to 450 ppm) and polymer:crosslinker ratio (20:1 to 40:1). An interesting observation of both projects is that oil response
was observed during the first treatment validating the possibility of CDG displacing viscous oils as reported by Diaz et al.

SPE 169705

(2008) and Castro et al. (2013). CDG Re-Treatment 2 (Case 2) was selected to provide additional evidences that CDG can be
formed and propagate in the reservoir.

Fig. 7Examples of injection logs for wells reporting retreatment with CDG

Comparing both CDG injection treatments (at time zero) of Case 2 it can be observed that build-up pressures are very
similar (Fig. 8). It is important to observe that each CDG phase (Phase I and Phase II) injected approximately 190,000 bbls
(Fig. 7). CDG injected during Phase I of the project involved a variable polymer concentration (300 to 450 ppm) at a constant
polymer:crosslinker ratio of 20:1. The second CDG treatment (Phase II) was implemented after 13 months of water injection
at an approximately constant injection rate of 1,076 bbl/d. Injection strategy for Phase II of the project considered a different
strategy injecting at a constant polymer concentration and polymer:crosslinker ratio of 300 ppm and 40:1, respectively.
Similarities of both CDG treatment injection logs (Fig. 8a) suggests that the first treatment (Phase I) did not change well
injectivity and the second treatment (Phase II) is flowing through the same high permeable interval (i.e., the first treatment
might be too small to generate a stronger water diversion) and/or CDG formation (chemical reaction) occurs at a given rate in
the reservoir, among other possibilities. As was observed in the four field cases described at the beginning of this section
(Fig. 3a), pressure build-up during the first 20,000 bbls of CDG injected is very similar and can be fitted with a similar
polynomial of 2nd order (Fig. 8b). This pattern (pressure buildup type curve) has been observed in more than 80% of the field
cases evaluated and may provide important information regarding the CDG (microgel) formation and characteristics of its
propagation in the reservoir. However, additional interpretations are required to infer possible gel formation supported by
numerical simulation studies.

Fig. 8Comparison of injection logs of CDG re-treatments of Case 2 shown in Fig. 7 for the total treatment of each phase (a) and
pressure build-up observed during the first 20,000 bbls of CDG injected (b)

SPE 169705

Based on laboratory studies, CDGs cannot be formed in the reservoir or should plug injectors because it cannot
propagate in the reservoir. CDG re-treatments presented in this section do not support a conclusion that face plugging of the
injectors occurred. However, these field cases do not necessarily validate the formation of CDG in the reservoir. The Hall
plot was used to continue evaluating CDG performance of Case 2 (Fig. 9). The Hall plot includes the injection history since
the injector started with water injection, which identified some well events (changes in injection rates) and the periods when
both CDG treatments were implemented.

Fig. 9Hall Plot of CDG retreatments of case 2 shown in Figs. 7 and 8

The first CDG treatment of 186,200 bbls was injected at 1,025 bbd/d. During the CDG injection, an increase in oil
production and a decrease in water production were observed. The operator ran an injection profile (ILT) before and after the
first CDG slug. The injection profile after the first phase of CDG injection showed a clear reduction of water intake in the
main thief zone and new intervals taking injection water that did not record any injectivity prior to CDG injection (Diaz et al.,
2008). Water injection rates were kept constant at 1,076 bbl/d after Phase I of the CDG project. It can be noticed that the Hall
plot shows a slight increase in positive skin typical of in-depth conformance (Choudhary et al., 2014) compared with water
injection at a similar injection rate. Based on the trends observed in the Hall plot (Fig. 9) and changes in injection profiles
observed right after the first treatment, it can be concluded that CDG formed and was displaced away from the injector
having a small influence near the injector wellbore. Additionally, a slight and continued increase in positive skin (Trend I in
Fig. 9) and no polymer production in offset producers also suggest that in-depth permeability reduction was still in place
before phase II of the CDG project started.
After approximately 13 months injecting water, Phase II of the project started with a slightly lower injection rate (1,002
bbl/d). The second CDG slug (192,729 bbls) was injected without any injectivity constraint, which can be confirmed with the
overlap of pressure build-up response of both CDG treatments described in Fig. 8. However, five months after the second
phase of CDG injection was completed, injection rates were decreased (700 bbl/d) due to the increase of wellhead injection
pressures caused by water diversion to lower permeability intervals and continued water injection below the parting pressure.
Injectivity reduction should be expected due to water diversion into low flow-capacity (unswept) zones. A sharper increase in
positive skin (Trend II in Fig. 9) and no polymer production in offset producers suggest that in-depth permeability reduction
occurred. It can be argued that a decrease in water injection rates may have a negative impact on oil production rates (voidage
ratio) of the pattern treated twice with CDG. However, project economics needs to take into account the benefits of CDG for
extending oil production life of waterfloods by reducing water re-cycling and contributing to an increase in final recovery
factors.
Pressure buildup and Hall plot data strongly suggest that co-injected polymer and crosslinker do react in the reservoir.
Case 2 represents a good example of CDG formation and its propagation in a low permeability reservoir (20 to 1,000 md)
with viscous oil (30 cp) at reservoir temperature of 113F (Diaz et al. 2008). Injected polymer solution at low concentrations
(300 to 450 ppm) in poor water quality cannot justify the build-up pressures recorded or increase in positive skin observed
during the project (Figs. 8 and 9). Additionally, tracer breakthrough in two of the six producers was reported between 50 and

10

SPE 169705

110 days. However, no polymer production was reported during the duration of the project. These field case histories clearly
contradict laboratory studies postulating that CDGs cannot be formed or propagate in the reservoir (Al-Assi et al. 2006;
PRRC 2013; Ranganathan et al. 1998; Seright 1994 and 2013).
The authors understand the difficulty of demonstrating how chemical reactions forming CDG occur in the reservoir but
this difficulty is not different than other chemical EOR processes. One possible explanation is the differences in scale
between core floods and reservoir volumes making it difficult to capture possible interactions at laboratory scale. For
example, viscosity buildup observed in bottle tests is not representative for numerical simulation studies. CDG viscosities
generated in bottle tests are extremely high compared to those measured at the outlet of coreflood or sand packs. This
observation suggests that chemical interaction between the polymer and the crosslinker in the porous media differs from
bottle test observation where all the reactants can interact without constraints. The pressure build-up signature (i.e., type
curve) observed may contribute to understanding the reaction mechanisms at reservoir conditions. In the meantime, CDG
kinetics has been investigated during history matching of recent CDG floods. These observations will be briefly discussed in
the following section of the paper.
CDG Simulation
The use of numerical simulation to potentially explain the observed characteristics of in-depth conformance technologies
has been reported by Garmeh et al. (2012), Garmeh and Manrique (2011), Seright et al. (2012) and Shi et al. (2011).
Proposed treatment methods include thermally active polymers (i.e., BrightWater) or chemical (i.e., CDG, microgels) indepth conformance technologies. CDGs consists of a cross-linked low concentration polymer that develops viscosity and
resistance factor with time during flow in the reservoir. To model delayed viscosification (observed in the injection logs for
the cases reviewed) and adsorption of this system, two approaches have been evaluated: the use of multiple regions (single
component) and gelation (chemical reaction).
Not all commercial simulators include chemical reactions to form gels. Therefore, a reasonably simple simulation
approach where CDG is modeled as a single component has been considered. This approach takes advantage of defining
multiple regions with increasing CDG viscosity and RRF moving away from the injector. The intention of this method is to
capture the delayed viscosification of CDG deep in the formation, which can be inferred from pressure buildup observed at
the injectors. Although this approach is simple and has faster running times as a result of not having to define chemical
reactions in highly detailed numerical models, this method might cause some numerical difficulties due to sudden changes in
fluid properties from one region to the other. To make this transition smoother multiple regions must be defined which
represents a disadvantage of this approach. However, this method can be used as a preliminary approach for projects without
access to numerical tools including gel (chemical reaction) options.
The second modeling approach is based on a chemical reaction to form CDG. In this approach polymer and crosslinker
are injected in the reservoir with specific concentrations as it happens in field operations. Each component (polymer and
crosslinker) is injected at a given concentration and viscosity (initially a lower viscosity than CDG as determined in the
laboratory) and will react to form CDG considering the following:

CDG activation can be controlled by the reaction rate coefficient


Reaction product (CDG) will have high viscosity and resistance against flow (RRF) due to adsorption/retention
Chemical reaction rate is tuned to build the viscosity vs. time curve (delayed viscosification)
Chemical reaction rate coefficient controls the viscosification timing of CDG

While the chemical reaction approach provides a smooth transition in terms of CDG formation, reaction stoichiometry is
not well understood. In this approach polymer:crosslinker ratio has a significant effect on the simulation results and in cases
when this ratio is changed (i.e., most common variable changed in field operations) all other parameters have to be adjusted
accordingly, which makes this approach relatively complicated. To overcome this potential complication the use of a two
component system has also been tested. In this case a chemical reaction is defined to convert CDG1 to CDG2 where CDG1
has lower viscosity and RRF compared to CDG2. However, additional efforts are required to improve the prediction of CDG
during pilot design phases.
Fig. 10 depicts an example of CDG history matching using the two different simulation approaches. The best approach to
history matching CDG performance is using pressure buildup reported at the (four) injectors as described previously in this
paper (Fig. 3 and Fig. 7). Fig. 10a represents an example using the simulation approach of multiple regions. Both models
(CDG Model 1 and CDG Model 2) include three regions. The main difference between both models is the approach to the
viscosity build-up and RRF from one region to the other. This simulation approach was also compared with straight polymer
injection (Green dashed line in Fig. 10a). Polymer injection was based on the polymer concentration (400 ppm) injected

SPE 169705

11

during the CDG project in the absence of crosslinker. Results suggest that pressure build-up observed at the injector cannot
be matched using low polymer concentrations.
The second history matched project was based on a variable polymer concentration (300 to 600 ppm) at a constant
polymer:crosslinker ratio (30:1) as was described in Fig. 3. In this case chemical reaction (gelation) approach was used and
reported by Garmeh and Manrique (2011). This gel option reasonably matched (Red line in Fig. 10b) historical well head
pressure data (Blue circles in Fig.10b) recorded during the pilot test. Similar results were obtained for all four injectors.
Again, this model was used to compare CDG vs. straight polymer flooding. CDG injection started at 600 ppm and to history
match historical wellhead pressure data polymer flooding was assumed as injecting the same volume of CDG using a
constant polymer solution of 600 ppm. Again, results suggest that pressure build-up generated by CDG cannot be matched
with straight polymer injection (Green dashed line in Fig. 10b).

Fig. 10Example CDG simulation approaches using multiple regions (a) and chemical reaction (b)

To history match wellhead pressure data depicted in Fig. 10b the use of variable skin factors was also considered
(assuming face plugging effects). However, this approach could not match injection or production rates during and after the
injection of CDG. Finally, a special case of polymer flooding was run to be able to match incremental oil recoveries gained
by the CDG pilot test. In this case CDG and polymer flooding generated similar incremental oil recoveries. However,
polymer flooding required approximately 16,000 pounds of additional polymer mass compared to CDG. CDG injection used
875 pounds of crosslinker, which represents important savings that can benefit project economics of this particular project.
Therefore, it could be suggested that CDG is not superior to polymer flooding based on these simulation results. However,
project economics will provide a benefit from CDG when comparing to straight polymer flooding. Additional benefits of
CDG are that field project experience has shown that less polymer will be produced compared to polymer flooding, which
reduces treatment costs of produced fluids. In other words, CDG can generate similar recoveries at lower CAPEX and OPEX.
However, it is important to mention that CDG cannot be formed under certain reservoir conditions. Finally, the authors
recognize that a comprehensive review of laboratory protocols needs to be revisited to better explain field observations.
Discussions and Closing Remarks

Injection logs, well injectivity, and Hall plots confirm that CDGs do not significantly reduce injectivity and can
propagate in the reservoir.

Field cases reviewed showed that large volumes of CDG can be injected below maximum pressure operating
conditions (i.e., below the fracture gradient or maximum capacity of surface facilities). The main variables to
manage the injection of large volumes of CDG include polymer:crosslinker ratios, polymer concentration, and
injection rates in a lesser extent.

The pressure responses that have been observed in different wells of the same field suggest that it may be possible to
develop a type-curve response that will provide valuable information for project design and field expansion.

Simulation results indicate that polymer flooding and CDG flooding may produce similar final oil recovery but
polymer flooding will require more polymer mass.

12

SPE 169705

The Hall plot represents a good diagnostic tool for performance evaluation of conformance or mobility control
methods including CDG. It gives solid indications of permeability, skin effects, and changes in drainage area (i.e.,
flow diversion) supporting project interpretation.

Detailed research and development efforts are required and ongoing to better explain possible mechanisms of CDG
technology.

Acknowledgments
The authors would like to thank TIORCO LLC for permission to publish this work.
The authors also gratefully acknowledge the contribution of operating companies during all phases of project
implementation and monitoring. Special thanks to Delia Daz y Nicolas Saez Abadia, for valuable discussions of CDG
experience in Argentina.
References
Al-Assi, A. A., Willhite, G. P., Green, D. W., McCool, C.S. (2006). Formation and Propagation of Gel Aggregates Using Partially
Hydrolyzed Polyacrylamide and Aluminum Citrate (SPE-100049-PA). SPE Journal, Vol. 14, No. 3, Sept. 2009.pp: 450461.
Buell, R. S., Kazemi, H., Poettmann, F. H. (1990). Analyzing Injectivity of Polymer Solutions With the Hall Plot (SPE-16963-PA).
SPE Reservoir Engineering, Vol. 15, No.1, Feb. 1990. pp: 4146.
Castro, R., Maya, G., Sandoval, J., Leon, J., Zapata, J., Lobo, A. Villadiego, D., Perdomo, L., Cabrera, F., Izadi, M., Romero, J.,
Norman, C., Manrique, E. (2013). Colloidal Dispersion Gels (CDG) in Dina Cretceos Field: From Pilot Design to Field
Implementation and Performance (SPE-165273). SPE Enhanced Ol Recovery Conference, Kuala Lumpur, Malaysia, July 24,
2013.
Chang, H. L., Sui, X., Xiao, L., Guo, Z., Yao, Y., Xiao, Y., Chen, G., Song, K., Mack, J. (2004). Successful Field Pilot of In-Depth
Colloidal Dispersion Gel (CDG) Technology in Daqing Oil Field (SPE-89460). SPE/DOE Symposium on Improved Oil
Recovery, Tulsa, OK, April 1720, 2004.
Choudhary, M., Parekh, B., Solis, H., Meyer, B., Shepstone, K., DeZabala, E., Prosteby, C., Manrique, E., Izadi, M., Larsen, D.
(2014). Reservoir In-Depth Waterflood Conformance: An Offshore Pilot Implementation (SPE-169132). To be presented at the
SPE IOR Symposium to be held in Tulsa, OK, April 1216, 2014.
Choudhuri, B., Kalbani, A., Cherukapalli, P. K., Ravula, C., Hashmi, K., Jaspers, H. (2013). Enhancing Value of Polymer Flood
Project with Proactive Well and Reservoir Management (SPE-165274). SPE Enhanced Oil Recovery Conference, Kuala Lumpur,
Malaysia, July 24, 2013.
Daz, D., Somaruga, C., Norman, C., Romero, J. (2008). Colloidal Dispersion Gels Improve Oil Recovery in a Heterogeneous
Argentina Waterflood (SPE-113320). SPE/DOE Symposium on Improved Oil Recovery, Tulsa, OK, April 2023, 2008.
Fletcher, A. J. P.; Flew, S.; Forsdyke, I. N.; Morgan, J. C.; Rogers, C.; Suttles, D. (1992). Deep diverting gels for very cost-effective
waterflood control. Journal of Petroleum Science and Engineering, 7 (1992). pp. 3343.
Frampton, H.; Morgan, J. C.; Cheung, S.; Munson, L. (2004). Development of a Novel Waterflood Conformance Control System
(SPE-89391). SPE/DOE Symposium on Improved Oil Recovery, Tulsa, OK, April 1721, 2004.
Gall, F. W. (1973). Subterranean Formation Permeability Correction. U.S. Patent 3,762,476 A. Oct. 2, 1973.
Garmeh, G., Manrique E. (2011). CDG Injection in a Mature Waterflood Reservoir: History Matching Approach and Field
Performance. SPE ATW History Matching: Field Experiences and Lessons Learned, Cartagena, Colombia, August 2931, 2011.
Garmeh, R., Izadi, M. Salehi, M., Romero, J. L., Thomas, C. P., Manrique, E. J. (2012). Thermally Active Polymer to Improve Sweep
Efficiency of Water Floods: Simulation and Pilot Design Approaches (SPE-144234-PA). SPE Reservoir Evaluation &
Engineering, Vol. 15, No. 1.pp: 8697.
Honarpour, M. M., Tomutsa, L. (1990). Injection/Production Monitoring: An Effective Method for Reservoir Characterization (SPE20262). SPE/DOE Enhanced Oil Recovery Symposium, Tulsa, OK, April 2225, 1990.
Mack, J. C., Smith, J. E. (1994). In-Depth Colloidal Dispersion Gels Improve Oil Recovery Efficiency (SPE-27780). SPE/DOE
Improved Oil Recovery Symposium, Tulsa, OK, April 1720, 1994.
Manrique, E., Lantz, M. (2011). Colloidal Dispersion Gels (CDG) and Their Application in Wyoming and Elsewhere. 3 rd Wyoming
IOR/EOR Conference, Jackson Hole, WY, September 1213, 2011.
Maya, G., Castro, R. (2013). Diseo, ejecucin, monitoreo y expansin de un proceso EOR mediante al inyeccin de geles de
dispersin coloidal. Petrotecnia (Revista del Instituto Argentino de Petrleo y Gas), No. 2, Abril 2013. pp: 1228.
Menconi, F., Giaccaglia, F., Ramirez, J. Berto, C. (2013). Inyeccin de geles en el yacimiento El Tordillo: Desde los pilotos hasta la

SPE 169705

13

Masificacin. Petrotecnia (Revista del Instituto Argentino de Petrleo y Gas), No. 2, Abril 2013. pp: 4255.
Moffitt, P. D., Zornes, D. R., Moradi-Araghi, A., McGovern, J. M. (1993). Application of Freshwater and Brine Polymer Flooding in
the North Burbank Unit, Osage County, Oklahoma (SPE-20466-PA). SPE Reservoir Engineering, Vol. 8, No. 2, May 1993. pp:
128134.
Mumallah, N. A. (1987). A Practical Method for the Evaluation of Weak Gels (SPE-15142-PA). Journal of Petroleum Technology,
Vol. 39, No. 2, Feb. 1987. pp: 195202.
Muruaga, E., Flores, M., Norman, C., Romero, J. (2008). Combining Bulk Gels and Colloidal Dispersion Gels for Improved
Volumetric Sweep Efficiency in a Mature Waterflood (SPE-113334). SPE/DOE Symposium on Improved Oil Recovery, Tulsa,
OK, April 2023, 2008.
Paponi, H. L., Hyrc, A., Paponi, H., Puliti, R. L., Gerlero, T., Hochenfellner, F. (2013). Design and Execution f a Polymer Injection
Pilot in Argentina (SPE-166078). SPE Annual Technical Conference and Exhibition, New Orleans, LA, Sept. 30Oct. 2, 2013.
Petroleum Recovery Research Center PRRC (2013). Are Colloidal Dispersion Gels Really a Viable Technology? New Mexico
Tech, http://baervan.nmt.edu/research_groups/reservoir_sweep_improvement/
Pritchett, J.; Frampton, H.; Brinkman, J.; Cheung, S.; Morgan, J.; Chang, K.T.; Williams, D. and Goodgame, J. (2003). Field
Application of a New In-Depth Waterflood Conformance Improvement Tool (SPE-84897). SPE International Improved Oil
Recovery Conference in Asia Pacific, Kuala Lumpur, Malaysia, Oct. 2021, 2003.
Ranganathan, R., Lewis, R., McCool, R., Green, D. W., Willhite, G. P. (1998). Experimental Study of the Gelation Behavior of a
Polyacrylamide/Aluminum Citrate Colloidal-Dispersion Gel System (SPE-52503-PA). SPE Journal, V. 3, No. 4, Dec. 1998. pp:
337343.
Saez, M., Paponi, H. M., Cabrera, F. A., Muniz, G., Romero, J. L., Norman, C. (2012). Improving Volumetric Efficiency in an
Unconsolidated Sandstone Reservoir With Sequential Injection of polymer Gels (SPE-150492). SPE IOR Symposium, Tulsa, OK,
April 1418, 2012.
Seright, R. S. (1994). Propagation of an Aluminum Citrate-HPAM Colloidal Dispersion Gel Through Berea Sandstone. Report No.
PRRC 9429, New Mexico Institute of Mining and Technology, New Mexico, June 1994.
Seright, R. S. (2013). Polymer Flooding and Sweep Improvement. Workshop 2: Introduction to EOR in Tensleep Reservoirs. EORI,
University of Wyoming, Cody, Oct. 2223, 2013. http://www.uwyo.edu/eori/_files/eorc_tab_meeting_july_2012/sheena%20%20tensleep%20workshop%20updates%20-%20xie.pdf
Seright, R. S., Zhang,G., Akanni, O. O., Wang, D. (2012). A Comparison of Polymer Flooding With In-Depth Profile Modification
(SPE-146087). Journal of Canadian Petroleum Technology, Vol. 51, No. 5, Sept. 2012.pp: 393-402.
Shi, J., Varavei, A., Huh, C., Delshad, M., Sepehrnoori, K., Li, X. (2011). Transport Model Implementation and Simulation of
Microgel Processes for Conformance and Mobility Control Purposes. Energy & Fuels, 25, Sept. 2011. pp: 50635075.
Singhal, A. (2011). Preliminary Review of IETP Projects Using Polymers. July, 2011. pp 135.
http://www.energy.gov.ab.ca/Oil/pdfs/IETPPolymerStudy.pdf
Spildo, K., Skauge, A., Aarra, M. G., Tweheyo, M. T. (2009). A New Polymer Application for North Sea Reservoirs (SPE-113460PA). SPE Reservoir Evaluation & Engineering, V. 12, No. 3, June 2009. pp: 427432.
Spildo, K., Skauge, A., Skauge, T. (2010). Propagation of Colloidal Dispersion Gels (CDG) in Laboratory Corefloods (SPE-129927).
SPE Improved Oil Recovery Symposium, Tulsa, OK, April 2428, 2010.
Vongvuthipornchai, S., Raghavan, R. (1987). Well Test Analysis of Data Dominated by Storage and Skin: Non-Newtonian PowerLaw Fluids (SPE-14454-PA). SPE Formation Evaluation, Vol. 2, No. 4, Dec. 1987. pp: 618628.
Wang, G. Y., Liu, Y. (2003). Applications of Weak Gels for In-Depth Profile Modification and Oil Displacement. Journal of Canadian
Petroleum Technology, Vol. 42, No. 6, June 2003. pp: 5461.
Zhidong, G., Chang, H. L., Chengyong, F., Jinliang, L., Weihua, H., Dianguo, S. (2011). Comparison of Oil Displacement
Characteristics Between CDG and Polymer Flooding in the Daqing Oilfield (SPE-144119). SPE Enhanced Oil Recovery
Conference, Kuala Lumpur, Malaysia, July 1921, 2011.

Vous aimerez peut-être aussi