Spe 206010 Pa

An Extended Unified Viscoelastic Model
for Predicting Polymer Apparent Viscosity

at Different Shear Rates
Mursal Zeynalli1*, Emad Walid Al-Shalabi1, and Waleed AlAmeri1
1
Petroleum Engineering Department, Khalifa University of Science and Technology, SAN Campus
Summary
Polymer flooding is one of the most commonly used chemical enhanced oil recovery (EOR) methods. Conventionally, this technique was
Downloaded from http://onepetro.org/REE/article-pdf/26/01/99/3066351/spe-206010-pa.pdf by Khalifa University of Science and Technology user on 21 February 2023
believed to improve macroscopic sweep efficiency by sweeping only bypassed oil. Nevertheless, recently it has been found that polymers
exhibiting viscoelastic behavior in the porous medium can also improve microscopic displacement efficiency resulting in higher addition-
al oil recovery. Therefore, an accurate prediction of the complex rheological response of polymers in porous media is crucial to obtain a
proper estimation of incremental oil to polymer flooding. In this paper, a novel viscoelastic model is proposed to comprehensively analyze
the polymer rheological behavior in porous media. This proposed model was developed and validated using 30 coreflooding tests obtained
from the literature and further verified against a few existing viscoelastic models.
The proposed viscoelastic model is considered an extension of the unified apparent viscosity model provided in the literature and is
termed as extended unified viscoelastic model (E-UVM). The main advantage of the proposed model is its ability to capture the polymer
mechanical degradation at ultimate shear rates primarily observed near wellbores. Moreover, the fitting parameters used in the model
were correlated to rock and polymer properties using machine learning technique, significantly reducing the need for time-consuming
coreflooding tests for future polymer screening works. Furthermore, the E-UVM was implemented in MATLAB Reservoir Simulation
Toolbox (MRST) and verified against the original shear model existing in the simulator. It is worth mentioning that the irreversible
viscosity drop for mechanical degradation regime was captured during implementing our model in the simulator. It was found that imple-
menting the E-UVM in MRST for polymer non-Newtonian behavior might be more practical than the original method. In addition, the
comparison between various viscosity models proposed earlier and E-UVM in the reservoir simulator showed that the latter model could
yield more reliable oil recovery predictions as the apparent viscosity is modeled properly in the mechanical degradation regime, unlike
UVM or Carreau models.
This study presents a novel viscoelastic model that is more comprehensive and representative as opposed to other models in the litera-
ture. Furthermore, the need to conduct an extensive coreflooding experiment can be reduced by virtue of developed correlations that may
be used to estimate model fitting parameters accounting for shear-thickening and mechanical degradation.
Introduction
Fossil fuel remains the primary source to satisfy the continuously accelerating energy demand despite the recent development in the
renewable energy sector (Abas et al. 2015). Total oil consumption reached around 100 million barrels per day in 2020 (IEA 2020).
Moreover, population growth and global economic expansion in 2015–2035 are forecasted to increase the total energy needs by around
38%, where oil is expected to contribute the highest fraction of 27.4% and retain its first place in energy supply (BP 2014). Therefore, the
petroleum industry needs to apply more advanced techniques to enhance production from existing fields or make new field discoveries to
satisfy future energetic requirements. EOR is an essential stage in oil production that may increase ultimate oil recovery to more than half
of the reservoir’s original oil in place (Aladasani and Bai 2010). It is implemented by injecting chemicals, solvents, or high-temperature
steam into the reservoir (Lake 1989; Meng et al. 2015; Ansari et al. 2015; Abdelfatah et al. 2017; Khurshid et al. 2020).
Polymer flooding is a mature chemical EOR technique that has been extensively used to reduce the mobility of the injectant fluid.
Conventionally, polymer flooding technology was believed to improve macroscopic sweep efficiency by producing bypassed oil. This
fraction of movable oil remained unflooded after water injection due to higher water mobility and/or reservoir heterogeneities (Sheng
2011). Nevertheless, recently it has been found that polymers exhibiting viscoelastic behavior in the porous medium can also improve
microscopic displacement efficiency resulting in higher additional oil recovery. Residual oil mobilization by viscoelastic polymers was
related to extensional flow resistance in high shear rate regions. The stretchability of elastic polymers during converging flow results in
considerable extensional viscosity and improves pore-scale production (Azad and Trivedi 2021).
Polymer Flooding as an EOR Technique. Polymer flooding can successfully reduce the remaining oil saturation after waterflooding
and improve oil recovery. Polymer injection projects outnumber other chemical EOR types for a broader range of reservoir conditions.
Despite the lower incremental oil recovery, polymer flooding seems more attractive for EOR investments due to lower operational risks.
In general, the EOR polymers are classified into two main groups: biopolymers and synthetic polymers. The former typically include
xanthan gum, guar gum, welan gum, lignin, schizophyllan, sceluroglucan, and cellulose. Contrarily, synthetic polymers are usually com-
posed of partially hydrolyzed polyacrylamide (HPAM) and its derivatives (Gbadamosi et al. 2019). There is a consensus among research-
ers that biopolymers possess superior temperature and salinity tolerance capabilities than polyacrylamides (Akstinat 1980; Davison and
Mentzer 1982; Jensen et al. 2018). However, these polymers are susceptible to biological degradation and have poor injectivity (Hashmet
*Corresponding author; email: mursalzeynally@gmail.com

Copyright © 2023 Society of Petroleum Engineers
This paper (SPE 206010) was accepted for presentation at the SPE Annual Technical Conference and Exhibition, Dubai, UAE, 21–23 September 2021, and revised for publication.
Original manuscript received for review 30 September 2021. Revised manuscript received for review 30 June 2022. Paper peer approved 5 July 2022. Supplementary materials are
available in support of this paper and have been published online under Supplementary Data at https://doi.org/10.2118/206010-PA. SPE is not responsible for the content or functionality
of supplementary materials supplied by the authors.
February 2023 SPE Reservoir Evaluation & Engineering 99

et al. 2017; Gbadamosi et al. 2019). On the other hand, comparatively lower cost, better control on molecular weight distribution, and ease
of manufacturing on-site are the factors favoring HPAM for polymer flooding (Kim et al. 2010; Diab and Al-Shalabi 2019).
As a result, HPAM has been used in more polymer flooding field projects than biological polymers, as reported in several studies
(Standnes and Skjevrak 2014; Sheng et al. 2015). Moreover, unlike biopolymers, viscoelastic synthetic polymers possessing a flexible,
random-coil conformation may obtain dilatant behavior in porous media where their apparent viscosities increase several folds with shear
rate (Wang et al. 2008). Because of these viscoelastic properties, synthetic HPAM polymers may result in additional oil recovery (Hincapie
et al. 2017).
Carbonate Reservoirs and Harsh Conditions Limiting Polymer Flooding. It is worth mentioning that although polymer flooding is
an extensively applied chemical EOR technique in sandstones, its applicability in carbonates is still limited. Carbonate reservoirs account
for about 60% of the worldwide oil reservoirs. However, EOR applications in carbonates represent only 18% based on the international
EOR project database comprising 1,507 projects (Manrique et al. 2010). Most of these carbonate EOR field cases are either continuous
miscible gas or water alternating gas injection. Additionally, Masalmeh et al. (2019) reported that no successful polymer flooding project
was performed in a carbonate reservoir at high temperature and high-salinity conditions heretofore.
Generally, carbonate rocks are characterized by complex features due to the mixed-to-oil wettability nature and high heterogeneity
with low permeability. Harsh conditions also prevail in carbonates, such as high temperatures (above 85°C), high salinity (above 100,000
ppm), and high hardness (above 1,000 ppm). These conditions are challenging for the application of polymer flooding in carbonates (Diab
and Al-Shalabi 2019). Additionally, high-permeable natural fracture network in heterogeneous carbonate reservoirs may cause injection
fluid channeling, leaving a considerable amount of oil behind in unswept parts of the reservoir (Azad and Sultan 2014). Recently, new
types of polymers have been developed for overcoming the harsh conditions in carbonates. These novel synthetic polymers incorporate
several monomers, such as N-vinylpyrrolidone, 2-acrylamido-2-methylpropane-sulfonate, and sodium acrylamido tertiobutyl sulfonate,
that protect the polymer from thermal and chemical degradations. However, the viscoelasticity of these synthetic polymers is not clearly
comprehended and requires further investigation and research.
Polymer Viscoelasticity and Mechanical Degradation

Polymer Viscoelastic Properties. In general, polymer flow in porous media can be subdivided into two separate flow regimes: shear
dominant and extensional dominant. Most of the polymer solutions demonstrate pseudoplastic flow behavior in the shear dominant
region. On the other hand, the shear-thickening or dilatant behavior of synthetic polymer solutions occurs in the extensional dominant
flow regime.
It is important to note that polymer solutions show only shear-thinning behavior in a rheometer, while some polymers with viscoelastic
properties may exhibit shear-thickening behavior above a critical shear rate in a porous medium. The reason for such flow behavior is the
repeated variations of the capillary diameters that result in converging and diverging extensional flows, and consequently, stretching and
contraction of the polymer chains occurs (Heemskerk et al. 1984; Sheng et al. 2015). At high shear rates, the frequency of the deformation
of a polymer molecule grows so high that the polymer does not have time to relax to its original conformation. The resultant change in
the shape is called an elastic strain, and it is thought to be responsible for the additional pressure drop and thickening of the flow (Delshad
et al. 2008; Clarke et al. 2016; Tahir et al. 2017).
One significant parameter to represent the viscoelastic flow in porous media is the Deborah number (‍De‍). It may be explained as the
ratio between elastic and viscous forces (Skauge et al. 2018). Large Deborah numbers lead to solid-like behavior, while fluid-like behavior
occurs if the Deborah number is small (Sorbie 1991). Therefore, a material can behave solid-like behavior either because of its significant
elasticity or/and the swift deformation process with large viscous forces. Alternatively, the Deborah number is defined as the ratio of the
materials characteristic time (relaxation time) to the characteristic flow time (residence time) (Macosko 1994):
r
De = , (1)
‍ t ‍
where ‍ r ‍is the relaxation time and ‍t ‍is the residence time. The polymer characteristic relaxation time is the time taken by the polymers to
return to their original conformation after being disturbed. On the other hand, the residence time is the characteristic time for the flow in
the porous media required for the polymers to flow from one constriction to another (Wilton and Torabi 2013). Viscoelastic effects become
more pronounced when the fluid suffers a swift deformation where the residence time is comparable to the polymer relaxation time. In the
Deborah number calculations, the residence time is usually considered as the inverse of the either shear rate or strain rate (Azad and
Trivedi 2019). The effective shear rate (‍P ‍) is usually calculated as follows (Cannella et al. 1988):
2 n 3
1 !
1 6 3n1 + 1 n1 1 4uw 7
P = = C 4 p 5, (2)
t 4n1 8Sw kw
‍ ‍
where ‍P ‍is the effective shear rate, ‍t ‍is the residence time, ‍n1‍is the shear-thinning index, ‍uw ‍is the Darcy’s velocity, ‍Sw ‍is the aqueous phase
saturation, ‍ ‍is the porosity, ‍kw ‍is the effective aqueous phase permeability, and C ‍ ‍is the shear correction factor accounting for the differ-
ence between the actual porous medium and the equivalent capillary tube. Cannella et al. (1988) have taken C ‍ ‍as 6 for a number of core-
flooding experiments. Generally, Wreath et al. (1990) reported that C ‍ ‍is a function of transport and petrophysical properties and presented
various expressions for estimating the effective shear rate.
The Deborah number is a function of several molecular parameters and fluid flow conditions in porous media. These factors affecting
polymer viscoelastic behavior are listed in Table 1. According to the table, polymer molecular weight, concentration, and injection rate
positively affect its relaxation time and correspondingly viscoelasticity. With the increase in molecular weight, the average length of the
polymer chain increases, and it becomes more challenging for the molecule to relax, increasing its relaxation time and the Deborah num-
ber (Hester et al. 1994; Jiang et al. 2003; Delshad et al. 2008). On the other hand, the positive effect of polymer concentration can be
explained by intensifying the intermolecular interactions with the increase in polymer concentration in the solution. Therefore, it becomes
more difficult for the polymer to return to its original conformation, resulting in a higher relaxation time and more pronounced viscoelastic
effects (Heemskerk et al. 1984; Wang et al. 2000). Moreover, polymer injection rate is also believed to increase the Deborah number by
reducing the residence time leading to the stronger viscoelastic behavior (Qi 2018).
100 February 2023 SPE Reservoir Evaluation & Engineering

Effect on Viscoelasticity
Factors Positive Negative Inconclusive
Polymer molecular weight p
‍ ‍
Polymer concentration p
‍ ‍
Polymer injection rate p
‍ ‍
Reservoir temperature p
‍ ‍
Water salinity p
‍ ‍
Reservoir permeability p
‍ ‍
Water hardness p
‍ ‍
Table 1—Factors affecting polymer viscoelasticity.
Nevertheless, high temperatures and salinities in the reservoir reduce the relaxation time of the polymers and deteriorate its viscoelas-
ticity. The temperature effect is related to the increased kinetic motion of the individual atoms, which cause faster conformational rear-
rangements of a polymer chain (Jiang et al. 2003; Skauge et al. 2018; Alfazazi et al. 2019). On the other hand, a high-salinity environment
causes the collapse of the polymer molecular chain and leads to a much smaller random-coil structure with lower oscillatory relaxation
time and Deborah number (Garrouch 1999). Another factor influencing the viscoelastic properties is the permeability of the reservoir.
Polymers flowing in porous media with low permeability demonstrate more pronounced viscoelasticity due to the reduced median pore
space and higher differential pressures that increase the deformation of the polymer molecules and induce the viscoelastic flow even at
comparatively lower flow rates (Heemskerk et al. 1984; Ranjbar et al. 1992).
Finally, brines with high concentrations of divalent cations are generally believed to degrade the polymer and reduce its relaxation time
and the Deborah number (Tahir et al. 2017). However, the opposite effect of solution hardness was also reported, related to the chemical
reactions between the ions and polymer molecules (Walter et al. 2019).
Mechanical Degradation of Viscoelastic Polymers. Mechanical degradation may significantly reduce viscoelastic polymer efficiency.
This degradation leads to the breakage of the polymer backbone in the high flow rate regions due to significant stresses exerted on a
solution and results in considerable viscosity loss. This process usually takes place in polymer handling equipment, pumps, chokes,
and perforations. However, the polymers may also be irreversibly degraded in the porous medium, particularly in the wellbore vicinity
associated with high shear rates (Sorbie 1991). During mechanical degradation, chain scission causes higher molecular weight polymer
components to break into smaller ones altering molecular weight distribution and reducing the viscosifying power of polymer solution
(Sorbie 1991; Azad and Trivedi 2019).
Furthermore, most of the biopolymers with rigid rod structures are shear-stable resisting shear degradation. At the same time, synthetic
viscoelastic polymers are pretty susceptible to shearing, and their viscosities may be considerably reduced if mechanical degradation
occurs (Diab and Al-Shalabi 2019). It is related to the flexible coil structure of synthetic polymers that leads to dilatant behavior at high
flow rates. As a result, ultimate elongational stresses are exerted on a polymer molecule, stretching it apart. Beyond a critical flow rate,
the elongational stresses prevail macromolecular covalent bonding forces and result in polymer breakdown (Odell and Keller 1986).
The degree of mechanical degradation in porous media primarily depends on several factors, including formation permeability, injec-
tion rate, polymer molecular weight, brine salinity, and hardness. The degradation effect becomes more pronounced in tight formations
with lower permeabilities as larger stresses are exerted on the polymer solution flowing through narrower pore throats, leading to severe
degradation and pore plugging. Similarly, polymer molecular weight exacerbates mechanical degradation. In other words, polymers with
higher Mw are more likely to be degraded than smaller molecular weight polymers due to higher flow resistance and larger mechanical
stresses (Taylor and Nasr-El-Din 1995). Furthermore, Maerker (1975) and Martin (1986) reported that brine salinity and hardness also
negatively affect the process. They mentioned that it is possible to mitigate the polymer degradation by diluting or softening the injection
brine.
In addition to the aforementioned factors, the effects of polymer concentration and temperature on mechanical degradation were also
reported in the literature. Polymer concentration has a controversial impact on shear degradation. For instance, Maerker (1975) reported
that concentration does not affect mechanical degradation, while Morris and Jackson (1978) claimed that the degradation is exacerbated
with the reduction in polymer concentration. In contrast, Clemens et al. (2013) stated that the onset for mechanical degradation is shifted
to lower velocities with polymer concentration, where more degradation occurs. On the other hand, the temperature is believed to have a
detrimental impact on polymer stability to shearing. Such effect was related to the tendency of polymer bonds to be ruptured at higher
temperatures (Choi and Kasza 1989; Hamouda and Moshood 2007). These factors with the corresponding effects are listed in Table 2.
Effect on Mechanical Degradation

Polymer molecular weight p
‍ ‍
Polymer concentration p
‍ ‍
Reservoir temperature p
‍ ‍
Table 2—Factors affecting polymer mechanical degradation.

Effect on Mechanical Degradation
Water salinity p
‍ ‍
Reservoir permeability p
‍ ‍
Water hardness p
‍ ‍
Injection rate p
‍ ‍
Table 2 (continued)—Factors affecting polymer mechanical degradation.
Flow of Viscoelastic Polymers in the Porous Medium. Fig. 1 illustrates different flow regimes exhibited by viscoelastic polymers during
the flow in porous media. At low shear rates, the entropic forces exceed the hydrodynamic drag forces generated from the fluid flow. Thus,
polymer configuration remains in a coil shape, and the polymer .
behaves like a Newtonian fluid with constant viscosity during Regime 1.
As the shear rate increases and passes the first critical rate (‍ cr1),
‍ the polymer molecules face the larger drag forces that untangle polymer
coils and align them along the flow direction reducing their resistance to flow (Skauge et al. 2018). Regime 2, dominated by shear forces,
is called the shear-thinning region where polymer solutions lose their viscosities with the shear rate.
Fig. 1—Schematic of viscoelastic fluid flow behavior (modified after Sheng 2011).
Furthermore, as the shear rate increases further and passes the second critical rate (γ̇cr2), the polymer exhibits shear-thickening behav-
ior. It is associated only with flexible synthetic polymers. The dilatancy in Regime 3, dominated by elongational forces, is observed only
at high shear rates (near injection wells) when polymers undergo deformation called elastic strain as they flow through a series of pore
bodies and pore throats (Sheng 2011). Heemskerk et al. (1984) reported that the critical shear rate for shear-thickening increases with
increasing permeability, temperature, and salinity of the solution and with decreasing molecular weight and concentration of the
polymer.
Once viscoelastic polymers undergo full stretching during the dilatant flow, they accumulate excessive stresses that eventually lead to
chain scission of the polymer. After the third critical point (γ̇cr3), the apparent viscosity reaching the maximum value drops due to mechan-
ical degradation (Regime 4). As mentioned earlier, the high molecular weight of the polymer and low permeability of the rock both
accelerate this mechanical degradation.
Viscoelastic Models. Apparent viscosity or the viscosity observed in porous media is one of the primary indications of polymer flooding
efficiency. Therefore, several apparent viscosity models were designed to estimate in-situ viscosity and predict polymer rheological
behavior in porous media, which was previously depicted in Fig. 1. Examples of these models are given below:
Unified Viscoelastic Model. The apparent viscosity model proposed by Delshad et al. (2008) is frequently called UVM. The model
defines the apparent viscosity for both shear-thinning and shear-thickening regions. In this model, the shear viscosity is modeled by
Carreau expression, while elastic viscosity is estimated by relating it to the Deborah number and correspondingly to the polymer relax-
ation time:
h 2 i n1 1 /2 h n 1 i
app = 1 + 0 1 1 + 1 P + max 1 exp 2 r P 2 , (3)
‍ ‍
where ‍app‍is the polymer solution apparent viscosity, ‍0‍and ‍1‍are the zero-shear rate and infinite-shear rate viscosities, respectively;
‍n1‍is the shear-thinning index, ‍1‍is the shear-thinning constant that represents a transition between zero-shear Newtonian and pseudoplas-
tic regions, ‍P ‍ is the effective shear rate, ‍max ‍ is the maximum elongational viscosity, ‍2‍ is the shear-thickening constant that is usually
taken as 0.01, ‍ r ‍is the relaxation time, and ‍n2‍is the shear-thickening (or strain hardening) index.
According to UVM model, the relaxation time (‍ r ‍) and strain hardening index (‍n2‍) identify the level of polymer shear-thickening
behavior. Moreover, UVM also considers the maximum elongational viscosity (‍max )‍ before the mechanical degradation effect appears.
The model extensional parameters (‍n2‍, ‍max ‍) may be estimated by fitting the model to coreflooding test results, while shear parameters

(‍1‍ , ‍0‍ , ‍1‍ , ‍n1‍) are determined through bulk rheology measurements. The UVM has been successfully implemented in injectivity
models to match the injection pressure data from coreflooding and field tests (Lotfollahi et al. 2016). Even though UVM overcomes the
drawbacks of an unlimited increase in viscosity with respect to the Deborah number, it still needs coreflooding data to predict the model
parameters. Moreover, the drop in apparent viscosity due to mechanical degradation at ultimate shear rates cannot be characterized by
UVM as the model provides the plateau value of ‍max ‍.
Azad-Trivedi Viscoelastic Model (AT-VEM). Azad and Trivedi (2018) developed another viscosity prediction model called Azad-
Trivedi Viscoelastic Model (AT-VEM):
h 2 i n1 1 /2 h n2 1.21 i
app = 1 + 0 1 1 + 1 P + 0.35
max 1 exp 2 r P , (4)
‍ ‍
It is based on the UVM; however, AT-VEM does not require coreflooding data to predict the empirical parameters. The extensional param-
eters may be determined by a capillary breakup extensional rheometer. Based on capillary breakup extensional rheometer measurements,
the authors found that the maximum elongational viscosity occurs at the Deborah number of 0.66. Additionally, the downscaling factor
0.35 was applied to ‍max ,‍ and subtrahend 1.2 was used with ‍n2‍ to scale down the pure elongation in the extensional bulk field to the
combination of elongation and shear encountered in the porous media.
Extended Viscoelastic Model (Stavland et al. 2010). This model determines apparent viscosity in both shear- and elongation-dominated
regions and accounts for polymer mechanical degradation at high shear rates. However, the model needs extensive coreflooding measure-
ments to properly model the polymer rheology in shear-thickening and mechanical degradation regimes:
j
m1 m2 x (5)
‍ app = 1 + 0 1 1 + r1 P
+ r 2 P
1 + d P
x ,‍
where ‍r1‍is the time constant, ‍r2‍is the inverse of elongation onset, ‍m1‍is the shear-thinning exponent, ‍m2‍is the elongational exponent, ‍d ‍
is the characteristic time constant for degradation, ‍x ‍is the is a constant that is usually taken as 2, and ‍j ‍is the tuning parameter for degra-
dation. Similarly, shear parameters may be found through shear rheology measurements. On the other hand, the parameters accounting
for elongation and mechanical degradation are estimated by coreflood data and effluent sample.
From the literature above, one can note that mechanical degradation cannot be captured in the latter models or it requires extensive
coreflooding data. Therefore, a novel viscoelastic model was proposed to comprehensively analyze the polymer rheological behavior in
porous media. This proposed model was validated using 30 coreflooding tests obtained from the literature and further verified against a
few existing viscoelastic models. Based on this model, an accurate prediction of the complex rheological response of polymers in porous
media was captured to obtain a proper and representative estimation of incremental oil to polymer flooding. Furthermore, it is worth
mentioning that the need to conduct an extensive coreflooding experiment was reduced by virtue of developed correlations that were used
to estimate model fitting parameters accounting for shear-thickening and mechanical degradation.
Methodology
The approach followed in this study to develop an extended viscoelastic model and further use it to simulate the effect of polymer flooding
on oil recovery is summarized in Fig. 2. First, the E-UVM was developed and validated using 30 coreflooding test results collected from
the literature. These measurements, mainly linear with two radial coreflooding tests, were performed in sandstone and carbonate cores
with corresponding rock properties. Moreover, both synthetic polymers and biopolymers with various molecular weights, concentrations,
and makeup water salinities were utilized in the measurements. The comparatively broader experimental conditions were used in fitting
aid in understanding and analyzing the effect of different factors on polymer in-situ behavior and controlling the corresponding fitting
parameters of the proposed E-UVM. Furthermore, the fitting parameters of the model were correlated to rock and polymer solution prop-
erties using the symbolic regression machine learning approach embedded in EUREQA software. This stage was implemented to reduce
the need to perform extensive coreflooding tests for predicting fitting model parameters.
Fig. 2—Flow chart of methodology used to develop the E-UVM and simulate polymer flooding at core scale.
Next, the developed apparent viscosity model was implemented in MRST, replacing the original shear model existing in the simulator.
Then, the implementation was verified against the original PLYSHEAR model. Finally, the effect of various rheological models on oil
recovery was also captured in this study. A description of each of these steps is presented in the following section.

Results and Discussion
This section includes a description and analysis of the developed E-UVM. The model validation against different experimental data from
the literature is also included. In addition, development of the empirical correlations for estimating the E-UVM model parameters is dis-
cussed. This is followed by core-scale simulations of the polymer flooding effect on oil recovery using E-UVM model.
Development of E-UVM. As discussed before, several analytical models were proposed to predict aqueous phase apparent viscosity
during polymer flooding. However, most of these models either cannot comprehensively characterize polymer behavior in porous media
or require extensive coreflooding data to estimate the model fitting parameters. Therefore, a new model is proposed, called E-UVM,
which captures all four regimes observed during viscoelastic polymer flooding, including Newtonian, shear-thinning, shear-thickening,
and mechanical degradation regimes.
As the name indicates, the E-UVM is based on the UVM published earlier (Delshad et al. 2008). In E-UVM, the Newtonian and shear-
thinning regimes observed at low shear rates are captured using the Carreau model. In contrast, the shear-thickening behavior of visco-
elastic polymers usually observed near the wellbores is modeled by relating elastic viscosity to the Deborah number. Although the original
UVM provides the plateau value of ‍max ‍, it cannot characterize the drop in apparent viscosity due to mechanical degradation at ultimate
shear rates. Nevertheless, it is essential to consider the mechanical degradation regime in the viscoelastic models since it may significantly
impact polymer injectivity and oil recovery predictions. Therefore, a degradation term is added in E-UVM to capture the reduction in
polymer apparent viscosity beyond critical shear rates. This term is generated empirically based on the collected coreflooding data from
the literature with the idea of introducing two parameters: one to control the onset for mechanical degradation and another one to control
the degree/severity of degradation. The E-UVM is shown below (Eq. 6):
0 n1 1 1 n3
2 h n2 1 i
app = @1 + 0 1 1 + 1 P 2 + max 1 exp 0.01r P A 1 + 3 P 2 2 , (6)
‍ ‍
Water
Mw Cp S Core Length Temperature Saturation
№ References Polymer Type (MDa) (g/L) (g/L) Core Type (cm) k (Darcy) Porosity (°C) (%) Setup
1 Southwick Exp1 20 2.5 0.048 NA 100

Glass Bead
and Manke HPAM 0.5 20 0.25 linear
2 Exp2 Pack 2.5 0.022 NA 100
(1988)
3 Delshad et al. (2008) Flopaam 3,630 S 20 1.5 20 Berea 20.32 NA NA NA 100 linear
4 BV1 Flopaam 3,630 S 20 20.04 (SFB) 20.32 0.647 0.23
5 BV2 AN-125 8 20.04 (SFB) 20.32 0.578 0.24

Magbagbeola
6 BV3 HJ 63020 20 1.5 20.04 (SFB) Berea 20.32 0.552 0.235 23 100 linear
(2008)
7 BV4 20 20.04 20.32 0.26
HJ 63020 0.243
(SSFB)
8 Exp1 Flopaam 3,830 S 20 3 14.5 0.573

Seright et al.
9 Exp2 Flopaam 3,830 S 20 1 41.95 Berea 14.5 0.573 0.21 25 100 linear
(2009)
10 Exp3 Xanthan K9D236 2 41.95 14.5 0.551
11 Stavland et al. (2010) HPAM 20 1.5 35 Berea 7 0.361 0.22 20 100 linear
12 Manichand Exp 1 18 1 Reservoir NA 4

Flopaam 3,830 S 0.5 0.3 25 100 linear
13 et al. (2013) Exp 2 1.35 Sandstone Core NA
14 Brakstad and 20 9.27 1.26

AMPS 1 103.3 Bentheimer 0.23 20 100 linear
Rosenkilde (2016)
15 Skauge et al. Exp1 18 28.95 2.34 0.228 linear

Flopaam 3,630 S 2 4.659 Bentheimer 22 100
16 (2016) Exp2 30-diameter 2.6 0.24 radial
17 Lohne et al. Exp1 15 7 0.1613 0.177

HPAM 3530 S 1.5 38.88 Berea 20 100 linear
18 (2017) Exp2 7 2.0188 0.235
19 Asen et al. (2019) HPAM 20 1 38.81 Bentheimer 3 2.17 0.21 20 100 linear
20 Skauge et al. Exp1 Hydrophobically NA 29.1 2.433 0.235 linear

1 10 Bentheimer 22 100
21 (2019) Exp2 modified polyacrylamide 30-diameter 2.09 0.234 radial
22 C-2 8 4.15 1.616 0.26 100
23 Masalmeh C-4 Reservoir 5.64 0.107 59 (Sor =

SAV10 1 180 0.291 50 linear
et al. (2019) carbonate core 41%)
24 C-6 6.62 0.121 0.304 100
25 CF-1 8 0.5 7.315 0.19852

0.183 25 100 linear
26 1 7.315
27 CF-2 0.5 Indiana 7.416 0.16513

Alfazazi et al. Sulfonated ATBS-
243.028 limestone 0.189 50 100 linear
28 (2021) based polymer 1 7.416
outcrop
29 CF-4 0.5 7.442 0.1826 61 (Sor =
0.185 50 linear
30 1 7.442 39%)
Table 3—Summary of experimental data from the literature used to validate E-UVM.

where shear-thickening constant ‍ 2‍) is set as 0.01, ‍3‍is the mechanical degradation constant that is usually in the range of 10−5-10−3, and
‍n3‍ is the mechanical degradation index that usually varies in the range of 0.3 to 1. Mechanical degradation constant (‍3)‍ controls the
degradation onset; higher ‍3‍shifts onset to the lower shear rates. Furthermore, the degree of degradation is controlled by the mechanical
degradation index (‍n3)‍ ; higher ‍n3‍in E-UVM is used to represent more severe degradation cases.
Validation of E-UVM. Several coreflooding results reported in the literature were extracted and used to validate the proposed model.
The experimental conditions for each study are listed in Table 3. The latter table summarizes the experimental data including polymer
molecular weight (Mw) and concentration (Cp), aqueous solution salinity (S), core length, permeability (k) and porosity (‍ ),
‍ experimental
temperature and setup as well as water saturation used during coreflooding tests. Moreover, polymer and rock types are also indicated in
this table. One should note that NA is used for the cases where data is not applicable or not available.
Furthermore, both the original UVM and E-UVM were fitted to the data from 30 experiments to highlight the improvement in polymer
rheological behavior prediction using the E-UVM. Mechanical degradation occurring at ultimate shear rates and corresponding viscosity
reduction were well captured using the E-UVM. Some examples of the model fitting to experimental data are illustrated in Figs. 3
through 9. The apparent viscosities indicated in these figures were measured during coreflood experiments. Moreover, the bulk viscosi-
ties, if available, were obtained from the bulk rheology tests conducted in corresponding studies.
Fig. 3—Fitting the models to Experiment 22 data [1 g/L-8 MDa-SAV10 at 180 g/L total dissolved solids (TDS), 1.616 D Reservoir
Carbonate Core].
Fig. 4—Fitting the models to Experiment 24 data (1 g/L-8 MDa-SAV10 at 180 g/L TDS, 0.121 D Reservoir Carbonate Core).

Fig. 5—Fitting the models to Experiment 8 data (1 g/L-20 MDa-Flopaam 3,830 S at 3 g/L TDS, 0.573 D Berea Sandstone Core).
Fig. 6—Fitting the models to Experiment 9 data (1 g/L-20 MDa-Flopaam 3,830 S at 41.95 g/L TDS, 0.573 D Berea Sandstone Core).
Fig. 7—Fitting the models to Experiment 10 data (1 g/L-2 MDa-Xanthan K9D236 at 41.95 g/L TDS, 0.551 D Berea Sandstone Core).

Fig. 8—Fitting the models to Experiment 27 data (0.5 g/L-8 MDa-ATBS-based polymer at 243.028 g/L TDS, 0.189 D Indiana Limestone
Outcrop).
Fig. 9—Fitting the models to Experiment 28 data (1 g/L-8 MDa-ATBS-based polymer at 243.028 g/L TDS, 0.189 D Indiana Limestone
Outcrop).
It is worth mentioning that Experiment 10 was conducted using xanthan polymer that does not exhibit shear-thickening in porous
media. UVM and E-UVM models can predict the rheological behavior of the biopolymer by simply setting some of the fitting parameters
to zero. As shown in Fig. 7, both models are in a match as there were no shear-thickening and mechanical degradation observed.
Factors Affecting Polymer Shear-Thickening and Mechanical Degradation. The effect of different factors on polymer dilatant
behavior and mechanical degradation was also analyzed. It was observed that decreasing rock permeability shifts the onset for shear-
thickening and mechanical degradation to the lower shear rates. The permeability effect was captured in Experiments 1 and 2 (Southwick
and Manke 1988), Experiments 22 and 24 (Masalmeh et al. 2019), and Experiments 17 and 18 (Lohne et al. 2017). For example,
Experiments 22 and 24 (Figs. 3 and 4, respectively) were conducted under the same conditions but using cores of different permeabilities.
It is seen from the figures that the onsets for shear-thickening and mechanical degradation in Experiment 22 (k = 1,616 md) are around
100 and 3,000 seconds−1, respectively. These critical shear rates for Experiment 24 conducted using a core with lower permeability (k =
121 md) are approximately 20 and 1,000 seconds−1. Similar observations are applicable for the other two sets of experiments mentioned
above (1–2 and 17–18). Such impact is related to a rapid increase in the stresses exerted on the polymer solution in low-permeable rock
enhancing polymer deformation and accelerating dilatancy and degradation.
The solution salinity is another factor affecting polymer rheological behavior in porous media. Experiments 8 and 9 (Figs. 5 and 6, respec-
tively) illustrate that the polymer solution with 3 g/L TDS exhibits stronger viscoelasticity than the one with the salinity of 41.95 g/L. It resulted
in earlier shear-thickening and higher maximum elongational viscosities (‍max‍) in porous media. Namely, the onset for dilatant behavior in
Experiment 8 is around 20 seconds−1, while for Experiment 9, it is ten times greater, confirming a negative salinity effect on the viscoelastic
onset. On the other hand, mechanical degradation also depends on salinity. It was found that the degradation becomes more severe with higher
salinity. This is reflected in the steeper slopes in the degradation regime, where more viscosity loss is observed with the shear rate. Particularly,
the slope was varied from −21.5 to −270 when the salinity increased from 3 g/L to 41.95 g/L, confirming more rapid degradation at higher
salinities. However, there was a negligible salinity impact on the mechanical degradation onset (≈ 1,200 seconds−1 for both).

In addition, the effect of polymer concentration was investigated by comparing Experiments 12–13, 25–26, 27–28, and 29–30. It was
found that an increase in polymer concentration can enhance its dilatant behavior. Particularly, the onset for shear-thickening was shifted
to lower shear rates for polymers at higher concentrations, as reported in the literature (Heemskerk et al. 1984; Wang et al. 2000). This
can be explained by the larger relaxation times observed at higher polymer concentrations. Additionally, the increment in concentration
led to higher maximum elongational viscosity (‍max ‍) and shear-thickening index (‍n2‍). For instance, the slope of polymer shear-thickening
increased from 10.6 to 23.8 with polymer concentration in Experiments 12 and13. In contrast, the effect of polymer concentration on
mechanical degradation was found to be questionable. The critical shear rates for mechanical degradation and corresponding mechanical
degradation constants (‍3)‍ did not considerably vary with polymer concentration in most of the experiments conducted at room tempera-
ture. However, Experiments 27 and 28 (Figs. 8 and 9) performed at temperature of 50°C revealed that higher polymer concentration
might accelerate mechanical degradation in terms of shear rates and result in more severe degradation process. Higher mechanical degra-
dation constant (‍3‍) and index (‍n3‍) were used for 1,000 than 500 ppm polymer solutions to accommodate smaller critical shear rates and
steeper slopes for mechanical degradation regime, respectively.
Furthermore, it is believed that radial coreflooding can better mimic the actual conditions existing in a reservoir. Therefore, Experiments
16 and 21 performed using radial coreflooding setup were analyzed. In particular, Experiment 16 was conducted using HPAM polymer.
In these experiments, less extensive polymer shear-thickening behavior was observed in the radial coreflood compared to the linear setup
performed under the same conditions. This was thought to be due to the various flow fields, which are constant in linear but gradually
decreasing in radial flooding. In contrast, Experiment 21 was conducted using Hydrophobically Modified Polyacrylamides (HMPAMs),
showing only shear-thinning for this polymer in radial coreflood, and no indication of shear-thickening was found. Similar to the previous
study, lack of dilatancy and better injectivity in radial coreflooding was related to the decreasing shear rates with the distance from the
center of the core. In addition, the hydrophobic nature of the hydrophobically modified polyacrylamide was thought to be another reason
of polymers exhibiting no shear-thickening and moderate shear degradation. Although the in-situ behavior of these two viscoelastic poly-
mers was not consistent, the experiments showed that the degree of shear-thickening and mechanical degradation might be significantly
reduced in radial setup, as opposed to the linear under the same conditions. However, more radial polymer flooding experiments need to
be conducted to better understand the effect of the experimental setup on polymer rheology in a porous medium, both qualitatively and
quantitatively.
Based on the discussion above, it was found that polymer properties, rock permeability, and salinity can significantly affect polymer
rheological behavior. Particularly in carbonates, the heterogeneity with low permeability may result in early shear-thickening behavior
and mechanical degradation (Figs. 3 and 4). Although the polymer elasticity and dilatant behavior are enhanced in low-permeable
rock, severe mechanical degradation may occur if injection rates and polymer molecular weight are not appropriately designed. The
proposed E-UVM with the embedded correlations that control its fitting parameters based on the rock and polymer properties can
estimate polymer rheology at various shear rates and can be used to select proper polymer design for the field implementation. The
techniques to select the polymers suitable for a given rock type are out of the scope of this paper and can be found elsewhere (Guo
2017). In addition, high brine salinity in carbonates may also deteriorate the viscoelastic behavior of polymers and accelerate mechan-
ical degradation.
Furthermore, reservoir temperature and polymer molecular weight were not discussed. However, these factors are believed to have a
considerable effect on polymer shear-thickening and degradation. Elevated temperatures in carbonates may result in substantial polymer
thermal degradation and impair the elastic properties. On the other hand, injected polymer molecular weight is believed to improve the
dilatant behavior. Nevertheless, injecting high molecular weight polymers in low-permeable rock may plug the pore space and negatively
affect injectivity. Therefore, further experimental studies are required to comprehensively investigate the effect of each factor on polymer
rheological behavior in carbonate rocks.
Moreover, the proposed E-UVM can estimate polymer rheology at various shear rates and can be used to select the proper polymer
design for field implementations. The factors discussed above have a particular effect on polymer rheology in porous media, and their
effect can be captured in E-UVM through the model fitting parameters. These parameters can either be estimated by fitting the model to
the preliminary coreflooding test data or by calculating them using the correlations proposed in this study. These correlations relating
fitting parameters to certain polymer and rock properties are discussed in the following section.
Empirical Correlations for Estimating the Model Parameters. This section discusses the model parameters used for fitting E-UVM
to the coreflooding data. Moreover, the empirical correlations generated using the machine learning technique to relate those fitting
parameters to rock and fluid properties are also included in this section. Polymer shear-thinning behavior in E-UVM is controlled by the
shear-thinning constant (‍1‍) and index (‍n1‍), as in the original Carreau model. On the other hand, the shear-thickening section of the model
depends on the Deborah number. It was found that the critical shear rate when the dilatant behavior is first observed is a function of poly-
mer relaxation time. As discussed in Table 1, the relaxation time of polymers strongly depends on fluid flow conditions in porous media
and polymer molecular parameters. Higher relaxation time leads to the earlier shear-thickening onset. Additionally, the slope for dilatant
behavior is controlled by another fitting parameter, shear-thickening index (‍n2‍). Higher polymer molecular weight and concentrations as
well as lower salinities increase ‍n2‍leading to more pronounced shear-thickening.
Furthermore, the model parameters responsible for mechanical degradation are also discussed. Namely, mechanical degradation con-
stant (‍3‍) increases mainly with polymer molecular weight and decreases with rock permeability. Neglecting this parameter in the model
converts the E-UVM to the original UVM model. On the other hand, the mechanical degradation index (‍n3‍) h was previously discussed
and can capture severe mechanical degradation cases with its steeper slope. Since the xanthan gum and other biopolymers usually do not
exhibit shear-thickening and mechanical degradation in porous media, ‍max ‍and ‍3‍for experiment 10 were tuned as zero, converting the
E-UVM to the Carreau model. All fitting parameters determined by matching the proposed E-UVM model to the coreflooding data from
literature are listed in Table 4.
Most of the existing viscoelastic models for predicting apparent polymer viscosity at various shear rates require extensive coreflooding
tests to estimate the fitting parameters. However, generally, conducting the coreflooding experiment is a time-consuming and costly pro-
cess. Besides, the cores obtained from a reservoir are usually not available as their extraction may be challenging and expensive. Therefore,
difficulties in conducting extensive coreflooding experiments strongly limit the application of traditional viscoelastic models. Consequently,
in this study, we attempted to reduce the need for coreflooding data to predict polymer rheological behavior in porous media as it might
lead to faster polymer screening for future studies and polymer field applications.
For this purpose, the E-UVM fitting parameters were related to rock and polymer properties using several correlations developed in
the EUREQA software. EUREQA is a software that detects latent mathematical relationships in the data using a machine learning method
called symbolic regression and provides a mathematical expression to represent that relationship (Informer Technologies 2021). Therefore,
the data from Tables 3 and 4 were imported and analyzed in the software. The output correlations with reasonable accuracy and

№ References ‍0‍ ‍1‍ ‍1‍ ‍n1‍ ‍max ‍ ‍2‍ ‍ r ‍ ‍n2‍ ‍3‍ ‍n3‍
1 Southwick and Exp1 20 1.02 0.32 0.662 24 0.01 0.225 2.18 0.00034 0.48
2 Manke (1988) Exp2 5 1.02 0.32 0.662 23.4 0.01 0.24 2.21 0.00056 0.39
3 Delshad et al. (2008) 13.7 1.02 0.221 0.755 108 0.01 0.035 2.15 NA NA
4 BV1 13.76 1.02 0.221 0.755 65 0.01 0.085 2.3 0 NA
5 Magbagbeola BV2 10.15 1.02 0.25 0.802 90 0.01 0.06 2.05 0 NA
6 (2008) BV3 6.54 1.02 0.066 0.828 21 0.01 0.045 2.3 0 NA
7 BV4 18.09 1.02 0.175 0.720 82 0.01 0.067 2.25 0 NA
8 Exp1 38 1.02 0.25 0.56 299 0.01 0.32 1.957 0.00029 0.83
Seright et al.
9 Exp2 5 1.02 0.1 0.81 26 0.01 0.13 3.4 0.00032 0.92
(2009)
10 Exp3 80 1.02 0.23 0.48 0 0.01 NA NA 0 NA
11 Stavland et al. (2010) 27 1.02 0.51 0.715 37.2 0.01 0.047 3.25 0.00014 0.56
12 Manichand et al. Exp 1 40 1.02 0.62 0.6 172 0.01 0.365 2.14 0.0009 0.43
13 (2013) Exp 2 114 1.02 0.68 0.59 237 0.01 0.38 2.29 0.00092 0.44
14 Brakstad and Rosenkilde 4 1.02 0.62 0.842 41.6 0.01 0.09 3.2 0.000132 0.694
(2016)
15 Skauge et al. Exp1 20 1.02 0.36 0.76 55 0.01 0.49 2.15 0 NA
16 (2016) Exp2 55 1.02 0.08 0.14 45 0.01 0.15 1.6 0 NA
17 Lohne et al. Exp1 4 1.02 0.25 0.79 13.8 0.01 0.22 2.88 0.00048 0.85
18 (2017) Exp2 6 1.02 0.45 0.87 46.2 0.01 0.098 3.35 0.00041 0.41
19 Asen et al. (2019) 10 1.02 0.15 0.74 27 0.01 0.15 3.15 0.000235 0.68
20 Skauge et al. Exp1 2.8 1.02 0.34 0.89 12 0.01 0.12 2.75 0.00085 0.49
21 (2019) Exp2 6.4 1.02 0.34 0.97 0 0.01 NA NA 0 NA
22 C-2 5.05 1.02 0.09 0.59 67 0.01 0.036 2.1 0.000335 0.9
Masalmeh et al.
23 C-4 32 1.02 0.23 0.54 17 0.01 0.03 2.06 0.00027 0.43
(2019)
24 C-6 5.05 1.02 0.14 0.87 12.4 0.01 0.25 2.05 0.00043 0.45
25 8 1.02 0.15 0.32 19.8 0.01 0.047 2.01 0.000071 0.72
CF-1
26 18 1.02 0.17 0.3 27 0.01 0.051 2.03 0.000072 0.72
27 Alfazazi et al. 3 1.02 0.34 0.95 12.6 0.01 0.028 2.15 0.000062 0.32
CF-2
28 (2021) 9 1.02 0.62 0.86 27.7 0.01 0.047 2.21 0.00012 0.34
29 13 1.02 0.14 0.52 10.4 0.01 0.012 2.56 0.000044 0.39
CF-4
30 24 1.02 0.18 0.51 18 0.01 0.016 2.61 0.000047 0.42
Table 4—Fitting parameters of E-UVM used to match experimental data from literature.
complexity were then selected among the developed expressions and used in this study. These expressions relating E-UVM fitting param-
eters to polymer and rock properties as well as solution salinities are presented below:

1.19 Cp 110 k
max = 0.226 kS + 339 259 1.15 103 ln, (7)
‍ S+k ‍
0.5
‍ r = 0.12 Mw + 0.00447 S + 0.01242 kMw 0.104 0.249 k 0.0769 S 7.5 108 Cp2 0.00444M2w ,‍ (8)
2 5
2
7
3
‍ n2 = 2.07 + 0.00102 S + 7.2 10 Mw S 0.0377 S 9.62 10 Mw S ,‍ (9)

0.00756 k 0.03316
3 = 3.17 105 Mw + 1.7 107 kCp + + 0.0008142 0.0003905 k, (10)
‍ M w kC p ‍

‍ n3 = 42.9 + 0.00656 Mw + 0.00236 S + 0.000274 S2 1.13 120 2 2.14 0.0293 S ,‍ (11)
where ‍k ‍is the absolute permeability in Darcy, ‍S ‍is the solution salinity in g/L, C
‍ p‍is the polymer concentration in ppm, ‍ ‍is the porosity
in fraction, and ‍Mw ‍is the polymer molecular weight in MDa.
These correlations may be used to predict the fitting parameters of the E-UVM with an acceptable error if the coreflooding data are not
present. Blind tests were conducted to test the accuracy and predictability of the developed correlations. For this purpose, the data from
experiments 9, 12, and 23 were not used in EUREQA as an input to develop the correlations and kept for validation. The results for these
experiments are shown in Figs. 10 through 12, respectively, where the E-UVM is plotted using original fitting parameters reported in
Table 3 (Case A) and modified fitting parameters calculated from the EUREQA correlations (Case B). It is worth mentioning that the
accuracy of the blind test was slightly reduced in Fig. 12 because the salinity of the polymer solution used in Experiment 23 was
extremely high and out of the suggested salinity range. However, there is a minor error in general, and E-UVM may be used to predict
polymer rheological behavior for future polymer screening even if extensive coreflooding measurements are not performed. The

Fig. 10—First blind test—E-UVM results for Experiment 9 using original fitting parameters (Case A) and modified fitting parameters
(Case B).
Fig. 11—Second blind test—E-UVM results for Experiment 12 using original fitting parameters (Case A) and modified fitting
parameters (Case B).
limitations of the proposed correlations and the ways to improve their accuracy are discussed at the end of the paper. Moreover, additional
information regarding the blind tests can be found elsewhere (Supplementary Material A, Fig. S-1).
Implementation of E-UVM in MRST. The core-scale polymer flooding simulation studies were performed in MRST. MRST is an open-
source software based on black-oil formulation with fully implicit time stepping and upstream weighting for spatial discretization. The
software has been developed into a practical platform for rapid prototyping and testing the new simulation methods and mathematical
models (Lie 2016; Bao et al. 2017). The core-scale simulation studies were performed mainly to implement the proposed extended unified
apparent viscosity model in the reservoir simulator and verify it against the original rheological model existing in MRST as well as for
polymer rheological studies. As was previously mentioned, E-UVM successfully captures all flow regimes exhibited by polymers at
different shear rates. Therefore, it was decided to use this model in simulation studies to predict polymer rheological behavior.
The core-scale simulation studies of this work were performed under representative conditions of the Middle East carbonate reservoirs.
The experimental two-phase flow data for oil and water as well as that of the rock were obtained from Chandrasekhar and Mohanty
(2013). These authors conducted a series of vertical coreflooding measurements to analyze the effect of brine ionic composition on oil
recovery for limestone reservoirs. The first coreflooding test was performed by injecting the formation brine with a TDS of 179,730 ppm
through heterogeneous carbonate core plugs with rock and fluid properties given in Tables 5 through 7. The formation water injection
was performed with an injection velocity of 1 ft/day at a temperatures of 120°C. The authors reported about 40% original oil in place
recovery through injecting 5 pore volumes of formation water. Additional information about the experimental study can be found else-
where (Chandrasekhar and Mohanty 2013). In addition to the experimental two-phase flow data, Table 8 lists the polymer parameters that
were assumed and used for polymer flooding simulations.

Fig. 12—Third blind test—E-
UVM results for Experiment 23 using original fitting parameters (Case A) and modified fitting
parameters (Case B).
Pore volume (cm3) 14.6 Length (cm) 7.8

Porosity 0.178 Permeability (md) 23.2
Cross sectional area (cm2) 10.987 Diameter (cm) 3.74
Swirr 0.21 Sorw 0.47
Injection Rate (mL/min) 0.045
Table 5—Core petrophysical properties (Chandrasekhar and Mohanty

2013).
Viscosity at 120°C (cp) 0.9

Density at 120°C (g/cm3) 0.82
API (degrees) 41.06
Table 6—Oil properties

(Chandrasekhar and Mohanty 2013).
Ion\brine Concentration (ppm)

Na+ 49,933
2+
Mg 3,248
Ca2+ 14,501
Cl− 111,810
2−
SO4 234
TDS 179,730
Table 7—Water properties

(Chandrasekhar and Mohanty 2013).
Core-Scale Simulation Model. The core-scale simulation model was designed similar to the work of Adegbite et al. (2018). The het-
erogeneous 2D Cartesian grid system incorporating 10×1×10 gridblocks was used to simulate viscoelastic polymer flooding. The model
dimensions are presented in Table 9. Moreover, the core heterogeneity was modeled using the log-normal permeability distribution in x-
and z- directions as well as a sequential Gaussian simulation method with a spherical variogram. The standard deviation of the log-normal
distribution can be related to Dykstra Parson’s coefficient of variation as follows (Dykstra and Parsons 1950):
‍ DP = 1 e ,‍ (12)

Polymer Type HPAM-Based
Molecular weight (MDa) 8
Degree of hydrolysis 0.28
Concentration (ppm) 500, 1,000, 1,500, 2,000
Table 8—Polymer properties used in polymer

flooding simulations.
where σ is the standard deviation and DP is the Dykstra Parson’s coefficient of variation, which is taken as 0.8 for the highly heteroge-
neous simulation model used. The permeability obtained ranged from 19 to 27.3 md with an average permeability of 23.2 md.
Parameter Value Comments
Number of gridblocks 100 2D Model (10×1×10)
Gridblock size (m) ‍ x = 0.0033145‍
Constant grid size in each
‍ = 0.033145‍
y direction
‍ z = 0.0078‍

Core model dimensions (m) 0.033145×0.033145×0.078 Length × width × height
Table 9—Core-scale simulation model data.
The permeability distribution of the model used is depicted in Fig. 13. The simulation model was designed vertically based on the
experimental work of Chandrasekhar and Mohanty (2013), with two horizontal wells; production well at the top and injection well at the
bottom of the model. Moreover, the production well was set at a constant bottomhole pressure of 50 psi, while the injection rates for the
injector during polymer flooding varied from 0.045 (≈ 1 ft/d) to 4.5 mL/min (≈ 100 ft/d) to capture the flow rate effect on viscoelastic
polymer flow.
Fig. 13—Core-scale simulation model with heterogeneous permeability.
Original Polymer Model Used in MRST. The polymer model used in MRST assumes that the polymer component is transported in the
aqueous phase through changing its mobility only. On the other hand, the oleic and gaseous phases are not affected by the polymers and
are represented with the standard three-phase black-oil equations. Polymers can change aqueous phase mobility in several ways. First,
polymers mixed in the aqueous phase can increase viscosity. Moreover, polymers can reduce the effective permeability to water by being
adsorbed on the rock surface or retained in the narrow pore throats. These factors are captured in MRST using the Todd-Longstaff mixing
model (Bao et al. 2017).
In addition to the mechanisms discussed above, aqueous phase mobility can also be dramatically altered with the shear rates. This
mobility change during polymer flooding is related to polymer non-Newtonian behavior, where the viscosity changes with water velocity.
Originally, the non-Newtonian behavior of polymers is captured in MRST by importing the data from external files/tables and iteratively
solving for shear multipliers, this is possible using default PLYSHEAR model/keyword or logarithm-based models, such as PLYSHLOG

and SHRATE. For this purpose, the user-prescribed factors (‍msh‍) are tabulated in the input files as a aqueous phase velocity or shear rate
function. Within this table, the velocity values should monotonically increase starting from zero. The first ‍msh‍factor that corresponds to
zero velocity is usually equal to one as negligible shear effects are observed at minimal fluxes. Then, the factors may decrease with veloc-
ities if the polymer exhibits shear-thinning or may increase in the shear-thickening case (Schlumberger 2013). When the viscosity depen-
dence on flux/shear rate is activated, the MRST model determines the ‍msh‍ factor corresponding to the specific shear-modified water
velocity (‍uw, sh)‍ and uses it to iteratively solve for new ‍uw, sh.‍ Eventually, the final shear-modified velocity and respective ‍msh‍are used
to estimate the shear multipliers (‍Z ‍) representing polymer non-Newtonian behavior in MRST. The more detailed information for
PLYSHEAR and other models used in MRST for polymer non-Newtonian behavior can be found elsewhere (Schlumberger 2013; Bao
et al. 2017).
Our study aimed to replace the original technique used in MRST to capture shear effects with the E-UVM that can independently from
input deck files estimate the apparent polymer viscosities and corresponding shear factors (‍Z ‍) at various shear rates. The implementation
of the model in MRST is described in the next section.
Proposed E-UVM in MRST. To implement the E-UVM in MRST, the water Darcy’s velocities (‍uw ‍) were calculated on each face ‍F ‍
of the discrete grid as follows:
qw,0 F
uw F = , (13)
‍ A F ‍

where ‍qw, 0 F ‍is the water flux across the interface ‍F ‍and ‍A F ‍is the face ‍F ‍area. Furthermore, the shear rate values (‍P ‍) used in E-UVM
were also computed on the faces of grids, as shown below:
2 n 0 13
1
6 3n1 + 1 n1 1 B 4uw F C7
P F = C 4 @q A5 , (14)
4n1 8 Sw kw F
‍ ‍
where C ‍ ‍is the shear correction factor, ‍n1‍is the shear-thinning index, ‍uw ‍is the water Darcy’s velocity, ‍Sw ‍is the aqueous phase saturation,
‍ ‍is the porosity, and ‍kw ‍is the effective aqueous phase permeability for the face ‍F ‍. Moreover, apparent viscosities (‍app)‍ were estimated
at interfaces using the E-UVM, as follows:
0 n1 1 1 n3
2 h n2 1 i 2
@
app F = 1 + 0 1 1 + 1 P F 2 + max 1 exp 2 r P F A 1 + 3 P F 2 . (15)
‍ ‍
One should note that the model fitting parameters can be predicted separately by fitting the E-UVM to the coreflooding data and adding
them to the code as constant input values. Alternatively, the fitting parameters can also be estimated using the EUREQA correlations (Eqs.
7 through 11). In the latter case, the correlations should be added to the code manually. Finally, the shear factors (‍Z ‍) at the faces were
calculated by dividing the apparent viscosity to zero-shear rate viscosity, as given below:

app F
Z F = . (16)
‍ 0 ‍
It is worth mentioning that with E-UVM, we modified only the shear factors representing polymer non-Newtonian behavior. In contrast,
the multipliers accounting for polymer mixing and permeability reduction were not affected by E-UVM implementation and were calcu-
lated similarly using Todd-Longstaff mixing model (Bao et al. 2017).
Furthermore, it is important to note that our implementation of E-UVM in the simulator accommodates the irreversible drop of viscos-
ity during mechanical degradation. Therefore, viscosities decrease with the shear rate increase, while they stay constant when shear rates
decrease back across the interfaces. One should note that this behavior is applicable only for mechanical degradation region of the pro-
posed model when the shear rates are higher than γ̇cr3.
Verification of the E-UVM against Original Implementation. To verify the E-UVM implementation, shear factors (‍Z )‍ were first
calculated using the E-UVM model by following the procedure mentioned in the previous section. The estimated shear factors were then
used to predict the corresponding ‍msh‍values as follows:
m Z 1
msh = , (17)
‍ m 1 ‍
where ‍Z ‍are the shear factors calculated using E-UVM model and ‍m‍are the multipliers depending on polymer concentration where no
shear effects encountered. ‍m‍ factors are not affected by E-UVM implementation and are estimated separately in MRST. Furthermore,
‍msh‍is the user-prescribed function; ‍msh‍= 1 if the polymer solution does not exhibit shearing and has a constant Newtonian viscosity, ‍msh‍
<1 if the polymer solution exhibits shear-thinning behavior, and ‍msh‍>1 when shear-thickening of polymers occurs. Next, the estimated
‍msh‍values were used to compute the corresponding shear-modified velocities (‍uw, sh), ‍ as shown below:
m uw,0
uw,sh = . (18)
‍ 1 + m 1 msh ‍
The ‍msh‍factor is one of the parameters controlling shear effects in the MRST model (PLYSHEAR model). It is usually tabulated in the
input deck file as a function of shear-modified water velocities. The polymer shear response strongly depends on the change in ‍msh‍values
with water velocities within the table. Therefore, it is possible to compare two rheological models (PLYSHEAR and E-UVM) by modi-
fying the input file table with new ‍msh‍and ‍uw, sh‍values estimated using Eqs. 17 and 18, respectively.
The verification of E-UVM against the original PLYSHEAR model was performed using the core-scale simulation model defined
earlier. In this analysis, a comparison between the two rheological models was performed for around 5 PVs of secondary polymer flood-
ing. Other input parameters for MRST are listed in the Table 10. The core-scale simulation model containing 100 gridblocks had 180
interfaces between those grids. Therefore, the calculation of ‍msh‍ factors in MRST using Eq. 17 revealed a matrix containing 180×1

elements, where each element of the matrix denoted an individual ‍msh‍factor per interface. Similarly, the shear-modified water velocities
were also given in the 180×1 matrix. As the water velocities usually monotonically increase in the original PLYSHEAR table, we filtered
the elements in the velocity matrix from the smallest to the largest with corresponding ‍msh‍ factors. The generated table from modified
‍uw, sh‍and ‍msh‍was then exported to the input file and used as a new table for the PLYSHEAR model.
Average permeability (md) 23.2 Dykstra-Parsons coefficient 0.8

Porosity 0.178 Polymer concentration (ppm) 1,500
Oil density (kg/m3) 820 Oil viscosity (cp) 0.9
Injection rate (mL/min) 0.045 (≈ 1 ft/ Producer BHP (psi) 50
day)
Sorw 0.47 Swirr 0.21
Salinity (g/L) 179.73
Table 10—Input parameters used to verify E-UVM implementation.
After modifying the table in the input file with updated ‍uw, sh‍ and ‍msh‍, polymer flooding simulation was run using the PLYSHEAR
model and compared with the results of E-UVM. Particularly, pressure drop and oil recovery curves obtained by these two models were
compared, as shown in Figs. 14 and 15, respectively. The figures show that the simulation results obtained by PLYSHEAR and E-UVM
for pressure drop and oil recovery are in a good agreement. This confirms the implementation of E-UVM to predict the polymer viscosi-
ties and shear multipliers directly from the correlation. Furthermore, the comparison of average inlet apparent viscosities predicted by the
two models can be found elsewhere (Supplementary Material B, Fig. S-2).
Fig. 14—Comparison of pressure drops for PLYSHEAR and E-UVM models.
Fig. 15—Comparison of oil recoveries for PLYSHEAR and E-UVM models.
Unlike the original model utilized in MRST to establish polymer non-Newtonian behavior (such as the PLYSHEAR model), the imple-
mentation of the E-UVM is independent of input deck files and does not require any initial multipliers tabulated as a function of aqueous

phase velocity according to the particular polymer behavior. The E-UVM can estimate polymer apparent viscosities and corresponding
shear factors at various shear rates by using the fitting parameters in the model. Preferentially, these parameters can be obtained by fitting
the E-UVM to the experimental coreflooding data or calculated using the developed correlations with some minor errors. This implemen-
tation avoids importing the data from external files, extrapolation or interpolation of the data, and iteratively solving for multipliers, which
considerably reduces the simulation time. Therefore, we believe that this method is more practical than the original technique used in
MRST.
Application of E-UVM in MRST. This section describes application of the proposed model by conducting polymer rheological studies
in the MRST reservoir simulator. As was discussed earlier, the E-UVM captures different regimes exhibited by polymers in porous media.
In E-UVM, the Newtonian and shear-thinning flow is governed by the Carreau model, shear-thickening behavior is captured by Deborah
number, and the degradation multiplier controls mechanical degradation of polymers. The model is based on UVM published earlier
(Delshad et al. 2008) that accommodates polymer Newtonian and shear-thinning behavior at low shear rates and shear-thickening at high
rates observed primarily near the wellbores. In addition, E-UVM also characterizes the viscosity drop due to degradation at ultimate shear
rates. In summary, the Carreau model defines only Newtonian and shear-thinning behavior. On the other hand, the UVM also accounts
for the shear-thickening of viscoelastic polymers in porous media. Finally, the E-UVM captures all previously mentioned regimes and
mechanical degradation of polymers.
For polymer rheological studies, viscoelastic polymer flooding was simulated in the tertiary mode of injection using the MRST simu-
lator. The secondary waterflooding was first simulated for 5 PVs at a rate of 0.045 mL/min and obtained about 40% original oil in place
of oil recovery. The remaining oil of 47% was assumed very close to the actual residual oil saturation. Thereafter, the tertiary polymer
flooding was conducted for the consecutive 5 PVs and showed the dependence of displacement efficiency on polymer solution viscosity
and injection rate. For tertiary polymer flooding simulation studies, all three models (Carreau, UVM, and E-UVM) were separately imple-
mented in MRST to distinguish the effect of different rheological regimes on oil recovery. The simulation results for Cp of 2,000 ppm at
0.045 mL/min are depicted in Figs. 16 through 18 for average inlet apparent viscosity, average trapping number, and incremental oil
recovery to polymer flooding, respectively. It is worth mentioning that the trapping numbers (‍NT )‍ were calculated for each gridblock
using Eq. 19:
q
NT = NB2 + N2C , (19)
‍ ˇ ‍ ˇ

ˇk g ˇ
ˇ w w o ˇ
NB = ˇ ˇ, (20)
ˇ ow ˇ
‍ ‍
uw
Nc = , (21)
‍ ow ‍
Fig. 16—Average inlet apparent viscosities for different rheological models at 0.045 mL/min and Cp = 2,000 ppm.

Fig. 17—Average trapping numbers for different rheological models at 0.045 mL/min and Cp = 2,000 ppm.
Fig. 18—Incremental oil recoveries for different rheological models at 0.045 mL/min and Cp = 2,000 ppm.
where ‍NB ‍ and ‍Nc ‍ are Bond and capillary numbers, respectively, ‍kw ‍ is the effective aqueous phase permeability, ‍g ‍ is the gravitational
constant, ‍‍is the apparent viscosity of displacing fluid, ‍ow ‍is the interfacial tension between the aqueous and oleic phases, and ‍w ‍and
‍o‍are the densities of aqueous and oleic phases, respectively. The gravity effects captured by the Bond number were negligible due to the
dimensions of the core-scale model. Additionally, the increment in aqueous phase viscosity using polymers increased the capillary num-
bers and improved displacement efficiency. Furthermore, it is worth mentioning that trapping number-dependent relative permeability
model was used in simulations (Brooks and Corey 1964; Delshad et al. 1986; Delshad 1990).
Moreover, as the onset for shear-thickening and mechanical degradation of polymers is observed at high shear rates, the main differ-
ence between the models arises near the wellbores. Therefore, we averaged the apparent viscosities at the cell interfaces near the injection
well and showed them in Fig. 16. It is worth mentioning that average shear rates in the wellbore vicinity exceeded the third critical shear
rate (γ̇cr3), and mechanical degradation occurred. Accordingly, the Carreau model yielded the lowest viscosities at high shear rates due to
excessive shear-thinning of polymers. In contrast, viscosities estimated by the UVM were the highest as the model considered the viscos-
ity increment due to shear-thickening. Notably, the E-UVM viscosities were lower as opposed to UVM, which can be explained by the
mechanical degradation occurring at ultimate shear rates. The calculated trapping numbers were consistent with the apparent viscosities
(Fig. 17), yielding the highest oil recovery for the UVM model (Fig. 18). On the contrary, the Carreau model did not result in oil mobili-
zation at 0.045 mL/min due to the negligible viscosifying power of polymers. Furthermore, the comparison of the relative injectivity of
the tertiary polymer flooding to secondary waterflooding at 0.045 mL/min is provided in Supplementary Material C (Fig. S-3) to support
the results reported above.
Furthermore, the simulations were repeated for a higher injection rate of 4.5 mL/min, and the results are presented in Figs. 19 and 20
for average inlet apparent viscosity and incremental oil recovery to polymer flooding, respectively. Moreover, the simulation results for

average trapping number and relative injectivity at 4.5 mL/min are provided elsewhere (Supplementary Material C, Fig. S-4 and S-5,
respectively). The increment in injection rate from 0.045 mL/min to 4.5 mL/min significantly increased the shear rates in the simulation
model and exacerbated mechanical degradation. The average inlet apparent viscosities estimated at a given shear rate were correspond-
ingly changed; however, the same trend of results still holds. For this case, the E-UVM viscosities decreased about five times due to
injection rate increment leading to severe mechanical degradation. On the other hand, the UVM viscosities, reaching their maximum
plateau values (‍max ‍) at 0.045 mL/min, were not notably increased with the injection rate (Fig. 19).
Fig. 19—Average inlet apparent viscosities for different rheological models at 4.5 mL/min and Cp = 2,000 ppm.
Fig. 20—Incremental oil recoveries for different rheological models at 4.5 mL/min and Cp = 2,000 ppm.
Nevertheless, an increase in injection rate improved oil mobilization for all models (Fig. 20). Residual oil saturation was reduced from
47% (Sorw) to 42.3, 39.5, and 32% for Carreau, E-UVM, and UVM models, respectively. Although the significant viscosity decrease for
the E-UVM and Carreau models, the incremental oil recoveries increased to 16 and 9%. The latter shows that at a constant injection rate,
the oil recovery is a function of apparent viscosity, and the higher viscosity predicted by the model yields correspondingly high oil recov-
ery. However, the injection rate was found to have a more predominant effect on oil recovery as opposed to apparent viscosities. It was
found that the increment in Darcy’s velocities with injection rate was much higher than the viscosity drop due to mechanical degradation.
This can be explained by the increase in trapping numbers according to Eq. 19, leading to higher residual oil mobilization and conse-
quently, higher incremental oil recovery.
Furthermore, the low injection rate of 0.045 mL/min, which is identical to 1 ft/D, was chosen to represent the flow in porous media.
Shear rates generated in the bulk of the simulation model (away from the injector) were at low ranges, and primarily shear-thinning
behavior was observed at the rate of 0.045 mL/min (shear-thickening took place in the cells close to the wellbores). All three models
predicted the same viscosities in these parts of the simulation model where only shear-thinning was observed. However, the shear rates at
interfaces near injection well were comparatively larger and higher than the critical shear rate for mechanical degradation (‍Pcr3)‍ even at
0.045 mL/min. Therefore, it caused an initial shear-thickening and consecutive drop in viscosity starting from ‍Pcr3‍. The latter caused the
main difference between the models (Carreau, UVM, and E-UVM) in these cells, as UVM did not capture mechanical degradation and
the Carreau model did not accommodate both shear-thickening and mechanical degradation. On the other hand, at the higher injection rate
of 4.5 mL/min, which is identical to 100 ft/D, the shear rates generated in the core-scale simulation model considerably increased, and
primarily shear-thickening were observed in the bulk of the model (the shear rates in some cells remote from the wellbores did not reach
the onset for shear-thickening, and only shear-thinning was observed there). Similarly, the shear rates at the “inflow” interfaces signifi-
cantly increased at 4.5 mL/min, and exacerbated polymer mechanical degradation was observed in those cells. A more extensive discus-
sion can be found in Supplementary Material C (Fig. S-6).
These observations confirmed that using a suitable rheological model is essential in polymer flooding simulation to predict more accu-
rate results. Therefore, for the regions where developed shear rates exceed the critical shear rate for the mechanical degradation, utilizing
UVM may result in overestimated oil recoveries as this model does not captures the viscosity drop. On the other hand, E-UVM accom-
modates all regimes and provides more accurate and representative results.

Proposed Model Limitations. The E-UVM developed in this study can capture the polymer rheology in porous media at various shear
rates and select the proper polymer design for laboratory or field implementations. However, there are several assumptions and limitations
of the proposed model, and embedded correlations are discussed. As E-UVM is used to estimate polymer apparent viscosity at different
shear rates, the only actual variable in this model is the shear rate. The formation properties, such as absolute permeability and porosity,
salinity, temperature as well as other polymer properties, are considered constant during a particular flooding test and do not have direct
control on E-UVM. However, these properties affect polymer behavior in porous media, and their impact on polymer apparent viscosity
is captured in E-UVM through the fitting parameters. Therefore, a preliminary coreflooding test needs to be conducted at the desired
experimental conditions, and the E-UVM fitting parameters, corresponding to the rock and fluid properties used in the experiment, can be
calculated. These fitting parameters can also be estimated using the correlations that we proposed.
The main limitation of the developed correlations is their applicability range. The best accuracy of these correlations was observed for
the ranges of rock and polymer properties with ‍k ‍of 200–2,000 md, ‍Mw ‍of 15 to 20 MDa, ‍S ‍of 10 to 50 g/L, C ‍ p‍of 1,000 to 1,500 ppm,
and ‍ ‍of 0.2 to 0.25. Accordingly, using these correlations for polymer flooding in tight rock and at higher salinities, usually observed in
carbonate rock, might result in inaccurate values. To overcome this issue and expand the application envelope of these correlations, it is
recommended to collect more experimental data as an input for the machine learning technique used. In addition, it will be even more
efficient to divide the collected data into several groups of solution salinity or rock permeability (these are two main parameters with the
most pronounced outliers in the range) and generate correlations for each group with considerably better accuracy. Furthermore, these
correlations relate E-UVM fitting parameters to the basic rock and polymer properties. However, more complex parameters also affect
polymer rheology and have not been considered during the development of correlations, such as molecular weight distribution, specific
monomers existing in a polymer, rock heterogeneity and anisotropy, experimental setup (linear or radial), and presence of oil. We believe
that the correlations can provide more reliable results for more diverse experimental conditions if the recommended improvements are
followed and the mentioned drawbacks are resolved.
As discussed earlier in the paper, the main advantage of the proposed model is its ability to capture the irreversible drop in polymer
viscosity due to mechanical degradation. The onset and degree of mechanical degradation are controlled by two model parameters, which
can be selected based on experimental observations under particular conditions. However, the effect of induced or natural fractures as well
as adsorption and presence of oil in porous media on polymer rheology and particularly on its mechanical degradation has not been con-
sidered in this model. It is also worth mentioning that the mechanical degradation of polymers considered in this study was mainly related
to porous media, while the degradation taking place in flowlines is out of the scope of the paper. It is worth mentioning that the previously
highlighted model limitations will be addressed in our future work.
Conclusions
In this study, a novel viscoelastic model was proposed for predicting polymer apparent viscosity at different shear rates. The proposed
model was validated against experimental data and successfully implemented in a reservoir simulator to highlight its effect on oil recov-
ery. The main findings of this work can be summarized as follows:
• The E-UVM, proposed in this study, captures all four regimes observed during viscoelastic polymer flooding; Newtonian, shear-
thinning, shear-thickening, and mechanical degradation regimes.
• The proposed model parameters were further analyzed; particularly, mechanical degradation constant (‍3)‍ controls the degradation
onset, while the degree of degradation is controlled by the mechanical degradation index (‍n3‍).
• The fitting parameters of the proposed viscoelastic model were related to rock and polymer properties using the symbolic regression
technique. The developed correlations have reasonable accuracy and complexity, which may reduce the need for performing extensive
coreflooding measurements.
• The proposed viscoelastic model was implemented in a reservoir simulator at core scales and verified against the original shear model.
It is believed that using E-UVM for shear factors may be more practical than the original implementation, which imports the data from
external files/tables and iteratively solves for shear multipliers.
• The comparison between Carreau, UVM, and E-UVM models in the reservoir simulator revealed that using the UVM model might
overestimate the oil recovery due to excessive shear-thickening without mechanical degradation evaluation.
Nomenclature
A[F] = face [F] area
C = shear correction factor
Cp = polymer concentration, ppm
d = characteristic time constant for degradation
De = Deborah number
g = gravitational constant
G′, G′′ = storage and loss moduli, respectively, Pa
j = tuning parameter for degradation
k = absolute permeability, D
kw, ko = permeabilities of aqueous and oleic phases, respectively, D
m1 = shear-thinning exponent
m2 = elongational exponent
msh = user-prescribed factor
mμ = concentration dependent multiplier
M = mobility ratio
Mw = polymer molecular weight, MDa
n1 = shear-thinning index
n2 = shear-thickening index
n3 = mechanical degradation index
NB = Bond number
NC = capillary number
NT = trapping number
qw, 0[F] = water flux across the interface [F]

r1 = time constant
r2 = inverse of elongation onset
S = solution salinity, g/L
Sw = water saturation
Swirr = irreducible water saturation
Sorw = residual oil saturation
t = residence time, seconds
uw = Darcy’s velocity, m/s
uw, sh = shear-modified Darcy’s velocity, m/s
uw [F] = Darcy’s velocities (uw) across the interface [F]
x = constant in the viscoelastic models (usually being equal to 2)
Z = shear factor
. −1
‍ =
‍ shear rate, seconds
λ1 = shear-thinning constant
λ2 = shear-thickening constant (usually being equal to 0.01)
λ3 = mechanical degradation constant
λw, λo = mobilities of aqueous and oleic phases, respectively, D/cp
μapp = apparent viscosity, cp
μelastic = elastic viscosity, cp
μmax = maximum elongational viscosity, cp
μshear = shear viscosity, cp
μw, μo = viscosities of aqueous and oleic phases, respectively, cp
μ0, μ∞ = zero-shear rate and infinity-shear rate viscosities, respectively, cp
ρo = density of oleic phase
ρw = density of aqueous phase
σ = standard deviation
σow = interfacial tension between the aqueous and oleic phases
τr = relaxation time, seconds
Φ = porosity
Acknowledgments
The authors wish to acknowledge Abu Dhabi National Oil Company (ADNOC) for funding this research and Khalifa University of Science
and Technology for the support and encouragement. The authors express their appreciation also to MRST developers for providing open-
source software that can be efficiently used for research purposes.
References
Abas, N., Kalair, A., and Khan, N. 2015. Review of Fossil Fuels and Future Energy Technologies. Futures 69: 31–49. https://doi.org/10.1016/j.futures.
2015.03.003.
Abdelfatah, E. R., Kang, K., Pournik, M. et al. 2017. Study of Nanoparticle Adsorption and Release in Porous Media Based on the DLVO Theory. Paper
presented at the SPE Latin America and Caribbean Petroleum Engineering Conference, Buenos Aires, Argentina, 17–19 May. SPE-185484-MS. https://
doi.org/10.2118/185484-MS.
Adegbite, J. O., Al-Shalabi, E. W., and Ghosh, B. 2018. Geochemical Modeling of Engineered Water Injection Effect on Oil Recovery from Carbonate
Cores. J Pet Sci Eng 170: 696–711. https://doi.org/10.1016/j.petrol.2018.06.079.
Akstinat, M. H. 1980. Polymers For Enhanced Oil Recovery In Reservoirs Of Extremely High Salinities And High Temperatures. Paper presented at the
SPE Oilfield and Geothermal Chemistry Symposium, Stanford, California, USA, 28–30 May. SPE-8979-MS. https://doi.org/10.2118/8979-MS.
Aladasani, A. and Bai, B. 2010. Recent Developments and Updated Screening Criteria of Enhanced Oil Recovery Techniques. Paper presented at the
International Oil and Gas Conference and Exhibition in China, Beijing, China, 8–10 June. SPE-130726-MS. https://doi.org/10.2118/130726-MS.
Alfazazi, U., Chacko Thomas, N., Al-Shalabi, E. W. et al. 2021. Investigation of Oil Presence and Wettability Restoration Effects on Sulfonated Polymer
Retention in Carbonates Under Harsh Conditions. Paper presented at the Abu Dhabi International Petroleum Exhibition & Conference, Abu Dhabi,
UAE, 15–18 November. SPE-207892-MS. https://doi.org/10.2118/207892-MS.
Alfazazi, U., Thomas, N. C., AlAmeri, W. et al. 2019. An Experimental Investigation of Polymer Performance in Harsh Carbonate Reservoir Conditions.
Paper presented at the SPE Gas & Oil Technology Showcase and Conference, Dubai, UAE, 21–23 October. SPE-198607-MS. https://doi.org/10.2118/
198607-MS.
Ansari, A., Haroun, M., Rahman, M. M. et al. 2015. Electrokinetic Driven Low-Concentration Acid Improved Oil Recovery in Abu Dhabi Tight Carbonate
Reservoirs. Electrochim Acta 181: 255–270. https://doi.org/10.1016/j.electacta.2015.04.174.
Asen, S. M., Stavland, A., Strand, D. et al. 2019. An Experimental Investigation of Polymer Mechanical Degradation at the Centimeter and Meter Scale.
SPE J. 24 (4): 1700–1713. SPE-190225-PA. https://doi.org/10.2118/190225-PA.
Azad, M. S. and Sultan, A. S. 2014. Extending the Applicability of Chemical EOR in High Salinity, High Temperature & Fractured Carbonate Reservoir
Through Viscoelastic Surfactants. Paper presented at the SPE Saudi Arabia Section Technical Symposium and Exhibition, Al-Khobar, Saudi Arabia,
21–24 April. SPE-172188-MS. https://doi.org/10.2118/172188-MS.
Azad, M. S. and Trivedi, J. J. 2018. Novel Viscoelastic Model for Predicting the Synthetic Polymer’s Viscoelastic Behavior in Porous Media Using Direct
Extensional Rheological Measurements. Fuel 235: 218–226. https://doi.org/10.1016/j.fuel.2018.06.030.
Azad, M. S. and Trivedi, J. J. 2019. Quantification of the Viscoelastic Effects During Polymer Flooding: A Critical Review. SPE J. 24 (6): 2731–2757.
SPE-195687-PA. https://doi.org/10.2118/195687-PA.
Azad, M. S. and Trivedi, J. J. 2021. Quantification of Sor Reduction during Polymer Flooding Using Extensional Capillary Number. SPE J. 26 (3):
1469–1498. SPE-204212-PA. https://doi.org/10.2118/204212-PA.
Bao, K., Lie, K. A., Møyner, O. et al. 2017. Fully Implicit Simulation of Polymer Flooding with MRST. Comput Geosci 21 (5–6): 1219–1244. https://doi.
org/10.1007/s10596-017-9624-5.
BP. 2014. Statistical Review of World Energy. Annual Report 63. London, United Kingdom, BP plc.

Brakstad, K. and Rosenkilde, C. 2016. Modelling Viscosity and Mechanical Degradation of Polyacrylamide Solutions in Porous Media. Paper presented at
the SPE Improved Oil Recovery Conference, Tulsa, Oklahoma, USA, 11–13 April. SPE-179593-MS. https://doi.org/10.2118/179593-MS.
Brooks, A. N. and Corey, A. T. 1964. Hydraulic Properties of Porous Media. In Hydrology Papers. Fort Collins, Colorado, USA: Colorado State University.
Cannella, W. J., Huh, C., and Seright, R. S. 1988. Prediction of Xanthan Rheology in Porous Media. Paper presented at the SPE Annual Technical
Conference and Exhibition, Houston, Texas, USA, 2–5 October. SPE-18089-MS. https://doi.org/10.2118/18089-MS.
Chandrasekhar, S. and Mohanty, K. K. 2013. Wettability Alteration with Brine Composition in High Temperature Carbonate Reservoirs. Paper presented
at the SPE Annual Technical Conference and Exhibition, New Orleans, Louisiana, USA, 30 September–2 October. SPE-166280-MS. https://doi.org/
10.2118/166280-MS.
Choi, U. S. and Kasza, K. E. 1989. Long Term Degradation of Dilute Polyacrylamide Solutions in Turbulent Flow. In Drag Reduction in Fluid Flows,
163–169. Davos, Switzerland: Argonne National Laboratory.
Clarke, A., Howe, A. M., Mitchell, J. et al. 2016. How Viscoelastic-Polymer Flooding Enhances Displacement Efficiency. SPE J. 21 (3): 0675–0687. SPE-
174654-PA. https://doi.org/10.2118/174654-PA.
Clemens, T., Deckers, M., Kornberger, M. et al. 2013. Polymer Solution Injection – Near Wellbore Dynamics and Displacement Efficiency, Pilot Test
Results, Matzen Field, Austria. Paper presented at the EAGE Annual Conference & Exhibition Incorporating SPE Europec, London, UK, 10–13 June.
SPE-164904-MS. https://doi.org/10.2118/164904-MS.
Davison, P. and Mentzer, E. 1982. Polymer Flooding in North Sea Reservoirs. SPE J. 22 (3): 353–362. SPE-9300-PA. https://doi.org/10.2118/9300-PA.
Delshad, M. 1990. Trapping of Micellar Fluids in Berea Sandstone. PhD dissertation, The University of Texas at Austin, Texas, USA.
Delshad, M., Bhuyan, D., Pope, G. A. et al. 1986. Effect of Capillary Number on the Residual Saturation of a Three-Phase Micellar Solution. Paper
presented at the SPE Enhanced Oil Recovery Symposium, Tulsa, Oklahoma, USA, 20–23 April. SPE-14911-MS. https://doi.org/10.2118/14911-
MS.
Delshad, M., Kim, D. H., Magbagbeola, O. A. et al. 2008. Mechanistic Interpretation and Utilization of Viscoelastic Behavior of Polymer Solutions for
Improved Polymer-Flood Efficiency. Paper presented at the SPE Symposium on Improved Oil Recovery, Tulsa, Oklahoma, USA, 20–23 April. SPE-
113620-MS. https://doi.org/10.2118/113620-MS.
Diab, W. N. and Al-Shalabi, E. W. 2019. Recent Developments in Polymer Flooding for Carbonate Reservoirs under Harsh Conditions. Paper presented at
the Offshore Technology Conference Brasil, Rio de Janeiro, Brazil, 29–31 October. OTC-29739-MS. https://doi.org/10.4043/29739-MS.
Dykstra, H. and Parsons, R. L. 1950. The Prediction of Oil Recovery by Water Flooding, Secondary Recovery of Oil in the United States, Second Edition,
160–174. Washington, DC, USA: American Petroleum Institute.
Garrouch, A. A. 1999. A Viscoelastic Model for Polymer Flow in Reservoir Rocks. Paper presented at the SPE Asia Pacific Oil and Gas Conference and
Exhibition, Jakarta, Indonesia, 20–22 April. SPE-54379-MS. https://doi.org/10.2118/54379-MS.
Gbadamosi, A. O., Junin, R., Manan, M. A. et al. 2019. An Overview of Chemical Enhanced Oil Recovery: Recent Advances and Prospects. Int Nano Lett
9 (3): 171–202. https://doi.org/10.1007/s40089-019-0272-8.
Guo, H. 2017. How to Select Polymer Molecular Weight and Concentration to Avoid Blocking in Polymer Flooding? Paper presented at the SPE
Symposium: Production Enhancement and Cost Optimisation, Kuala Lumpur, Malaysia, 7–8 November. SPE-189255-MS. https://doi.org/10.2118/
189255-MS.
Hamouda, A. A. and Moshood, O. T. 2007. Effect of Temperature on Degradation of Polymer Drag Reduction and Heat Transfer in Non-Newtonian Fluid.
Paper presented at the International Symposium on Oilfield Chemistry, Houston, Texas, USA, 28 February–2 March. SPE-106081-MS. https://doi.org/
10.2118/106081-MS.
Hashmet, M. R., Qaiser, Y., Mathew, E. S. et al. 2017. Injection of Polymer for Improved Sweep Efficiency in High Temperature High Salinity Carbonate
Reservoirs: Linear X-Ray Aided Flood Front Monitoring. Paper presented at the SPE Kingdom of Saudi Arabia Annual Technical Symposium and
Exhibition, Dammam, Saudi Arabia, 24–27 April. SPE-188125-MS. https://doi.org/10.2118/188125-MS.
Heemskerk, J., Rosmalen, R., Janssen-van, R. et al. 1984. Quantification of Viscoelastic Effects of Polyacrylamide Solutions. Paper presented at the SPE
Enhanced Oil Recovery Symposium, Tulsa, Oklahoma, USA, 15–18 April. SPE-12652-MS. https://doi.org/10.2118/12652-MS.
Hester, R. D., Flesher, L. M., and McCormick, C. L. 1994. Polymer Solution Extension Viscosity Effects During Reservoir Flooding. Paper presented at the
SPE/DOE Improved Oil Recovery Symposium, Tulsa, Oklahoma, USA, 17–20 April. SPE-27823-MS. https://doi.org/10.2118/27823-MS.
Hincapie, R. E., Rock, A., Wegner, J. et al. 2017. Oil Mobilization by Viscoelastic Flow Instabilities Effects during Polymer EOR: A Pore-Scale Visualization
Approach. Paper presented at the SPE Latin America and Caribbean Petroleum Engineering Conference, Buenos Aires, Argentina, 17–19 May. SPE-
IEA. 2020. Oil Market Report. Internal Energy Agency, Paris, ‎France.
Informer Technologies. 2021. https://eureqa.software.informer.com/0.9.
Jensen, T., Kadhum, M., Kozlowicz, B. et al. 2018. Chemical EOR Under Harsh Conditions: Scleroglucan As A Viable Commercial Solution. Paper
presented at the SPE Improved Oil Recovery Conference, Tulsa, Oklahoma, USA, 14–18 April. SPE-190216-MS. https://doi.org/10.2118/190216-MS.
Jiang, B., Keffer, D. J., Edwards, B. J. et al. 2003. Modeling Shear Thickening in Dilute Polymer Solutions: Temperature, Concentration, and Molecular
Weight Dependencies. J Appl Polym Sci 90 (11): 2997–3011. https://doi.org/10.1002/app.12950.
Khurshid, I., Al-Shalabi, E. W., and Alameri, W. 2020. Influence of Water Composition on Formation Damage and Related Oil Recovery in Carbonates: A
Geochemical Study. J Pet Sci Eng 195: 107715. https://doi.org/10.1016/j.petrol.2020.107715.
Kim, D. H., Lee, S., Ahn, C. H. et al. 2010. Development of a Viscoelastic Property Database for EOR Polymers. Paper presented at the SPE Improved Oil
Recovery Symposium, Tulsa, Oklahoma, USA, 24–28 April. SPE-129971-MS. https://doi.org/10.2118/129971-MS.
Lake, L. W. 1989. Enhanced Oil Recovery. Englewood Cliffs, New Jersey, USA: Prentice Hall.
Lie, K. A. 2016. An Introduction to Reservoir Simulation Using MATLAB: User Guide for the MATLAB Reservoir Simulation Toolbox (MRST). Sor-
Trondelag, Norway: SINTEF ICT.
Lohne, A., Nødland, O., Stavland, A. et al. 2017. A Model for Non-Newtonian Flow in Porous Media at Different Flow Regimes. Comput Geosci 21 (5–6):
1289–1312. https://doi.org/10.1007/s10596-017-9692-6.
Lotfollahi, M., Farajzadeh, R., Delshad, M. et al. 2016. Mechanistic Simulation of Polymer Injectivity in Field Tests. SPE J. 21 (4): 1178–1191. SPE-
174665-PA. https://doi.org/10.2118/174665-PA.
Macosko, K. W. 1994. Rheology: Principles, Measurements and Applications. New York, New York, USA: Wiley VCH.
Maerker, J. M. 1975. Shear Degradation of Partially Hydrolyzed Polyacrylamide Solutions. SPE J. 15 (4): 311–322. SPE-5101-PA. https://doi.org/10.
2118/5101-PA.
Magbagbeola, O. A. 2008. Quantification of the Viscoelastic Behavior of High Molecular Weight Polymers Used for Chemical Enhanced Oil Recovery.
MS thesis, The University of Texas at Austin, Texas, USA.
Manichand, R. N., Moe Soe Let, K. P., Gil, L. et al. 2013. Effective Propagation of HPAM Solutions Through the Tambaredjo Reservoir During a Polymer
Flood. SPE Prod & Oper 28 (4): 358–368. SPE-164121-PA. https://doi.org/10.2118/164121-PA.

Manrique, E., Thomas, C., Ravikiran, R. et al. 2010. EOR: Current Status and Opportunities. Paper presented at the SPE Improved Oil Recovery
Symposium, Tulsa, Oklahoma, USA, 24–28 April. SPE-130113-MS. https://doi.org/10.2118/130113-MS.
Martin, F. D. 1986. Mechanical Degradation of Polyacrylamide Solutions in Core Plugs From Several Carbonate Reservoirs. SPE Form Eval 1 (2):
139–150. SPE-12651-PA. https://doi.org/10.2118/12651-PA.
Masalmeh, S., AlSumaiti, A., Gaillard, N. et al. 2019. Extending Polymer Flooding Towards High-Temperature and High-Salinity Carbonate Reservoirs.
Paper presented at the Abu Dhabi International Petroleum Exhibition & Conference, Abu Dhabi, UAE, 11–14 November. SPE-197647-MS. https://doi.
org/10.2118/197647-MS.
Meng, W., Haroun, M. R., Sarma, H. K. et al. 2015. A Novel Approach of Using Phosphate-Spiked Smart Brines to Alter Wettability in Mixed Oil-Wet
Carbonate Reservoirs. Paper presented at the Abu Dhabi International Petroleum Exhibition and Conference, Abu Dhabi, UAE, 9–12 November. SPE-
Morris, C. W. and Jackson, K. M. 1978. Mechanical Degradation Of Polyacrylamide Solutions In Porous Media. Paper presented at the SPE Symposium
on Improved Methods of Oil Recovery, Tulsa, Oklahoma, USA, 16–17 April. SPE-7064-MS. https://doi.org/10.2118/7064-MS.
Odell, J. A. and Keller, A. 1986. Flow-Induced Chain Fracture of Isolated Linear Macromolecules in Solution. J Polym Sci B Polym Phys 24 (9): 1889–
1916. https://doi.org/10.1002/polb.1986.090240901.
Qi, P. 2018. The Effect of Polymer Viscoelasticity on Residual Oil Saturation. PhD dissertation, The University of Texas at Austin, Texas, USA.
Ranjbar, M., Rupp, J., Pusch, G. et al. 1992. Quantification and Optimization of Viscoelastic Effects of Polymer Solutions for Enhanced Oil Recovery.
Paper presented at the SPE/DOE Enhanced Oil Recovery Symposium, Tulsa, Oklahoma, USA, 22–24 April. SPE-24154-MS. https://doi.org/10.2118/
24154-MS.
Schlumberger. 2013. Eclipse Technical Manual. Schlumberger, Houston, Texas, USA. Version 2013. 2.
Seright, R. S., Seheult, M., and Talashek, T. 2009. Injectivity Characteristics of EOR Polymers. SPE Res Eval & Eng 12 (5): 783–792. SPE-115142-PA.
https://doi.org/10.2118/115142-PA.
Sheng, J. J. 2011. Modern Chemical Enhanced Oil Recovery: Theory and Practice. Houston, Texas, USA: Gulf Professional Publishing.
Sheng, J. J., Leonhardt, B., and Azri, N. 2015. Status of Polymer-Flooding Technology. J Can Pet Technol 54 (2): 116–126. SPE-174541-PA. https://doi.
org/10.2118/174541-PA.
Skauge, T., Djurhuus, K., Zimmermann, T. et al. 2019. Radial Injectivity of an Associative Polymer for EOR. In IOR 2019 – 20th European Symposium
on Improved Oil Recovery, 1–23. Houten, The Netherlands‎: European Association of Geoscientists & Engineers. https://doi.org/10.3997/2214-4609.
201900143.
Skauge, T., Skauge, A., Salmo, I. C. et al. 2016. Radial and Linear Polymer Flow - Influence on Injectivity. Paper presented at the SPE Improved Oil
Recovery Conference, Tulsa, Oklahoma, USA, 11–13 April. SPE-179694-MS. https://doi.org/10.2118/179694-MS.
Skauge, A., Zamani, N., Jacobsen, J. G. et al. 2018. Polymer Flow in Porous Media. Colloids Interfaces 2 (3): 1–27. https://doi.org/10.3390/colloids2030027.
Sorbie, K. S. 1991. Polymer-Improved Oil Recovery. Glasgow and London: Blackie and Son Ltd. https://doi.org/10.1007/978-94-011-3044-8.
Southwick, J. G. and Manke, C. W. 1988. Molecular Degradation, Injectivity, and Elastic Properties of Polymer Solutions. SPE Res Eng 3 (4): 1193–1201.
SPE-15652-PA. https://doi.org/10.2118/15652-PA.
Standnes, D. C. and Skjevrak, I. 2014. Literature Review of Implemented Polymer Field Projects. J Pet Sci Eng 122: 761–775. https://doi.org/10.1016/j.
petrol.2014.08.024.
Stavland, A., Jonsbråten, H. C., Lohne, A. et al. 2010. Polymer Flooding – Flow Properties in Porous Media Versus Rheological Parameters. Paper
presented at the SPE EUROPEC/EAGE Annual Conference and Exhibition, Barcelona, Spain, 14–17 June. SPE-131103-MS. https://doi.org/10.2118/
131103-MS.
Tahir, M., Hincapie, R. E., Be, M. et al. 2017. Experimental Evaluation of Polymer Viscoelasticity During Flow in Porous Media: Elongational and Shear
Analysis. Paper presented at the SPE Europec Featured at 79th EAGE Conference and Exhibition, Paris, France, 12–15 June. SPE-185823-MS. https://
doi.org/10.3997/2214-4609.201701589.
Taylor, K. C. and Nasr-El-Din, H. A. 1995. Water-Soluble Hydrophobically Associating Polymers for Improved Oil Recovery: A Literature Review. Paper
presented at the SPE International Symposium on Oilfield Chemistry, San Antonio, Texas, USA, 14–17 February. SPE-29008-MS. https://doi.org/10.
2118/29008-MS.
Walter, A. V., Jimenez, L. N., Dinic, J. et al. 2019. Effect of Salt Valency and Concentration on Shear and Extensional Rheology of Aqueous Polyelectrolyte
Solutions for Enhanced Oil Recovery. Rheol Acta 58 (3–4): 145–157. https://doi.org/10.1007/s00397-019-01130-6.
Wang, D., Cheng, J., Yang, Q. et al. 2000. Viscous-Elastic Polymer Can Increase Microscale Displacement Efficiency in Cores. Paper presented at the SPE
Annual Technical Conference and Exhibition, Dallas, Texas, USA, 1–4 October. SPE-63227-MS. https://doi.org/10.2118/63227-MS.
Wang, D., Han, P., Shao, Z. et al. 2008. Sweep-Improvement Options for the Daqing Oil Field. SPE Res Eval & Eng 11 (1): 18–26. SPE-99441-PA. https://
doi.org/10.2118/99441-PA.
Wilton, R. R. and Torabi, F. 2013. Rheological Assessment of the Elasticity of Polymers for Enhanced Heavy Oil Recovery. Paper presented at the SPE
Heavy Oil Conference-Canada, Calgary, Alberta, Canada, 11–13 June. SPE-165488-MS. https://doi.org/10.2118/165488-MS.
Wreath, D., Pope, G. A., and Sepehrnoori, K. 1990. Dependence of Polymer Apparent Viscosity on the Permeable Media and Flow Conditions. In Situ 14
(3): 263–283.

Spe 206010 Pa

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Spe 206010 Pa

Uploaded by

Copyright:

Available Formats

An Extended Unified Viscoelastic Model

for Predicting Polymer Apparent Viscosity

*Corresponding author; email: mursalzeynally@gmail.com

February 2023 SPE Reservoir Evaluation & Engineering 99

Polymer Viscoelasticity and Mechanical Degradation

100 February 2023 SPE Reservoir Evaluation & Engineering

Effect on Mechanical Degradation

Table 2—Factors affecting polymer mechanical degradation.

February 2023 SPE Reservoir Evaluation & Engineering 101

Table 2 (continued)—Factors affecting polymer mechanical degradation.

102 February 2023 SPE Reservoir Evaluation & Engineering

February 2023 SPE Reservoir Evaluation & Engineering 103

1 Southwick Exp1 20 2.5 0.048 NA 100

4 BV1 Flopaam 3,630 S 20 20.04 (SFB) 20.32 0.647 0.23

5 BV2 AN-­125 8 20.04 (SFB) 20.32 0.578 0.24

8 Exp1 Flopaam 3,830 S 20 3 14.5 0.573

12 Manichand Exp 1 18 1 Reservoir NA 4

14 Brakstad and 20 9.27 1.26

15 Skauge et al. Exp1 18 28.95 2.34 0.228 linear

17 Lohne et al. Exp1 15 7 0.1613 0.177

20 Skauge et al. Exp1 Hydrophobically NA 29.1 2.433 0.235 linear

22 C-­2 8 4.15 1.616 0.26 100

23 Masalmeh C-­4 Reservoir 5.64 0.107 59 (Sor =

24 C-­6 6.62 0.121 0.304 100

25 CF-­1 8 0.5 7.315 0.19852

27 CF-­2 0.5 Indiana 7.416 0.16513

104 February 2023 SPE Reservoir Evaluation & Engineering

February 2023 SPE Reservoir Evaluation & Engineering 105

106 February 2023 SPE Reservoir Evaluation & Engineering

February 2023 SPE Reservoir Evaluation & Engineering 107

108 February 2023 SPE Reservoir Evaluation & Engineering

February 2023 SPE Reservoir Evaluation & Engineering 109

110 February 2023 SPE Reservoir Evaluation & Engineering

Pore volume (cm3) 14.6 Length (cm) 7.8

Table 5—Core petrophysical properties (Chandrasekhar and Mohanty

Viscosity at 120°C (cp) 0.9

Table 6—Oil properties

Ion\brine Concentration (ppm)

Table 7—Water properties

February 2023 SPE Reservoir Evaluation & Engineering 111

Table 8—Polymer properties used in polymer

Table 9—Core-­scale simulation model data.

Fig. 13—Core-­scale simulation model with heterogeneous permeability.

112 February 2023 SPE Reservoir Evaluation & Engineering

February 2023 SPE Reservoir Evaluation & Engineering 113

Average permeability (md) 23.2 Dykstra-­Parsons coefficient 0.8

Fig. 14—Comparison of pressure drops for PLYSHEAR and E-­UVM models.

Fig. 15—Comparison of oil recoveries for PLYSHEAR and E-­UVM models.

114 February 2023 SPE Reservoir Evaluation & Engineering

February 2023 SPE Reservoir Evaluation & Engineering 115

116 February 2023 SPE Reservoir Evaluation & Engineering

February 2023 SPE Reservoir Evaluation & Engineering 117

118 February 2023 SPE Reservoir Evaluation & Engineering

February 2023 SPE Reservoir Evaluation & Engineering 119

120 February 2023 SPE Reservoir Evaluation & Engineering

February 2023 SPE Reservoir Evaluation & Engineering 121

You might also like

5 BV2 AN-125 8 20.04 (SFB) 20.32 0.578 0.24

22 C-2 8 4.15 1.616 0.26 100

23 Masalmeh C-4 Reservoir 5.64 0.107 59 (Sor =

24 C-6 6.62 0.121 0.304 100

25 CF-1 8 0.5 7.315 0.19852

27 CF-2 0.5 Indiana 7.416 0.16513

Table 9—Core-scale simulation model data.

Fig. 13—Core-scale simulation model with heterogeneous permeability.

Average permeability (md) 23.2 Dykstra-Parsons coefficient 0.8

Fig. 14—Comparison of pressure drops for PLYSHEAR and E-UVM models.

Fig. 15—Comparison of oil recoveries for PLYSHEAR and E-UVM models.