226 (2023) 211753

Contents lists available at ScienceDirect

Geoenergy Science and Engineering

journal homepage: www.sciencedirect.com/journal/geoenergy-science-and-engineering

Nanoparticles assisted polymer flooding: Comprehensive assessment and

empirical correlation
Khalaf G. Salem a, *, Mahmoud A. Tantawy b, Ahmed A. Gawish b, Sayed Gomaa c,
A.N. El-hoshoudy d, e, **
a
Department of Reservoir Engineering, South Valley Egyptian Petroleum Holding Company (GANOPE), Cairo, Egypt
b
Department of Petroleum Engineering, Faculty of Petroleum and Mining Engineering, Suez University, Suez 11252, Egypt
c
Mining and Petroleum Engineering Department, Faculty of Engineering, Al-Azhar University, Cairo 11837, Egypt
d
PVT Lab, Production Department, Egyptian Petroleum Research Institute, 11727, Cairo, Egypt
e
PVT Service Center, Egyptian Petroleum Research Institute, 11727, Cairo, Egypt

A R T I C L E I N F O A B S T R A C T

Keywords: Conventional Polymer flooding is considered the most promising chemical-enhanced oil recovery by reducing
NPs assisted polymer flooding water mobility, and hence increasing the sweep efficiency. However, there are certain drawbacks associated with
Polymeric nanofluid polymer flooding as the degradation of its molecular structure in severe reservoir environments. Recently, it is
Enhanced oil recovery
experimentally adopted that attachment of nanoparticles (NPs) toward polymer macromolecule may ameliorate
Statistical analysis
And artificial neural network (ANN)
the flooding performance and hence increase the oil recovery further than conventional polymer flooding. The
potential of NPs assisted polymer flooding has been extensively studied in porous media through core flooding.
The target of this work is to statistically scrutinize the previous core flooding experiments and investigate the
parameters that affect the incremental oil recovery by NPs assisted polymer flooding. To accomplish this target, a
robust investigation of the outcome of over 350 core-flood tests from the literature was gathered to investigate
the impacts of rock-oil properties, flooding conditions, polymer, and NPs characteristics on incremental oil re­
covery. Additionally, Fractional factorial design (FFD) was used to analyze core flood results and screen the
effects of reservoir rock, fluid, polymer, and NPs properties on the performance of assisted polymer flooding.
Furthermore, an empirical model was built using an Artificial neural network (ANN) to predict the incremental
oil recovery for polymeric nanofluid. The results of the investigation indicated that NPs assisted polymer flooding
with optimum concentration of both polymer and NPs may result in an incremental recovery of 18% of oil-in-
place. Furthermore, the results of FFD revealed that NPs concentration, polymer concentration, and rock
permeability are the most significant parameters of oil recovery. The results outcome showed that there is an
optimal concentration of both polymer and NPs assisted polymer. In addition, the results revealed that core-
flooded viscous oil needs a highly permeable rock to guarantee successful flooding. The coefficient of determi­
nation (R2) between the real and calculated incremental oil from the ANN model was established to be 0.953 and
0.952 with an error of 5.6 and 8.7% respectively for the training and testing approaches. Such statistical in­
vestigations and ANN approaches afford new insights and guidelines for preliminary assessment, designing, and
execution of Nanohybrid polymer upcoming projects on a field scale.

recovery or synonymously known as enhanced oil recovery (EOR)

1. Introduction
Practically, it is well-established that great quantities (~60%) of
original oil-in-place (OOIP) remain in reservoirs after both primary and
secondary recovery stages (Muggeridge et al., 2014). To meet the de­
mand for world energy and maximize the recovery factor (RF), tertiary
techniques can be considered and utilized (Alvarado and Manrique,
2010; Lake, 1989; Muggeridge et al., 2014). Among the most mature,
effective, and successful chemical-enhanced oil recovery (CEOR)
methods are polymer flooding (Corredor et al., 2019d; El-hoshoudy,
2022; Ruiz-Cañas et al., 2020; Tavakkoli et al., 2022), which involves
the injection of polymer solution into the reservoir with a suitable

* Corresponding author.
** Corresponding author.PVT Lab, Production Department, Egyptian Petroleum Research Institute, 11727, Cairo, Egypt
E-mail address: Khalaf.SaIb@pme.suezuni.edu.eg (K.G. Salem).

https://doi.org/10.1016/j.geoen.2023.211753
Received 28 December 2022; Received in revised form 21 March 2023; Accepted 30 March 2023
Available online 5 April 2023
2949-8910/© 2023 Elsevier B.V. All rights reserved.
K.G. Salem et al.
Abbreviations
OOIP Original oil in place
RF Recovery factor
EOR Enhanced Oil Recovery
CEOR Chemical Enhanced Oil Recovery
M mobility ratio
NPs Nanoparticles
IFT Interfacial tension
SNP Silica Nanoparticles
SiO2 Silicon Oxide
Wt. % weight percent
C Celsius
HPAM Hydrolyzed polyacrylamide
PAM Polyacrylamide
ANN Artificial neural network
Al2O3 Aluminum Oxide
k permeability
mD Milli Darcy
cp centipoise
PV Pore volume
concentration to diminish the mobility ratio of the water-oil system (M)
as much as possible, and hence increase the sweep efficiency (Firozjaii
and Saghafi, 2020; Gbadamosi et al., 2019c; Lake, 1989; Sheng et al.,
2015; Standnes and Skjevrak, 2014). In addition, the combination of
polymer and surfactant resulted in improved oil recovery (Gbadamosi
et al., 2019c; Hakiki et al., 2015). Although polymer flooding provides
better results, it also has certain drawbacks, especially under severe
reservoir circumstances of high temperature and salinity. The degra­
dation and retention/adsorption of the polymer are some of the con­
straints affecting the performance of polymer flooding (El-hoshoudy,
2022; Gbadamosi et al., 2022a; Gbadamosi et al., 2019c; Sheng et al.,
2015). Some literature discusses the effect of hydrophobic monomers on
the polymer structure. In this regard, El-Hoshoudy et al. (2017); Liu
et al. (2017b); and Zhu et al. (2014) stated that the introduction of
hydrophobic groups onto the backbone of conventional polymers led to
an improvement in the stability of polymeric solutions. Hence, hydro­
phobically associative polyacrylamide, or what is abbreviated as
HAPAM, has improved characteristics in terms of thermal stability, salt
tolerance, and rheology modifier. However, pilot investigations have
shown that HAPAM has some problems in terms of injectivity and sol­
ubility (El-Hoshoudy et al., 2017; Liu et al., 2017b; Zhu et al., 2014).
Researchers have proved that the optimal grafting of nanoparticles
(NPs) into polymer structure can guide to boost the performance of
polymer characteristics such as rheological behavior, thermal endur­
ance, and chemical resistance further than the conventional polymeric
material (Corredor et al., 2019d; El-hoshoudy, 2022; Gbadamosi et al.,
2019c; Ruiz-Cañas et al., 2020). Furthermore, the synergy of nano­
particles and polymer resulted in interfacial tension (IFT) reduction (Cao
et al., 2018a; Fan et al., 2021; Maurya et al., 2017; Saha et al., 2018),
wettability alteration (El-Hoshoudy et al., 2016; Hu et al., 2021; Liu
et al., 2022; Maurya et al., 2017; Saha et al., 2018), and improvement of
polymer viscoelasticity (Aliabadian et al., 2020; CHEN et al., 2020;
Haruna et al., 2020), which led to boost the incremental oil recovery.
Thus, the developed new class of enhanced oil recovery is referred to as
nanoparticle-assisted polymer flooding or polymeric nanofluid (Agi
et al., 2020a; Agi et al., 2020b; El-hoshoudy, 2022; Gbadamosi et al.,
2022b; Gbadamosi et al., 2018; Saha et al., 2018).
Most of the publications about the polymeric nanofluid solution re­
ported viscosity modifications of the polymer solution. Liu J et al. (Liu
et al., 2022) developed nano-hybrid polymer (polymer-SNP) through
in-situ polymerization. The authors found that polymer-SNP held
2
Geoenergy Science and Engineering 226 (2023) 211753
wt.% Weight percent
WF Waterflooding
CSNF crystalline starch nanofluid
F Fahrenheit
XG xanthan gum
TiO2 Titanium oxide
FFD Fractional factorial design
ANOVA Analysis of variance
DF Degrees of Freedom
Seq SS The sequential sum of squares
Adj SS The adjusted sum of squares
Adj MS The adjusted mean of squares
AAPE absolute average error
MSE mean squared error
RMSE root mean squared error
R2 coefficient of determination
W1i Weights between the input layer and hidden layer
W2i weights between the hidden layer and output layer
b1i biases for neurons of hidden layer
b2 the bias of the output layer
superior thickening efficiency at the analogous polymer concentration.
Additionally, silica nanoparticles were homogeneously filled into the
polymer configuration and achieved an enhancement of polymer vis­
cosity. Furthermore, they investigated the effect of high temperature on
polymer behavior. Moreover, the results revealed that the viscosity of
polymeric nano solution caught the best temperature endurance, and its
values were twice that of the traditional polymer solution. Nanjun Lai
et al. (2019) synthesized an innovative nano-SiO2 polymer by free
radical polymerization. In addition, they observed that the viscosity of
SiO2-nanocomposite showed extreme temperature endurance, salt
tolerance, and mechanical shear resistance when compared with hy­
drolyzed polyacrylamide. In the case of pure polymer without NPs such
as HPAM, the elevated temperature can hasten the thermal degradation
amongst the polymer elements because of the one-dimensionality of the
polymer structure. Similarly, Hu et al. (2017) developed hydrolyzed
polyacrylamide-based SiO2 NPs as a hybrid polymer EOR. They exam­
ined rheological characteristics at various temperatures and salinities,
where the enclosure of silica NPs implicitly increased the polymer vis­
cosity in elevated environments of salinity and temperature. Addition­
ally, the hybrid polymer solution demonstrated thermal stability
compared with conventional polymer. Such experimental findings
proved that incorporation of NPs into polymer matrix led to enhance­
ment of polymer solution viscosity.
Jie Cao et al. (2018a) prepared an amino-functionalized nano-silica
polymeric system and investigated IFT reduction using the pendant drop
technique. They found that the NPs significantly diminish the IFT be­
tween the crude oil and polymer solution. Moreover, the results from
their experiment showed that no reduction in IFT could be obtained
when the NPs concentrations exceed the critical value. The critical
concentrations of NPs were 0.1 wt% and 0.05 wt% for polymer nano
silica and polymer amino nano silica respectively. El-hoshoudy et al.
(El-Hoshoudy et al., 2016) prepared a nanocomposite polymer
(HPAM-SiO2) where silica was introduced through the polymer’s
backbone configuration. They investigated the effects of nanocomposite
concentration on IFT reduction between oil and brine at ambient tem­
perature. The results showed that the nanocomposite polymer can
decline the IFT to a lower value than those between the oil and brine
solution. IFT reduction was attributed to the presence of amphiphilic
structure which promotes intermolecular association, and the presence
of hydrophobic groups in a polymer backbone structure, which con­
tributes to making aggregates or micelle to minimize their exposure to
K.G. Salem et al.
Fig. 1. Framework of research methodology.
Table 1
Effective parameters collected in the Dataset.
Type of acquired data Characteristics
Core properties Core type, geometrics, porosity, permeability,
pore volume, and initial saturations.
Crude oil properties Oil viscosity, oil gravity
Flood conditions and Salinity, temperature, injection rate, and slug size
injection scheme
Polymer and Nanoparticles Polymer type, polymer concentration, NPs type,
characteristics and NPs concentration.
Results of core floods Water flood oil recovery, total recovery, and
incremental oil recovery.
water and consequently leads to IFT reduction.
Additionally, it was reported that the nanoparticles-assisted polymer
flooding alters wettability in porous media owing to the inclusion of NPs
in the polymer structure. Guangli Fan et al. (2021) investigated the
addition of nano-silica for water and polymer solutions to reduce contact
angle and alter the rock wettability to water-wet. Furthermore, they
found that the contact angle had dramatically reduced in low concen­
trations of silica. It was observed that the optimum concentration of
nano silica was 3000 mg/L due to no change observed in the contact
angle over this dose. Contact angles were measured by Rahul Saha et al.
(2018) to address wettability alteration induced by silica nanoparticles
loaded on xanthan Gum polymer. The results demonstrated that the
contact angle decreased with the increase of NPs concentration and
wettability converted from oil-wet to strongly water-wet as well. The
change in the contact angle (86.2–51.2◦ ) was not significantly observed
at 0 wt% NPs whereas the highest change of contact angle (50.4–18.84◦ )
was observed at 0.5 wt% NPs concentration. The authors stated that the
electrostatic repulsion influenced within the NPs was strong at greater
NP concentrations, causing the particles to spread along the solid surface
and lowering the contact angle.
Numerous core flood experiments on the application of polymeric
nanofluid for purpose of enhanced oil recovery have appeared to be
feasible in harsh reservoir conditions. Jie Cao et al. (Cao et al., 2019)
performed core flooding with polymer-nano silica in artificial cores. The
hybrid polymer exhibited around 17% of OOIP as incremental recovery
after a water flood. The authors stated that the performance of the
nano-silica/polymer solution could be attributed to the improved
rheological properties and the interfacial activity of nano-silica as well.
Ponnapati et al. (Ponnapati et al., 2011) performed displacement tests
with polymer-nano silica in sandstone. The Polymer nanohybrid could
improve the oil recovery and yield around 8% of OOIP as incremental
3
Geoenergy Science and Engineering 226 (2023) 211753
recovery. The authors hypothesized that Polymer nanohybrid would
block the small pores and increase the pressure drop to move the trapped
oil into larger pores. Moreover, Oil displacement tests of Al2O3 poly­
meric nanofluid were performed by Afeez O. Gbadamosi et al. (2019a) in
a sandstone core at 90 ◦ C. The obtained results were compared to those
of SiO2-polymer and hydrolyzed polyacrylamide (HPAM without NPs).
HPAM, SiO2-polymer, and Al2O3-polymer achieved a cumulative oil
recovery of 54.7%, 60.81%, and 65.3% OOIP respectively. These results
indicated that Al2O3-polymer had an additional oil recovery of 10.6%
more than traditional HPAM. Achinta Bera et al. (2020) point out that
the addition of Silica NPs to Guar Gum biopolymer provided an addi­
tional oil recovery of 44.3% of OOIP. The maximum oil recovery for the
Nano-polymer solution was attributed to wettability alteration and
improved rheological properties as well. The above overview indicates
that the performance of nanohybrid polymer flooding can yield incre­
mental oil recovery further than conventional polymer flooding.
Despite the extremely successful laboratory experiments for nano-
polymer flooding, there is a lack of application of this new technique
on a field scale. Consequently, this research is an attempt to statistically
investigate the findings of numerous core flood tests from the literature
to acquire guidelines when planning, designing, and implementing up­
coming nanohybrid polymer projects on a field scale. To the best of our
knowledge, the database gathered for this research is so far the most
comprehensive of its kind, and it can be used to further optimization of
these polymeric nanofluids for the EOR process in porous media. This is
supported by that these data include various properties of rock, oil,
nanoparticles, and polymer, which represents a data bank that is useful
in setting and deriving fundamentals and general rules taken into
consideration when implementing a field scale for nano-polymer pro­
jects. The main objective of this study is to conduct a comprehensive
assessment of incremental oil recovery using nanoparticles assisted
polymer flooding based on core flooding experiments from the litera­
ture. More than 350 core flooding experiments from prestige published
articles were considered for analysis. The comprehensive assessment for
nanoparticles-assisted polymer flooding was performed using statistical
approaches such as main effect plots, analysis of variance (ANOVA), and
fractional factorial design (FFD). Furthermore, an empirical correlation
for the incremental recovery factor of polymeric nanofluid-EOR using
ANN based on the gathered results of core flooding was also developed.
2. 2-Materials and methods
This paper aims at investigating and analyzing the results of the
previously published core flood experiments (more than 350 core flood
K.G. Salem et al.
Fig. 2. (a) Histogram of pore volumes of cores from the literature; (b) Histogram of porosities of cores from the literature (c) Histogram permeability of cores from
the literature. (d) Histogram oil saturation of cores from the literature.
tests) for NPs assisted polymer flooding. Consequently, the findings of
the present research are dependable. The objective is to shed light on the
impact of numerous factors (oil, rock properties, flooding conditions,
polymer, and NPs characteristics) on incremental oil recovery. Addi­
tionally, statistical analysis using screening factorial design was
accomplished to detect the highest meaningful parameters that had a
great impact on the incremental oil recovery of polymeric nanofluid.
Fig. 1 shows a methodology approach pursued in this research. Firstly,
the available data and results of published experimental tests were
compiled and revised. Furthermore, the database was visualized and
described using histograms, including all rock, fluid, polymer, and NPs
characteristics together with the results of these displacement tests. In
addition, the results of the gathered data were investigated, and
4
Geoenergy Science and Engineering 226 (2023) 211753
relationships between incremental/additional oil recovery versus rock-
fluid properties, polymer, and NPs were established using a scatter
plot and main effects plot. Moreover, factorial screening was imple­
mented to ascertain the most effective oil recovery parameters for
assisted polymer flooding using Pareto chart, normal, and half-normal
plots. Finally, the artificial neural network (ANN) was utilized to
calculate the incremental RF after secondary recovery for polymeric
nanofluid.
2.1. Data gathering
Datasets were compiled from the core flood experiments of
Nanoparticle-assisted polymer flooding published in the literature. This
K.G. Salem et al. Geoenergy Science and Engineering 226 (2023) 211753

Fig. 3. (a) Histogram of oil viscosity from the literature. (b) Histogram oil gravity from the literature.

Fig. 4. (a) Histogram of brine concentration from the literature. (b) Histogram of slug size from the literature; (c) Histogram of flooding temperature from the
literature. (d) Histogram of injection rate from the literature.

work gathered data from 60 publications with more than 350 experi­
ments conducted in laboratory core flooding/displacement tests in
sandstone and carbonate cores that used Nanoparticles assisted polymer
flooding. (Abdullahi et al., 2019; Agi et al., 2020a; Agi et al., 2019; Agi
et al., 2020b; Akbar et al., 2021; Akbar et al., 2020; Ali et al., 2019; Ali
et al., 2021; Aliabadian et al., 2020; Bera et al., 2020; Cao et al., 2019;
Cao et al., 2018a; Cao et al., 2018b; CHEN et al., 2020; Cheng et al.,
2017; Cheraghian and Khalilinezhad, 2015; Corredor et al., 2019a;
Corredor et al., 2019b; Corredor et al., 2019c; Corredor et al., 2021;
El-Hoshoudy et al., 2016; El-hoshoudy et al., 2019; Elhaei et al., 2021;
Elkady, 2016; Fan et al., 2021; Gbadamosi et al., 2022b; Gbadamosi
et al., 2019a; Gbadamosi et al., 2019b; Hamdi et al., 2022; Hamdi et al.,
2020; Haruna et al., 2020; Helmi et al., 2020; Hu et al., 2018; Hu et al.,
2019; Hu et al., 2021; Hu et al., 2017; Izadi et al., 2019; Keykhosravi

5
K.G. Salem et al. Geoenergy Science and Engineering 226 (2023) 211753

Fig. 5. (a)Distribution of polymer used in core floods of nanoparticles assisted polymer flooding. (b) Distribution of nanoparticles used in core floods of Nano­
particles assisted polymer flooding.

Fig. 6. (a) Histogram of polymer concentration from the literature. (b) Histogram of NPs size from the literature; (c) Histogram of NPs concentration from
the literature.

et al., 2021; Kumar et al., 2020; Lai et al., 2019; Lai et al., 2016; Liu
et al., 2022; Liu et al., 2017a; Liu et al., 2017b; Long et al., 2019a; Long
et al., 2019b; Maghzi et al., 2014; Maghzi et al., 2013; Maghzi et al.,
2011; Orodu et al., 2019; Orodu et al., 2018; Ponnapati et al., 2011; Pu
et al., 2015; Rezaei et al., 2016; Rueda et al., 2020; Saha et al., 2018;
Soliman et al., 2020; Sowunmi et al., 2022; Zahiri et al., 2022; Zheng
et al., 2017; Zhu et al., 2014). Numerous parameters were introduced in
the literature as parameters affecting the performance of Nanoparticle
assisted polymer flooding. The core flood experiments had a diverse
range of core and fluid properties in addition to various parameters of
flood conditions such as temperature, slug size, injection rate, salinity
concentration, polymer concentration, and nanoparticle concentration.
The brief of the collected data, incorporating the core properties, oil
properties, injection scheme, polymer, NPs, and results of core floods, is
presented in Table 1. Additionally, Appendix A (Table A1) describes the
gathered data of the core flooding experiments from the published
articles.

6
K.G. Salem et al. Geoenergy Science and Engineering 226 (2023) 211753

Fig. 7. (a) Histogram of incremental oil recovery from the literature. (b) Histogram of total oil recovery from the literature.

Table 2
Statistical evaluation of the datasets used in this research work.
Variable Unit Mean StDev Minimum Median Maximum Skewness Kurtosis

Permeability mD 867 1523.9 0.2 251.7 8266 2.64 6.95

oil viscosity cp 250.5 558.1 2 15.2 2400 2.66 6.64
Salinity ppm 28,380 21,823 500 22,000 138,000 1.62 3.8
Slug size PV 1.7684 1.689 0.2 1 7.2 1.35 0.76
flooding temperature. C 54.28 28.56 25 50 120 0.53 − 1
polymer concentration ppm 1569.3 1635.1 0 1000 5000 0.85 − 0.38
NPs concentration wt.% 0.4485 0.9613 0 0.1 5 3.03 9.04
Incremental oil recovery % Of OOIP 17.218 12.55 2.35 12.25 64.9 1.09 0.39

Fig. 8. (a) Scatterplot of incremental oil recovery vs permeability. (b) Main effect plot of permeability on incremental oil recovery.

3. Results & discussion
3.1. Rock and fluid properties
Different core models were stated in the literature aimed at con­
ducting core flooding experiments of nanoparticles-assisted polymer
flooding. These core models were sand pack and artificial cores, sand­
stone cores, carbonate cores, and glass micromodel. Most of the models
were conducted using sandstone cores (138 experiments) and sand
packs model (127 experiments) while few experiments were conducted
using carbonate cores (13 experiments). Around 43% of the tests were
conducted in the sandstone cores including Berea, outcrops, and
Reservoir cores. 40% of the tests were conducted in sand packs and
artificial cores. Few tests (13%) were conducted using micromodel,
while even fewer were conducted on carbonate cores (4%). In this work,
Histograms were employed to display the dataset graphically and to
describe the sampling distribution of the dataset. Additionally, the his­
togram demonstrates the frequency of the dataset on the y-axis and the
variables being measured on the x-axis. Fig. 2a, b, c, and d show the
histograms of the core properties (pore volume, porosity, permeability,
and oil saturation). The datasets cover a range of porosity from 6 up to
44% but most of the data points fall between 15 and 35%. The sand pack
and artificial cores had high values of porosity. Furthermore, the
permeability data had a wide range of values. The highest permeability
recorded in the literature was 8266 mD for the sand pack model,
whereas the lowest permeability recorded in the literature was 0.2 mD

7
K.G. Salem et al.
for carbonate rock. All cores flooded with heavy oil had permeability
values almost above 1000 mD. The majority dataset of pore volume falls
between 5 and 50 cc. All datasets for initial oil saturation above 50% and
the most frequent data was 80%. Fig. 3 a and b display oil viscosity and
oil gravity that have been reported in the literature. The maximum oil
viscosity recorded in the dataset was 2400 cp, while the minimum vis­
cosity was 2 cp. This indicates that the tested core floods were conducted
with light, medium, and heavy oil as this notice endorsed by data of oil
gravity which ranged from 12 to 42 API.
3.2. Flooding condition data
Fig. 4a shows brine/salinity concentration; most values varied from
0 ppm to 138,000 ppm. Some studies used deionized water in which the
salinity was zero. The chemical slug size by pore volume (PV) from the
previous studies is shown in Fig. 4b. Notably, the lowest and highest slug
sizes were 0.1 PV and 7 PV. The most frequent value for the injection
volume was 0.5 pore volume. Generally, most of the previous in­
vestigations ceased the injection when the water cut reached 99%).
Fig. 4c summarizes the distributions of injection flow rate where the
most frequent values ranged from 0.1 to 0.5 mL/min. However, the
highest value was 3 mL/min, and most literature recorded low values for
injection flow rate. Fig. 4d depicts that the flooding experiments of
Nanohybrid polymer were conducted at different values of temperature.
The temperature values ranged from ambient temperature to the highest
value of 120 ◦ C.
3.3. Polymer and NPs characteristics
Both synthetic polymers (such as HPAM and PAM) and biopolymers
(Such as Xanthan, Guar Gum, and Arabic Gum) were investigated in core
flooding of nanoparticles-assisted polymer flooding as depicted in
Fig. 5a. HPAM/PAM and Xanthan gum were the most frequently tested
during the experiments as shown in Fig. 5a. The molecular weight for
these tested polymers varied from 1 to 25 million Dalton as reported in
the literature. Moreover, numerous nanoparticles were tested such as
silicon oxide, aluminum oxide, titanium oxide, and graphene oxide.
Silicon oxide and aluminum oxide were the most frequently investigated
for the lab investigations as depicted in Fig. 5b. The most significant
parameters for nanofluid polymeric flooding are polymer and nano­
particle concentration. Fig. 6a shows the polymer concentration re­
ported in the literature. The polymer concentration is in the range of
500–3000 ppm. The nanoparticles concentration is considered a key
parameter for the incremental oil recovery induced by polymeric
nanofluids. Fig. 6b illustrates the histogram of NPs concentrations from
the reported studies. The data revealed that even though a few
Fig. 9. (a) Scatterplot of incremental oil recovery vs slug size. (b) Main effect plot of slug size on incremental oil recovery.
8
Geoenergy Science and Engineering 226 (2023) 211753
experiments were conducted with high concentrations, the majority of
core floods employed NP concentrations less than 0.5 wt%. Additionally,
most of the experiments utilized NPs size below 50 nm as shown in
Fig. 6c. One of the core flood objectives of experiments with nano-
polymer is to ascertain their potential results and appraise their tech­
nical feasibility. Fig. 7a and b depict the histograms of the incremental
oil recovery over water flooding (WF) and total oil recovery from the
reported studies. Based on the gathered data, the average oil recovery
after water flooding (WF) was 43%, while the mean incremental/addi­
tional oil recovery was 18%.
3.4. Data selection/inclusion for investigation
The selection criteria for the inclusion of the gathered data into the
proposed investigation were based on two aspects. The first was that the
selected parameter should be depicted and offered in most of the gathered
core flood experiments. The second involves the significance of the pa­
rameters on the incremental oil recovery from a reservoir engineering
perspective. For instance, Porosity was excluded because most porosity
values of core floods were in the same ranges. Additionally, the initial oil
saturation was excluded as this factor doesn’t affect the tertiary recovery
with the same degree of residual oil saturation. If the gathered data bank
had large datasets of residual oil saturation, it would be a significant
parameter in the assessment of incremental oil recovery from polymeric
nanofluid. Furthermore, it’s well known that the polymer’s viscosity is
changed or degraded over time during the injection. In all core floods of
polymer flooding, polymer concentration is considered a key input
parameter in all experiments, as stated in the literature. So, we prefer to
analyze the most accurate and reported parameter, which is taken into
consideration through any core flooding experiments of polymer. Hence,
for these reasons, we selected seven parameters for evaluation and exam­
ined their impact on incremental oil recovery in the following sections.
To capture a descriptive evaluation of the collected data, the statis­
tical parameters for the gathered data that encompass mean, minimum,
maximum, standard deviation, and skewness were presented in Table 2.
Such statistical parameters offer a good understanding of the data
distribution.
3.5. Factors affecting the incremental oil recovery of NPs assisted polymer
flooding
Different single parameters of the gathered data (such as perme­
ability, oil viscosity, salinity, slug size, temperature, polymer concen­
tration, and NPs concentration) were individually investigated using the
main effect and scatter plots to study their impact on the incremental oil
recovery by polymeric nanofluid flooding.
K.G. Salem et al. Geoenergy Science and Engineering 226 (2023) 211753

Fig. 10. (a) Scatterplot of incremental oil recovery vs salinity. (b) Main effect plot of salinity on incremental oil recovery.

Fig. 11. (a) Scatterplot of incremental oil recovery vs salinity. (b) Main effect plot of salinity on incremental oil recovery.

Fig. 12. (a) Scatterplot of incremental oil recovery vs polymer concentration. (b) Main effect plot of polymer concentration on incremental oil recovery.

9
K.G. Salem et al.
3.5.1. Effect of permeability
Fig. 8-a illustrates a cross plot of permeability (MD) vs the incre­
mental oil recovery after waterflooding for all the tests performed. As
shown in the scatterplot, the permeability had a broad range of values
due to flooding different types of cores and fluids. It is shown that there
are two clusters of datasets: low and high values of cores permeability.
As an illustration, core floods in the case of heavy oil needed high
permeable rock to guarantee successful flooding. consequently, perme­
ability values over 1000 md (Darcy level) are necessary criteria to attain
high incremental oil recovery in case of high viscous oil flooding (Cor­
redor et al., 2019b, 2019c). Furthermore, Permeability has a significant
effect on incremental recovery as seen in Fig. 8a and b. It was observed
that several successful core flood tests have been performed in low
permeability reservoirs (Akbar et al., 2021; Lai et al., 2019; Long et al.,
2019a, 2019b). From the gathered datasets, Polymeric nanofluid had the
prospective to perform in both low and high-permeability media.
It should be highlighted that injection of a high-ranking molecular
weight polymer in low permeable rock/reservoirs leads to minimal in­
cremental oil recovery due to mechanical trapping of the polymer, and
accordingly the high amount of retention. Nevertheless, nanoparticles
are microscopic/tiny particles that can pass easily across minuscule
pores. Consequently, a nanoparticle-assisted polymer solution retains
less in porous media than a conventional polymer. This enhancement led
to improved polymer performance and better oil recovery in low
permeability cores almost 10 md as reported by Reza Elhaei et al.
(2021). Overall, the main effect plot (Fig. 8b) shows that incremental oil
recovery increases with an increase in permeability values.
3.5.2. Effect of the injected/slug size
Determination of optimum injected/slug size is a critical factor to
make polymeric nanofluid more economically profitable and technically
feasible. Typically, from a practical point of view, One PV of injected
polymer is frequently thought to recover incremental/additional oil
over waterflooding (Bera et al., 2020). However, in laboratory experi­
ments, the injection continued to recover more oil and achieve a water
cut near 99%. E. Helmi et al. (2020) investigated different slug sizes of
nanocomposite polymer using a sand pack model (0.2, 0.4, 0.6, and 0.8
PV) with constant values for nanocomposite concentration, salinity, and
flooding temperature, which were 2000 ppm, 30,000 ppm, and 40 ◦ C,
respectively. The authors observed that the oil recovery was higher with
the large slug size of the nanocomposite polymer than with the small
nanocomposite polymer. The cumulative oil recovery was 79.2, 88.5,
92.7 and 96.7% for 0.2, 0.4, 0.6, and 0.8 PV slug sizes respectively.
Additionally, their study showed that the ideal size of the nano polymer
slug was 40% of pore volume, which achieved 88.5% as ultimate oil
Fig. 13. (a) Scatterplot of incremental oil recovery vs polymer concentration. (b) Main effect plot of polymer concentration on incremental oil recovery.
10
Geoenergy Science and Engineering 226 (2023) 211753
recovery at a flooding temperature of 40 ◦ C. The scatter plot of incre­
mental oil recovery versus slug size depicted most of the slug size data
had a low trend ranging from 0.5 to 3 PV with a broad range of oil re­
covery as shown in Fig. 9a. Generally, as depicted in Fig. 9b, the slug size
had fair significance on the incremental recovery because most experi­
ments tested with moderate slug size due to experimental and economic
aspects.
3.5.3. Effect of the salinity
From the gathered data, the efficacy of polymeric nanofluid was
examined at numerous values of salinity from zero to 140,000 ppm as
revealed in Fig. 10a. Generally, the water salinity and divalent cations
can negatively affect the polymer performance in the reservoir. Never­
theless, the synergistic power of the polymer and NPs as a Nanohybrid
polymer diminishes the inverse impact of salt on the rheological prop­
erties of the polymer, and hence on incremental oil recovery (Hamdi
et al., 2022; Haruna et al., 2020; Izadi et al., 2019; Zhu et al., 2014).
Furthermore, the higher incremental oil recovery for nanocomposite
polymer flooding in elevated salinity and temperature could be
endorsed by the viscosity improvement of the nanohybrid polymer
contrasted to pure polymer, and its potential to behave in harsh condi­
tions without missing much of its viscosity. Ali maghzi et al. (Maghzi
et al., 2014) observed at the analogous salinity conditions, the recovery
factor with a nano-polymer solution is almost 10% higher than that of
pure polymer. It should be highlighted that when saline is launched in a
pure polymer, the ionic charge of the polymer (– CO–O− ) will be
neutralized by the cationic charge and consequently electrostatic re­
pulsions of polymer will be reduced. On the contrary in polymer mixed
with NPs, an additional positive charge will be created due to the
presence of NPs and thus endorse electrostatic repulsion among the
polymer classes. Maje Alhaji Haruna et al. (2020) conducted six exper­
iments to study the effect of the addition of NPs for polymer flooding on
incremental oil recovery at harsh conditions of salinity. Two different
displacing fluids were used, one was API brine with a salinity of almost
100,000 ppm and the second was formation brine with a salinity of
almost 138,000 ppm. Initially, the first 3 series of experiments were
done manipulating API brine exhibited incremental oil recovery at
4.30%, 7.61%, and 11.28% for pure polymer, SiO2-polymer, and
modified SiO2-polymer respectively. Additionally, the second 3 sets of
experiments were conducted using formation brine exhibited incre­
mental oil recovery at 3.56%, 5.23%, and 10.79% for pure polymer,
SiO2-polymer, and modified SiO2-polymer respectively. Such results
indicate that the nanohybrid polymer shows greater performance at
harsh salinity conditions. Overall, from the main effect plot (Fig. 10b),
minimal incremental oil recovery (below the mean values) was observed
K.G. Salem et al.
Fig. 14. (a) Scatterplot of incremental oil recovery vs NPs concentration. (b) Main effect plot of NPs concentration on incremental oil recovery.
at severe salinity conditions.
3.5.4. Effect of the temperature
The efficacy of polymeric nanofluid was investigated at different
values of reservoir temperature in the literature. Based on the gathered
dataset, the maximum temperature of nano-polymer flooded oil reser­
voirs was 120 ◦ C (Agi et al., 2020b). As shown in Fig. 11a, the data
dispersion was triggered by the discrepancy in the performance from
various core flood temperatures from ambient conditions to 120 ◦ C. E.
Helmi et al. (2020) studied the impact of temperature on the perfor­
mance of nanocomposite polymer flooding. In their study, they inves­
tigated different flooding temperatures of 40, 60, 80, 90, and 100 ◦ C
with constant values for slug size, nanocomposite concentration, and
salinity concentration, which were 40% PV, 2000 ppm, and 30,000 ppm,
respectively. They found that increasing the flooding temperature
resulted in a reduction of the ultimate oil recovery using nanocomposite
polymer flooding. Additionally, they observed that nanocomposite
polymer endured temperatures up to 90 ◦ C (194 ◦ F), and no incremental
oil recovery was recorded after this temperature due to the occurrence of
thermal degradation for nanocomposite polymer. Additionally, core
Fig. 15. Pareto chart for results of all datasets of polymeric nanofluid.
11
Geoenergy Science and Engineering 226 (2023) 211753
flooding tests were performed by Augustine Agi et al. (2020b) to
simulate polymeric nanofluid flooding the Sarawak oil field at a harsh
temperature condition of 120 ◦ C. Incremental oil recovery was recorded
as 13%, 17%, and 23% for SiO2 polymeric nanofluid, Al2O3 polymeric
nanofluid, and crystalline starch nanofluid (CSNF) respectively. In
comparison to SiO2PNF and Al2O3PNF, the microscopic image of the
emulsion formed reveals the long-term stability of the emulsion and
confirms the potential of CSNF to tolerate high-temperature degrada­
tion. According to the findings, the CSNF could be a good alternative to
conventional polymer flooding. Overall, as shown from the main effect
plot Fig. 11b, significant incremental oil recovery was detected at
moderate temperatures while a minimal incremental oil recovery
(below the mean value) was observed at high-temperature conditions.
3.5.5. Effect of polymer concentration
Typically, a high concentration of polymer is considered necessary to
reach the preferred viscosity of the displacing fluid (Bera et al., 2020). In
addition, a higher value of polymer viscosity can be attained by esca­
lating either the molecular weight or polymer concentration (Saboor­
ian-Jooybari et al., 2016). Fig. 12a shows the incremental oil recovery
K.G. Salem et al. Geoenergy Science and Engineering 226 (2023) 211753

Fig. 16. Pareto chart for results of polymeric nanofluid flooded heavy oil.

Fig. 17. Pareto chart for results of polymeric nanofluid flooded light oil.

vs. the polymer concentration. Overall, it was indicated that higher in­
Table 3
cremental oil recovery was achieved as polymer concentration
ANOVA analysis for all datasets of polymeric nanofluid (First Scenario).
increased. Both synthetic polymers and natural polymers were utilized
in core flood tests of polymeric nanofluid. The utilized polymer con­ Variable DOF Adj. SS Adj. MS F P
Value Value
centration depends on its type. It should be highlighted that in the case
of biopolymers, a polymer that has superior concentration is utilized in Model 7 19700.6 2814.37 25.46 0
Permeability, md 1 2390 2390.04 21.62 0
experiments compared to a synthetic polymer such as polyacrylamide
oil viscosity, cp 1 0.1 0.12 0 0.974
and its derivative. The interpretation for that is biopolymers generally salinity, ppm 1 71 70.97 0.64 0.424
have low molecular weight, and it is a well-established fact that the slug size, PV 1 353.4 353.4 3.2 0.075
viscosity of polymer boosts with the augmentation of its molecular flooding temperature, C 1 5.8 5.76 0.05 0.82
weight and vice versa. Commonly, it was reported that the molecular polymer concentration, 1 554.2 554.16 5.01 0.026
ppm
weight of xanthan gum and PAM/HPAM were 2 M Da and 9–25 M Da. NPs concentration, wt.% 1 3834.1 3834.11 34.69 0
Hence, a higher concentration is required in case of slight molecular
weight to attain the desired viscosity of the polymer. In the case of
medium and high permeability conditions, a polymer that shoulders

12
K.G. Salem et al. Geoenergy Science and Engineering 226 (2023) 211753

Table 4 2021). Ali Maghzi et al. (2014) stated that the viscosity of the polymeric
ANOVA analysis for polymeric nanofluid flooded heavy oil (Second Scenario). nanofluid was boosted by increasing the NPs concentration. The authors
Variable DOF Adj. SS Adj. MS F Value P Value explained that the phenomenon was assigned to the formation of
ion-dipole interaction between the cations and NPs. Results of core
Model 7 4905.6 700.8 7.23 0
Permeability, md 1 1389 1389 14.34 0 flooding conducted by Afeez Gbadamosi et al. (2022b) demonstrated
oil viscosity, cp 1 139.9 139.87 1.44 0.233 that the polymeric nanofluids of Al2O3-xanthan gum, TiO2- xanthan
salinity, ppm 1 665.4 665.38 6.87 0.011 gum, and SiO2- xanthan gum had incremental oil recovery up to 28.4%
slug size, PV 1 44 43.97 0.45 0.502 OOIP after WF. In addition, they observed that usage of metal oxide
flooding temperature 1 3.7 3.69 0.04 0.846
polymer concentration, 1 3.5 3.53 0.04 0.849
polymer nanofluids (TiO2-XG, and Al2O3-XG) exhibited higher re­
ppm coveries than SiO2-XG. On contrary, Yanqiu Hu et al. (2021) stated that
NPs concentration, wt.% 1 1613.9 1613.91 16.66 0 SiO2 –polymer flooding delivered improved oil recovery than
Al2O3–polymer. Concerning the existence of SiO2 nanoparticles, IFT for
oil-water and oil-polymer was decreased by increasing the NPs con­
Table 5 centration Nevertheless, after 0.2 wt% of NPs concentration, there was
ANOVA analysis for polymeric nanofluid flooded light oil (Third Scenario). no considerable alteration in contact angle. Thus, it’s noticeable that
Variable DOF Adj. SS Adj. MS F Value P Value
laboratory tests for each particular reservoir are essential to establish the
optimal NPs concentration before the execution of polymeric nanofluid
Model 7 3699.2 528.46 5.44 0
in the field scale. Overall, the main effect plot (Fig. 14b) shows the in­
Permeability, md 1 1643.2 1643.15 16.92 0
oil viscosity, cp 1 126.3 126.35 1.3 0.256 cremental oil recovery significantly improves with an increase in NPs
salinity, ppm 1 87.7 87.67 0.9 0.343 concentration.
slug size, PV 1 93.3 93.33 0.96 0.328
flooding temperature, C 1 2.1 2.12 0.02 0.883 3.5.8. Effect of heterogeneity
polymer concentration, 1 117.4 117.42 1.21 0.273
ppm
Although the effect of heterogeneity is a crucial factor for oil
NPs concentration, wt.% 1 256.9 256.86 2.65 0.106 exploitation, unfortunately not represented numerically in core flooding
experiments. Thus, this assessment has some constraints in terms of data
gathering due to a lack of some data such as heterogeneity. Hence, a
lower polymer concentration but ends up with high displacement effi­ larger dataset in terms of heterogeneity input parameters would result in
ciency will be more beneficial as a hyperbranched polymer synthesized a more precise assessment of incremental oil recovery of nanoparticle-
by Liu et al., 2017a, 2017b. It was clear from the main effect plot assisted polymer flooding. Indubitably, heterogeneity is a crucial fac­
(Fig. 12b) that the oil recovery is positively influenced by the increase in tor in oil recovery through a reservoir. It is a measure of the permeability
polymer concentration. variation in the layered reservoirs or cores. However, lab-core flooding
represents or simulates one layer or one dimension, and accordingly,
3.5.6. Effect of oil viscosity one value for permeability is assumed. Thus, there were no available
Although most displacement tests were performed using light oil, data about the heterogeneity parameter (permeability variation) in the
several successful experiments flooded successfully using heavy oil core flooding, and it was not considered in experiments (in the litera­
(Aliabadian et al., 2020; Corredor et al., 2019a, 2019b, 2021; Maghzi ture). In a core flood lab-scale investigation, heterogeneity could be
et al., 2011, 2013, 2014; Saha et al., 2018). In other words, Core floods taken into consideration in one case if the investigator injects polymer or
of polymeric nanofluid were successfully tested in low, medium, and fluid in multiple parallel cores (2 or 3 cores) at the same time, and this
heavy oil reservoirs. For a long period, Polymer injection was once case is rarely reported in lab investigations of core flooding (Long et al.,
thought to be an ineffective EOR method for heavy oils (Saboor­ 2019b). In general, the heterogeneity parameter is evaluated through
ian-Jooybari et al., 2016). However, successful projects in the field have numerical simulation in layered reservoirs. This is generally done by
opened up possibilities for the use of polymer injection in heavy oil. At generating permeability distributions in the i, j, and k directions and
the field scale, the common cases were reported in Seal fields and populating them into the simulation grids. Additionally, the heteroge­
Pelican Lake with oil viscosities of 5000 and 5400 cp, respectively neity parameter is commonly represented through numerical simulation
(Saboorian-Jooybari et al., 2016; Sheng et al., 2015). As shown in Fig. 13 using the Dykstra-Parsons coefficient.
a, the highest oil viscosity reported in core flood tests of polymeric
nanofluid was 2400. Additionally, the data dispersion was likely caused 4. Fractional factorial design and analysis of variance
by the discrepancy in oil viscosities from different core flood tests
(Fig. 13 a). As shown from the effect plot (Fig. 13b), The results reveal It was clear from previous discussions that different factors influence
that higher incremental oil recovery was in the case of viscous oil. the performance of polymeric nanofluid EOR. Consequently, fractional
Factually, this phenomenon is attributed to the recovery factor after WF factorial design (FFD) was applied to screen the highest significant
being relatively low (~10% of OOIP) due to the unfavorable-mobility factors of the previous experimental design. Minitab software was
ratio, which causes viscous fingering and low sweep efficiency (Ali employed to apply these statistical assessments (Minitab, 2020). In
et al., 2019). general, one of the significant statistical tools for examining the impact
of numerous controllable elements on interest response is the fractional
3.5.7. Effect of the NPs concentration and NPs type factorial design (FFD) (Bengar et al., 2017; Snosy et al., 2020, 2021).
The concentration and type of NPs in nano polymeric fluids are Moreover, FFD is a widely used approach for experimental design to
significant factors, and they will directly affect the flooding performance improve and streamline lab investigations.
and hence the oil recovery. From the scatter plot (Fig. 14a), the most The Plackett-Burman design was selected to screen the important
reported concentrations were 0.05–3 wt%. In some experiments, the parameters of core flood experiments that affect incremental oil recov­
scholars aimed at establishing the optimum concentration of NPs and ery concerning their main effects neither than interaction effects.
their advantage on IFT reduction and wettability modification as well Furthermore, the results of FFD are shown in the Pareto figure. The
(Hu et al., 2021). The reported results reveal that the oil recovery im­ Pareto chart displays the effects of the factors from the greatest impact to
proves with an increase of NPs concentration up to the optimum limit. the lowest impact using horizontal bars. In addition, a reference line is
Nevertheless, it was observed that NPs concentrations greater than 3 wt plotted on the Pareto chart to indicate which effects are statistically
% might result in a decrease in reservoir permeability (Eltoum et al., significant.

13
K.G. Salem et al. Geoenergy Science and Engineering 226 (2023) 211753

Table 6
Statistical analysis of the data employed for ANN training and testing processes.
Variable Mean St. Dev. Minimum Maximum

Training testing training testing training testing training testing

permeability, md 1236.43 1238.27 1992.22 1925.23 0.68 0.84 8266 6660

oil viscosity, cp 366.07 353.86 725.72 718.27 2 2 2400 2400
salinity, PPM 21797.04 22858.20 15813.02 20844.16 500 1000 138,000 138,000
Slug size, PV 1.41 1.10 1.49 0.96 0.25 0.2 7.6 5
flooding temperature, C 46.45 46.12 25.40 23.67 25 25 120 120
polymer concentration, PPM 2635.98 3654.88 5133.17 7644.56 0 0 50,000 50,000
NPs concentration, wt.% 0.46 0.48 0.88 0.93 0 0 5 5
Incremental oil recovery, % OOIP 17.65 19.52 13.61 15.05 0 1.6 50.2 64.9

Fig. 18. Architecture of the ANN model.

In the present study, the analysis was performed using rock perme­
ability, oil viscosity, salinity, slug size, flooding temperature, polymer
concentration, and NPs concentration. Considering there was a signifi­
cant discrepancy between the results of cores flooded with heavy oil and
cores flooded with light oil. Accordingly, the statistical analysis was
performed via three scenarios. The first analysis was accomplished using
all the gathered data flooding heavy and light oil. The second analysis
was performed only using data flooding heavy oil, while the third
analysis was performed using only data flooding light oil. Fig. 15 dis­
plays the Pareto chart generated by FFD for the results of core floods of
polymeric nanofluid. NPs concentration, reservoir permeability, and
polymer concentration are found to have the most significant on the
performance of incremental oil recovery. Fig. 16 depicts the FFD output/
Pareto chart for the results of polymeric nanofluid flooded heavy oil.
NPs and rock permeability were found to have the most dominant per­
formance in incremental oil recovery for polymeric nanofluids. Fig. 17
displays the Pareto chart generated by FFD for the results of polymeric
nanofluid flooded light oil. On the Pareto graph, the factors that
outperform the reference line are statistically significant. Thus, it is
obvious that NPs and rock permeability were found the most dominant
parameters in the oil recovery of polymeric nanofluid. The statistical
analysis using FFD is matched with the aforementioned discussion.
Furthermore, rock permeability outperformed the reference line in all
three analyses. Results of FFD indicate that the NPs concentration,
polymer concentration, and reservoir permeability were the most sig­
nificant parameters and assessment efforts should therefore be concen­
trated on these parameters.
The analysis of variance (ANOVA) as shown in Tables 3–5 was used
to acquire the quantitative interpretation for the investigated parame­
ters affecting incremental oil recovery of polymeric nanofluid. As shown
in the first scenario (all gathered data), rock permeability, polymer
concentration, and NPs concentration give a significant impact on

14
K.G. Salem et al. Geoenergy Science and Engineering 226 (2023) 211753

Table 7 Table 8
Summary of the different parameters employed to build the ANN model (Jetter, Evaluation summary of the proposed ANN model.
2014). Type of data R2 AAPE, % MSE RMSE
parameter Options/ranges Optimized
Training 0.953 5.6 8.55 2.92
parameters
Testing 0.952 8.7 10.5 3.24
Neurons of input layer Based on inputs 7
Neurons of the output Based on outputs 1
layer incremental oil recovery based on the P- the value of less than 0.05.
Hidden-layers Number 1 Furthermore, the rock permeability exhibits the highest F- value (21.62)
Neurons of hidden layer 1–40 30
which implies the strong influence of the parameter compared with
The transfer function of 1. TAN HYP LOGISTIC
the hidden layer 2. LOGISTIC other parameters. For the second scenario (cores flooded heavy oil), rock
3. IDENTICAL permeability and NPs concentration give significant results to incre­
4. BINARY mental oil recovery based on the P- value less than 0.05 and the F- values
5. IDENTICAL 0 TO 1 of 14.34 and 2.65 respectively. For the third scenario (cores flooded
learning algorithms 1. Std. BP (No Loopback RPROP (Full
support) Loopback support)
light oil), rock permeability was the alone parameter that had a P- value
2. BP (Full Loopback support) less than 0.05 and the greatest F-value (16.92) which also reflects its
3. Std. BP with Momentum significance on oil recovery.
(No Loopback support) To sum up, the interpretation of the ANOVA was consistent with the
4. BP with Momentum (Full
FFD analysis. As attested by the results of FFD and the analysis of
Loopback support)
5. RPROP (Full Loopback variance (ANOVA), rock permeability, NP concentration, and polymer
support) concentration should be considered highly important when planning
6. cascade Correlation and designing polymer nanohybrid projects at both the core and field
scales.

Fig. 19. Actual vs predicted incremental RF of polymeric nanofluid using ANN model for training data sets.

Fig. 20. Actual vs predicted incremental RF of polymeric nanofluid using ANN model for testing data sets.

15
K.G. Salem et al.
Fig. 21. The actual, and ANN-based incremental oil recovery of polymeric nanofluid for the training dataset.
Fig. 22. The actual, and ANN-based incremental RF of polymeric nanofluid for the testing dataset.
5. Machine learning
Machine learning applications have gained a lot of attention in the
oil industry because of their accuracy. Different machine learning al­
gorithms were used, such as the adaptive neuro-fuzzy inference system
(ANFIS) (Ibrahim et al., 2022; Shayan Nasr et al., 2021), random forest
(Hakiki and Arifurrahman, 2023) and artificial neural network (ANN)
(Gomaa et al., 2022a; Salem et al., 2018). It should be highlighted that
the prediction performance of machine learning algorithms differs from
one specific case to another. In this regard, different types of machine
learning were employed by Hakiki and Arifurrahman (2023) to predict
polymer gelation time. Their results reported that random forest ma­
chine learning performed better in comparison to the artificial neural
network (ANN), K nearest neighbor (KNN), and convolutional neural
network (CNN). Furthermore, Ali et al. (2022) utilized four machine
learning algorithms to predict the saturation pressure of reservoir fluids:
Gradient Boosting Machine (LGBM), K-Nearest Neighbor (KNN), Sup­
port vector machine (SVM), and artificial neural network (ANN). Their
results stated that LGBM and ANN were the highest predictive efficiency.
Overall, as one of the most efficient and widely accepted machine
learning techniques, ANN was utilized in the present study to predict the
incremental oil recovery of polymeric nanofluid.
5.1. Development of artificial neural network
In the context of simplifying the artificial neural network (ANN),
ANN simulates the construction and operation of a human brain system.
ANN has become well known for modeling numerous geoscience and
petroleum engineering problems with a high ranking of accuracy
(Abdulaziz and Hawary, 2020; Abdulaziz et al., 2019; Al-AbdulJabbar
et al., 2022; Alsabaa et al., 2021; El-hoshoudy et al., 2022; Gomaa et al.,
16
Geoenergy Science and Engineering 226 (2023) 211753
2022a; Gomaa et al., 2022b; Gouda et al., 2021; Hassan et al., 2019;
Ibrahim et al., 2022; Kalam et al., 2021; Salem et al., 2018). Generally,
The ANN topology is made up of three layers: input, hidden, and output
respectively as well as each layer consists of neurons. The input layer
receives the data, sends it to the hidden layer for processing, and then
sends it to the output layer. The layers consist of neurons (nodes) that
handle all the data processing, controlled by a transfer function. In
addition, feed-forward neural networks are regarded as the most
prominent network topology due to the convenience of numerous
training algorithms, flexible design, and proper description. Further­
more, the biases and weights of neurons ought to be released from the
ANN model to develop an empirical equation for the prediction of in­
cremental oil recovery. The most significant input parameters affecting
the incremental recovery were selected. ANN was employed in the core
flood dataset to predict the incremental oil recovery of Nanoparticle
assisted polymer flooding using the MemBrain neural network simulator
(Jetter, 2014). In this respect, the database was gathered from the pre­
vious core flood experiments. The compiled datasets included 329 data
points for permeability, oil viscosity, slug/injected size, polymer con­
centration, the salinity of media, NPs concentration in addition to in­
cremental oil recovery. Before the model construction, the assembled
data was split into the training process (75% of points) and the testing
process (25% of points). The training data was employed in the devel­
opment stage of the model, while the testing data was considered to
verify and assess the reliability/consistency of the model. Table 6 shows
some simple statistics evaluated for the data employed in ANN training
and testing processes. To achieve the optimal model, numerous struc­
tural layouts with different parameters should be examined. The optimal
design and architecture of the ANN models were determined after these
observations as presented in Fig. 18 and Table 7.
The quality of the developed model was assessed utilizing absolute
K.G. Salem et al. Geoenergy Science and Engineering 226 (2023) 211753

Table 9
Weights and Biases of ANN-based model for incremental RF prediction of polymeric nanofluid.
[ [ ]
∑
N
1
RFincr. = W2i + b2 (equation 5)
i=1 1 + e (W1i,1− A+W1i,2 B+W1i,3 C+W1i,4 D+W1i,5 E+W1i,6 F+W1i,7 G.+b1i )2̂

i W1i, j W2i b1, j b2

j=1 j=2 j=3 j=4 j=5 j=6 j=7

1 0.351 0.376 − 3.865 0.298 0.766 3.749 2.415 0.566 0.362 0.114
2 0.359 − 0.462 0.275 0.261 0.161 0.003 0.350 0.445 − 0.011
3 − 3.917 − 6.030 6.786 − 4.245 − 4.700 − 111.824 3.817 − 1.156 − 1.097
4 − 2.236 0.275 − 0.200 2.324 − 1.727 8.035 1.243 − 0.610 − 0.345
5 1.979 18.623 − 3.061 3.892 − 1.232 12.884 − 8.006 − 0.797 0.123
6 − 0.294 − 32.869 − 30.095 76.224 5.098 5.948 − 13.797 − 0.608 − 1.462
7 0.249 − 19.475 2.578 − 4.147 − 0.615 − 10.543 3.068 − 2.005 − 0.258
8 − 28.790 − 3.806 30.701 2.905 12.349 − 21.590 188.992 0.892 2.599
9 − 374.674 8.405 − 2.577 5.072 − 1.605 − 13.430 18.414 − 1.111 1.064
10 5.766 − 10.777 0.527 − 0.635 − 5.733 − 29.963 4.653 1.159 − 0.620
11 14.559 − 14.985 − 5.455 19.082 − 34.449 − 0.087 20.347 − 0.805 0.280
12 − 24.292 4.734 − 8.644 44.879 6.638 62.538 67.399 0.280 0.116
13 − 52.170 − 4.723 13.557 − 0.731 − 3.845 40.193 − 140.163 − 0.395 − 1.013
14 − 1.113 0.391 − 1.047 − 1.283 0.002 0.277 − 0.046 − 0.339 − 0.513
15 − 1.894 − 0.191 0.942 1.693 0.769 0.457 − 2.086 0.877 0.177
16 − 7.538 − 2.513 3.397 − 2.003 − 9.389 − 22.125 4.105 − 1.276 − 1.080
17 33.609 0.388 2.732 − 9.230 − 10.695 − 79.723 − 6.278 0.799 − 7.287
18 − 6.705 1.175 5.598 − 3.326 − 2.847 29.020 − 61.855 1.176 − 1.165
19 0.014 0.408 0.264 0.139 − 0.474 − 0.330 0.342 − 0.171 − 0.090
20 − 0.272 − 0.738 − 0.087 0.425 0.360 0.364 0.035 0.148 0.059
21 − 0.008 − 0.392 − 0.339 − 0.145 − 0.404 − 0.512 0.047 − 0.249 − 0.200
22 − 5.824 3.869 − 3.684 − 1.741 6.737 − 10.764 − 2.295 − 1.609 0.061
23 8.911 − 28.776 7.333 − 7.302 0.890 29.674 − 36.976 − 1.096 1.500
24 − 1.034 − 0.435 9.236 − 27.332 − 9.522 9.872 32.401 0.452 − 2.908
25 − 17.846 16.352 9.296 7.866 4.416 − 25.960 84.921 − 0.864 1.701
26 − 12.485 − 0.696 2.437 1.376 0.450 − 4.973 3.362 1.161 0.445
27 11.667 − 7.183 − 3.770 11.313 − 1.251 − 32.327 1.910 − 0.651 − 1.872
28 0.719 − 0.847 − 0.942 1.248 0.363 − 0.748 − 0.561 0.344 − 0.272
29 165.044 1.038 − 2.603 56.133 15.869 − 157.800 45.484 0.984 4.921
30 − 27.001 22.928 16.150 − 1.382 0.629 − 0.923 − 144.648 − 0.692 2.940

√̅̅̅̅̅̅̅̅̅̅
Table 10 RMSE = MSE (Equation 3)
Terms and symbols of the developed correlation.
N (
∑ )
Symbol/ Name/definition Unit yi actual − yi predicted 2̂
Factor
R2 = 1 − N (
i=1
(Equation 4)
∑ )
RF incr. Incremental recovery factor for polymeric % OOIP yi average actual − yi predicted 2̂
nanofluid i=1
A Rock permeability md
B Oil viscosity cp where N symbolizes several datasets; y i Actual, y i Predicted, and y i average
C Salinity ppm
Actual symbolizes actual incremental RF, predicted incremental RF, and
D Slug size PV
an average of actual values of incremental RF, respectively.
E Flooding temperature C
F Polymer concentration ppm
G NPs concentration Wt.%
W1i Weights between the input layer and hidden dimensionless 5.2. ANN model results
layer
W2i weights between the hidden layer and output dimensionless As previously mentioned, ANN was applied to 329 datasets to train
layer
and verify the model. Before ANN construction, the gathered dataset is
b1i biases for neurons of hidden layer dimensionless
b2 the bias of the output layer dimensionless randomly split into two categories (training and testing). 75% of the
data (245 points) is dedicated to the training process, while 25% of the
data (82 points) is assigned to testing and verification of the proposed
average error (AAPE), mean squared error (MSE), root mean squared ANN model. Typically, an ANN-based model was built using one -hidden
error (RMSE), and coefficient of determination (R2). AAPE, MSE, RMSE, layer and thirty neurons. Figs. 19 and 20 show plots for this predicted
and R2 were determined as follows (Ibrahim et al., 2022; Kalam et al., incremental RF in the ANN model versus the corresponding actual in­
2021; Shayan Nasr et al., 2021): cremental RF that exhibits good correlation coefficients. The correlation
∑
N coefficients between actual output and predicted outputs approximate
yi actual − yi
0.953 and 0.952 respectively in both training and testing approaches.
predicted
yi actual
× 100
AAPE = i=1 (Equation 1) Thus, the training and testing datasets have an AAPE of 5.6 and 8.7%
N
respectively. The evaluation summary of the proposed model is tabu­
N (
∑ ) lated in Table 8. Furthermore, Figs. 21 and 22 compare the incremental
yi actual − yi predicted 2̂ RF calculated using the ANN model to the actual incremental RF values.
MSE = i=1
(Equation 2) Thus, there are consistent stable trends between the actual incremental
N
RF and that calculated by the ANN model.

17
K.G. Salem et al. Geoenergy Science and Engineering 226 (2023) 211753

5.3. Development of an empirical correlation for the proposed ANN
As part of applying ANN, the development of a new empirical cor­
relation was one of the key outputs of the present study that can be
applied to calculate incremental RF for polymeric nanofluid. The
generated ANN model’s optimized weights and biases served as the
foundation for the creation of the new correlation. The model was
developed and regulated by one input layer (seven neurons), one-hidden
layer (thirty neurons), and an output layer (one neuron). Based on the
logistic transfer function, the novel correlation will be mathematically
formulated as shown in Equation (5). Additionally, the final updated
biases and weights of the established incremental RF model are pre­
sented in Table 9. Accordingly, these weights and biases enumerated in
Table 9 will be utilized as a substitution in Equation (5). Table 10
summarizes the symbols and terms of the developed correlation.
Such an ANN approach affords new insights to predict the incre­
mental oil recovery of polymeric nanofluid by ANN rather than time-
consuming and costly core flood experiments.
6. Conclusions
In this research, a robust investigation was conducted and data from
the literature were analyzed to develop new findings for NPs assisted
polymer flooding. Additionally, a new empirical correlation was
developed using ANN to predict the incremental oil recovery. The
following conclusions can be depicted:
• Because of the main effect plots and FFD, nanoparticles concentra­
tion, rock permeability, and polymer concentration are detected as
the parameters with the greatest impact on the incremental oil re­
covery of polymeric nanofluid.
• There is an optimum NPs concentration, optimum polymer concen­
tration, and optimum injected slug size for each reservoir rock.
Careful design of nano polymeric flooding for these parameters will
ameliorate the efficiency of flooding and yield high incremental oil
recovery.
• Determination of the optimum NPs and polymer concentrations is
essential for each specific reservoir using experimental work. The
common optimum NPs concentration is 0.1 wt%.
• Experimentally, polymeric nanofluid can result in a mean incre­
mental RF of 18% of OOIP.
• Although nano polymeric flooding is successful in low, medium, and
high permeability rocks, permeability values over 1000 md are a
necessary criterion to attain high incremental RF in case of heavy oil
flooding.
• An empirical correlation was developed to calculate the incremental
oil recovery factor of polymeric nanofluid EOR.
• The coefficient of determination (R2) of ANN-based correlation be­
tween the actual and calculated incremental recovery factor from the
ANN model was established to be 0.955 and 0.952 with AAPE of 5.6
and 8.7% respectively during the training and testing processes.
• It was clear that polymeric nanofluid in carbonate reservoirs has not
been adequately investigated and requires more extensive research.
Declaration of competing interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
Data availability
Data will be made available on request.

Appendix A
Table A.1
core-flood test results of the nanoparticles assisted polymer flooding compiled from the published articles.

Reference No. of K, md oil Slug Temperature Salinity, polymer NPs Incremental

Experiments viscosity, size, PV ,C ppm concentration., concentration., oil
cp ppm wt.% Recovery, %
OOIP

Hu et al. (2021) 3 450–680 20 3 60 6000 500 0.3 14–19

Elhaei et al. (2021) 3 10.34–12.9 20 3 NR 30,000 2000 0.05–0.1 5.25–7.97
Helmi et al. (2020) 9 129–157 4.48 0.2–0.8 40–90 30,000 2000 0.2 0.24–49.9
Orodu et al. (2019) 7 245–293.1 4.62 0.2–0.8 Ambient 30,000 30,000-50000 0.5–1.5 2.63–12.44
El-hoshoudy et al. 4 82.73 20.58 0.1–0.5 80 80,000 500–2000 0.2 7.4–19.2
(2019)
Cao et al. (2018b) 3 319.79–360.56 15.8 0.3 70 20,000 2500 0.1 6–16.3
Gbadamosi et al. 3 204 240 1 90 34,000 2000 0.1 30.6–40.2
(2019a)
Abdullahi et al. 5 3700–3900 217 2.5 70 138,000 2000 0.1 50.2–64.9
(2019)
Haruna et al. 6 100 42 2 85 10,000 1000 0.6 3.56–11.28
(2020)
Zheng et al. (2017) 4 NR 479.8 0.6 80 5835 600 0.6–2 7.08–10.54
Hamdi et al. (2022) 5 150–395.2 3.93 2 90 30,000 50 0.005 5–19
Kumar et al. (2020) 3 5014–6318 1190 1–2 85 11,000 5000 0.14 14.05–40.99
Saha et al. (2018) 5 746–1002 20.1 0.5 30 8890 5000 0.1–0.5 14.47–20.82
Corredor et al. 21 5640–6770 2400 1 25 3000–10000 4000 0.2 24–39
(2019b)
Aliabadian et al. 5 5000 940 0.5 24 NR 500 0.2 10.03–36.13
(2020)
Gbadamosi et al. 4 165.6–169.1 11 0.5 90 35,000 2000 0.1 26.3–37.6
(2019b)
El-Hoshoudy et al. 4 45.39 9.28 3 90 40,000 1000–4000 0.2 10.6–26
(2016)
Zahiri et al. (2022) 3 154–285 NR 1–2 30 42,000- 2000 0.5 11.4–17
60000
Rezaei et al. (2016) 3 294–311 45 1.3 70 2000 2000 0.1 1–10
(continued on next page)

18
K.G. Salem et al. Geoenergy Science and Engineering 226 (2023) 211753

Table A.1 (continued )

Reference No. of K, md oil Slug Temperature Salinity, polymer NPs Incremental
Experiments viscosity, size, PV ,C ppm concentration., concentration., oil
cp ppm wt.% Recovery, %
OOIP

Sowunmi et al. 40 97–298.15 14.5 0.5 30 NR 1000–10000 0.1–0.5 2.7–35.7

(2022)
Gbadamosi et al. 4 165.8–169.1 61 0.33 80 30,000 5000 0.5 25.2–28.4
(2022b)
Agi et al. (2020b) 4 201 10 0.5 120 22,000 2000 0.1–0.2 13–23
Bera et al. (2020) 7 200–205 4.95 3 50 2000 4000 0.2 10.73–44.28
Keykhosravi et al. 3 66–87 28 7 75 50,000 2000 0.5 12–25
(2021)
Cheraghian (2016) 5 283 1320 1 60 20,000 3150 0–2.5 39.1–43.3
Yousefvand and 3 NR NR 1 NR 30,000 800 0.5 9.69–18.37
Jafari (2015)
Rueda et al. (2020) 3 2500 21.9 3 80 30,000 160–250 0.1 9–12
Hu et al. (2018) 2 894–926 6.9 0.3 45 4743 1500 0.3 7.12–13.25
Corredor et al. 17 5276–6686 2400 1 25 10,000- 4000 0.2–0.4 29–50
(2019c) 30000
Maghzi et al. 27 2500 1000 1 25 7400–24000 1000–3000 0.1–5 31.8–40.3
(2014)
Lai et al. (2019) 3 74.89–81.18 5.3 0.3 70 5727 1000 0.5 8.12–18.74
Long et al. (2019b) 8 4.18–75.82 15.2 0.5 40 5380 1000 0.05 4.16–23.61
Liu et al. (2017a) 4 213.6–403.5 5.13 0.3 82 39931.6 1500–2000 0.15 9.2–20
Liu et al. (2017b) 4 193–356 10.5 0.3 45 14989.7 1100–1750 0.37–0.48 12–24
Cheraghian and 4 284 1320 1 NR 20,000 3150 0.9 11–19
Khalilinezhad
(2015)
Maghzi et al. 2 NR 85 NR NR NR 1000 0.1 10
(2013)
Liu et al. (2022) 2 49.8–60.8 2 0.3 90 85,000 1000–1500 1.5 2.35–7.82
Elkady (2016) 5 1317–8266 21 NR NR 30,000 1000 0.02 3.4–8.65
Zhu et al. (2014) 2 1497–1498 39.2 0.3 85 32,868 5000 0.5 5.44–10.57
Pu et al. (2015) 3 160.1–168.5 20.6 0.25 70 11,439 1750 0.39 9.9–16.3
Lai et al. (2016) 2 304.3–311.6 NR 0.3 25 9784 1500 0.03 8.85–15.47
Fan et al. (2021) 3 450 315 2 60 6000 NR 0.3 5–10
Maghzi et al. 2 6000 85 NR NR NR 1000 0.02 10
(2011)
Soliman et al. 10 347–688 8.96 0.5–1.3 25 20,000 1000–2000 0.2 3.12–31.85
(2020)
Corredor et al. 3 450–478 60 0.4 56 10,490 750 0.055 19.3–21.5
(2021)
Ye et al. (2013) 2 926.28–937.28 52.5 1.8 65 23,000 2000 0.5 14.22–20.1
Cao et al. (2018a) 2 1492–1513 30.6 0.3 90 9374 2500 0.15 2.8–7.75
Ponnapati et al. 2 380 103 3.4 30 10,000 5000–10000 0.5 7.9–11
(2011)
Hu et al. (2019) 2 1015–1021 NR 0.3 25 4773 1250 0.6 6.56–10.08
Cao et al. (2019) 3 334.41–371.8 19.28 0.3 90 NR 2500 0.0625 5–17.3
Akbar et al. (2020) 2 1710–2000 68 0.6 70 NR 5000 3 26.95–41.04
Akbar et al. (2021) 8 1500–3430 68 0.5 70 NR 5000 3 24.6–50.11
Agi et al. (2020a) 3 102.53–167.43 NR 0.5 120 22,000 2000 0.2 11–26
Agi et al. (2019) 3 102.53–167.43 10 0.5 120 22,000 2000 0.1 11–20
Orodu et al. (2018) 4 245–278.8 4.62 0.3 30 30,000 3000–5000 3 2.63–7.81
Long et al. (2019a) 4 5.64–7.15 15.2 0.5 45 5380 1500 0.05 10.28–21.97
Ali et al. (2021) 8 3.2–4.2 NR 1.5 75 500–5000 1500 0.05 5.3–10.5
El-Hoshoudy et al. 2 45.39 9.28 0.5 NR NR 1500 0.2 20.8–26
(2015)
Hamdi et al. (2020) 3 150 8.13 2 90 30,000 100–500 0.001–0.005 15–19
CHEN et al. (2020) 3 1329–1340 537 0.5 60 10,000 1800 0.2–0.5 22.7–28.7
Cheng et al. (2017) 2 NR 479.8 0.5 80 NR 4999 5 12.61–23.22
NR: Not Reported.

References Agi, A., et al., 2019. The synergy of the flow behaviour and disperse phase of cellulose
nanoparticles in enhancing oil recovery at reservoir condition. PLoS One 14 (9),
e0220778.
Abdulaziz, A.M., Hawary, S.S., 2020. Prediction of carbonate diagenesis from well logs
Agi, A., et al., 2020b. Comparing natural and synthetic polymeric nanofluids in a mid-
using artificial neural network: an innovative technique to understand complex
permeability sandstone reservoir condition. J. Mol. Liq. 317, 113947.
carbonate systems. Ain Shams Eng. J. 11 (4), 1387–1401.
Akbar, I., et al., 2021. Nano-suspension combined effect with polymer gels for enhanced
Abdulaziz, A.M., Mahdi, H.A., Sayyouh, M.H., 2019. Prediction of reservoir quality using
oil recovery in low permeability reservoir. Arabian J. Geosci. 14 (15), 1–16.
well logs and seismic attributes analysis with an artificial neural network: a case
Akbar, I., et al., 2020. Experimental investigation of chemical flooding using
study from Farrud Reservoir, Al-Ghani Field, Libya. J. Appl. Geophys. 161, 239–254.
nanoparticles and polymer on displacement of crude oil for enhanced oil recovery.
Abdullahi, M.B., Rajaei, K., Junin, R., Bayat, A.E., 2019. Appraising the impact of metal-
Int. J. Chem. Eng. 2020.
oxide nanoparticles on rheological properties of HPAM in different electrolyte
Al-AbdulJabbar, A., Mahmoud, A.A., Elkatatny, S., Abughaban, M., 2022. Artificial
solutions for enhanced oil recovery. J. Petrol. Sci. Eng. 172, 1057–1068.
neural networks-based correlation for evaluating the rate of penetration in a vertical
Agi, A., et al., 2020a. Application of polymeric nanofluid in enhancing oil recovery at
carbonate formation for an entire oil field. J. Petrol. Sci. Eng. 208, 109693.
reservoir condition. J. Petrol. Sci. Eng. 194, 107476.
Ali, F., et al., 2022. Predicting saturation pressure of reservoir fluids using machine
learning techniques. Petrol. Sci. Technol. 1–21.

19
K.G. Salem et al.
Ali, J.A., Kolo, K., Manshad, A.K., Stephen, K.D., 2019. Potential application of low-
salinity polymeric-nanofluid in carbonate oil reservoirs: IFT reduction, wettability
alteration, rheology and emulsification characteristics. J. Mol. Liq. 284, 735–747.
Ali, J.A., Kolo, K., Manshad, A.K., Stephen, K.D., 2021. Emerging applications of TiO2/
SiO2/poly (acrylamide) nanocomposites within the engineered water EOR in
carbonate reservoirs. J. Mol. Liq. 322, 114943.
Aliabadian, E., et al., 2020. Application of graphene oxide nanosheets and HPAM
aqueous dispersion for improving heavy oil recovery: effect of localized
functionalization. Fuel 265, 116918.
Alsabaa, A., Gamal, H., Elkatatny, S., Abdulraheem, A., 2021. New correlations for better
monitoring the all-oil mud rheology by employing artificial neural networks. Flow
Meas. Instrum. 78, 101914.
Alvarado, V., Manrique, E., 2010. Enhanced oil recovery: an update review. Energies 3
(9), 1529–1575.
Bengar, A., Moradi, S., Ganjeh-Ghazvini, M., Shokrollahi, A., 2017. Optimized polymer
flooding projects via combination of experimental design and reservoir simulation.
Petroleum 3 (4), 461–469.
Bera, A., Shah, S., Shah, M., Agarwal, J., Vij, R.K., 2020. Mechanistic study on silica
nanoparticles-assisted guar gum polymer flooding for enhanced oil recovery in
sandstone reservoirs. Colloids Surf. A Physicochem. Eng. Asp. 598, 124833.
Cao, J., et al., 2019. Studies on the rheological properties of amphiphilic nanosilica and a
partially hydrolyzed polyacrylamide hybrid for enhanced oil recovery. Chem. Eng.
Sci. 206, 146–155.
Cao, J., et al., 2018a. Application of amino-functionalized nanosilica in improving the
thermal stability of acrylamide-based polymer for enhanced oil recovery. Energy
Fuel. 32 (1), 246–254.
Cao, J., et al., 2018b. Aqueous hybrids of amino-functionalized nanosilica and
acrylamide-based polymer for enhanced oil recovery. RSC Adv. 8 (66),
38056–38064.
Chen, W.-h., Wang, Y.-f., He, Z.-p., Ding, M.-c., 2020. Stability, rheology and
displacement performance of nano-SiO2/HPAM/NaCl dispersion systems. J. Fuel
Chem. Technol. 48 (5), 568–576.
Cheng, Y., et al., 2017. Water-dispersible reactive nanosilica and poly (2-acrylamido-2-
methyl-1-propanesulfonic acid sodium) nanohybrid as potential oil displacement
agent for enhanced oil recovery. Energy Fuel. 31 (6), 6345–6351.
Cheraghian, G., 2016. Effect of nano titanium dioxide on heavy oil recovery during
polymer flooding. Petrol. Sci. Technol. 34 (7), 633–641.
Cheraghian, G., Khalilinezhad, S., 2015. Effect of nanoclay on heavy oil recovery during
polymer flooding. Petrol. Sci. Technol. 33 (9), 999–1007.
Corredor, L.M., et al., 2019a. Heavy oil recovery by surface modified silica nanoparticle/
HPAM nanofluids. Fuel 252, 622–634.
Corredor, L.M., Husein, M.M., Maini, B.B., 2019b. Effect of hydrophobic and hydrophilic
metal oxide nanoparticles on the performance of xanthan gum solutions for heavy oil
recovery. Nanomaterials 9 (1), 94.
Corredor, L.M., Husein, M.M., Maini, B.B., 2019c. Impact of PAM-grafted nanoparticles
on the performance of hydrolyzed polyacrylamide solutions for heavy oil recovery at
different salinities. Ind. Eng. Chem. Res. 58 (23), 9888–9899.
Corredor, L.M., Husein, M.M., Maini, B.B., 2019d. A review of polymer nanohybrids for
oil recovery. Adv. Colloid Interface Sci. 272, 102018.
Corredor, L.M., et al., 2021. Oil displacement efficiency of a silica/HPAM nanohybrid.
Energy Fuel. 35 (16), 13077–13085.
El-hoshoudy, A., 2022. Experimental and theoretical investigation for synthetic
polymers, biopolymers and polymeric nanocomposites application in enhanced oil
recovery operations. Arabian J. Sci. Eng. 1–29.
El-hoshoudy, A., Ahmed, A., Gomaa, S., Abdelhady, A., 2022. An artificial neural
network model for predicting the hydrate formation temperature. Arabian J. Sci.
Eng. 47 (9), 11599–11608.
El-Hoshoudy, A., et al., 2017. Evaluation of solution and rheological properties for
hydrophobically associated polyacrylamide copolymer as a promised enhanced oil
recovery candidate. Egyptian Journal of Petroleum 26 (3), 779–785.
El-Hoshoudy, A., Desouky, S., Betiha, M., Alsabagh, A., 2016. Use of 1-vinyl imidazole
based surfmers for preparation of polyacrylamide–SiO2 nanocomposite through aza-
Michael addition copolymerization reaction for rock wettability alteration. Fuel 170,
161–175.
El-Hoshoudy, A., et al., 2015. Investigation of optimum polymerization conditions for
synthesis of cross-linked polyacrylamide-amphoteric surfmer nanocomposites for
polymer flooding in sandstone reservoirs. International Journal of Polymer Science
2015.
El-hoshoudy, A., et al., 2019. Enhanced oil recovery using polyacrylates/ACTF
crosslinked composite: preparation, characterization and coreflood investigation.
J. Petrol. Sci. Eng. 181, 106236.
Elhaei, R., Kharrat, R., Madani, M., 2021. Stability, flocculation, and rheological
behavior of silica suspension-augmented polyacrylamide and the possibility to
improve polymer flooding functionality. J. Mol. Liq. 322, 114572.
Elkady, M., 2016. Application of nanotechnology in EOR, polymer-nano flooding the
nearest future of chemical EOR. In: SPE Annual Technical Conference and
Exhibition. OnePetro.
Eltoum, H., Yang, Y.-L., Hou, J.-R., 2021. The effect of nanoparticles on reservoir
wettability alteration: a critical review. Petrol. Sci. 18 (1), 136–153.
Fan, G., Li, M., Chen, X., Palyanitsina, A., Timoshin, A., 2021. Polymer-Nanosilica-
assisted to evaluate oil recovery performances in sandstone reservoirs. Energy Rep.
7, 2588–2593.
Firozjaii, A.M., Saghafi, H.R., 2020. Review on chemical enhanced oil recovery using
polymer flooding: fundamentals, experimental and numerical simulation. Petroleum
6 (2), 115–122.
20
Geoenergy Science and Engineering 226 (2023) 211753
Gbadamosi, A., et al., 2022a. Application of polymers for chemical enhanced oil
recovery: a review. Polymers 14 (7), 1433.
Gbadamosi, A., et al., 2022b. Mechanistic study of nanoparticles-assisted xanthan gum
polymer flooding for enhanced oil recovery: a comparative study. J. Pet. Explor.
Prod. Technol. 12 (1), 207–213.
Gbadamosi, A.O., et al., 2019a. Effect of aluminium oxide nanoparticles on oilfield
polyacrylamide: rheology, interfacial tension, wettability and oil displacement
studies. J. Mol. Liq. 296, 111863.
Gbadamosi, A.O., et al., 2019b. Synergistic application of aluminium oxide nanoparticles
and oilfield polyacrylamide for enhanced oil recovery. J. Petrol. Sci. Eng. 182,
106345.
Gbadamosi, A.O., Junin, R., Manan, M.A., Agi, A., Yusuff, A.S., 2019c. An overview of
chemical enhanced oil recovery: recent advances and prospects. Int. Nano Lett. 9 (3),
171–202.
Gbadamosi, A.O., et al., 2018. Recent advances and prospects in polymeric nanofluids
application for enhanced oil recovery. J. Ind. Eng. Chem. 66, 1–19.
Gomaa, S., Emara, R., Mahmoud, O., El-Hoshoudy, A., 2022a. New correlations to
calculate vertical sweep efficiency in oil reservoirs using nonlinear multiple
regression and artificial neural network. Journal of King Saud University-
Engineering Sciences 34 (7), 368–375.
Gomaa, S., Soliman, A.A., Mohamed, A., Emara, R., Attia, A.M., 2022b. New correlation
for calculating water saturation based on permeability, porosity, and resistivity
index in carbonate reservoirs. ACS Omega 7 (4), 3549–3556.
Gouda, A., Khaled, S., Gomaa, S., Attia, A.M., 2021. Prediction of the rheological
properties of invert emulsion mud using an artificial neural network. ACS Omega 6
(48), 32948–32959.
Hakiki, F., Arifurrahman, F., 2023. Cross-linked and responsive polymer: gelation model
and review. J. Ind. Eng. Chem. 119, 532–549.
Hakiki, F., Maharsi, D.A., Marhaendrajana, T., 2015. Surfactant-polymer coreflood
simulation and uncertainty analysis derived from laboratory study. Journal of
Engineering & Technological Sciences 47 (6).
Hamdi, S.S., Al-Kayiem, H.H., Alsabah, M.S., Muhsan, A.S., 2022. A comparative study of
dispersed and grafted nanofluids of graphene nanoplatelets with natural polymer in
high salinity brine for enhanced oil recovery. J. Petrol. Sci. Eng. 210, 110004.
Hamdi, S.S., Al-Kayiem, H.H., Muhsan, A.S., 2020. Natural polymer non-covalently
grafted graphene nanoplatelets for improved oil recovery process: a micromodel
evaluation. J. Mol. Liq. 310, 113076.
Haruna, M.A., et al., 2020. Nanoparticle modified polyacrylamide for enhanced oil
recovery at harsh conditions. Fuel 268, 117186.
Hassan, A., Elkatatny, S., Abdulraheem, A., 2019. Intelligent prediction of minimum
miscibility pressure (MMP) during CO2 flooding using artificial intelligence
techniques. Sustainability 11 (24), 7020.
Helmi, M., Abu El Ela, M., Desouky, S., Sayyouh, M., 2020. Effects of nanocomposite
polymer flooding on Egyptian crude oil recovery. J. Pet. Explor. Prod. Technol. 10
(8), 3937–3945.
Hu, X., et al., 2018. Preparation and properties of nanocomposites of β-cyclodextrin-
functionalized polyacrylamide and its application for enhancing oil recovery. RSC
Adv. 8 (53), 30491–30501.
Hu, X., Zhao, X., Ke, Y., 2019. Effects of silica nanoparticle on the solution properties of
hydrophobically associating polymer based on acrylamide and β-cyclodextrin.
J. Mol. Liq. 296, 111885.
Hu, Y., Zhao, Z., Dong, H., Vladimirovna Mikhailova, M., Davarpanah, A., 2021. Hybrid
application of nanoparticles and polymer in enhanced oil recovery processes.
Polymers 13 (9), 1414.
Hu, Z., Haruna, M., Gao, H., Nourafkan, E., Wen, D., 2017. Rheological properties of
partially hydrolyzed polyacrylamide seeded by nanoparticles. Ind. Eng. Chem. Res.
56 (12), 3456–3463.
Ibrahim, A.F., Elkatatny, S., Al Ramadan, M., 2022. Prediction of water saturation in
tight gas sandstone formation using artificial intelligence. ACS Omega 7 (1),
215–222.
Izadi, N., Koochi, M.M., Amrollahi, A., Pourkhalil, M., 2019. Investigation of
functionalized polyelectrolyte polymer-coated Fe3O4 nanoparticles stabilized in
high salinity brine at high temperatures as an EOR agent. J. Petrol. Sci. Eng. 178,
1079–1091.
Jetter, T., 2014. MemBrain Neural Network Simulator Manual.
Kalam, S., et al., 2021. Data-driven modeling approach to predict the recovery
performance of low-salinity waterfloods. Nat. Resour. Res. 30 (2), 1697–1717.
Keykhosravi, A., Vanani, M.B., Aghayari, C., 2021. TiO2 nanoparticle-induced Xanthan
Gum Polymer for EOR: assessing the underlying mechanisms in oil-wet carbonates.
J. Petrol. Sci. Eng. 204, 108756.
Kumar, S., Tiwari, R., Husein, M., Kumar, N., Yadav, U., 2020. Enhancing the
performance of HPAM polymer flooding using nano CuO/nanoclay blend. Processes
8 (8), 907.
Lai, N., et al., 2019. Feasibility study of applying modified nano-SiO2 hyperbranched
copolymers for enhanced oil recovery in low-mid permeability reservoirs. Polymers
11 (9), 1483.
Lai, N., et al., 2016. Preparation and properties of hyperbranched polymer containing
functionalized Nano-SiO2 for low-moderate permeability reservoirs. Russ. J. Appl.
Chem. 89 (10), 1681–1693.
Lake, L.W., 1989. Enhanced Oil Recovery.
Liu, J., Wang, S., He, L., Hu, P., Gao, Y., 2022. Preparation and properties of nano-silica
hybrid hydrophobic associated polyacrylamide for polymer flooding. J. Petrol. Sci.
Eng. 208, 109434.
Liu, R., Pu, W.-F., Du, D.-J., 2017a. Synthesis and characterization of core–shell
associative polymer that prepared by oilfield formation water for chemical flooding.
J. Ind. Eng. Chem. 46, 80–90.
K.G. Salem et al.
Liu, R., Pu, W., Sheng, J.J., Du, D., 2017b. Star-like hydrophobically associative
polyacrylamide for enhanced oil recovery: comprehensive properties in harsh
reservoir conditions. J. Taiwan Inst. Chem. Eng. 80, 639–649.
Long, Y., et al., 2019a. Swelling mechanism of core–shell polymeric nanoparticles and
their application in enhanced oil recovery for low-permeability reservoirs. Energy
Fuel. 33 (4), 3077–3088.
Long, Y., et al., 2019b. Enhanced oil recovery by a suspension of core-shell polymeric
nanoparticles in heterogeneous low-permeability oil reservoirs. Nanomaterials 9 (4),
600.
Maghzi, A., Kharrat, R., Mohebbi, A., Ghazanfari, M.H., 2014. The impact of silica
nanoparticles on the performance of polymer solution in presence of salts in polymer
flooding for heavy oil recovery. Fuel 123, 123–132.
Maghzi, A., Mohebbi, A., Kharrat, R., Ghazanfari, M., 2013. An experimental
investigation of silica nanoparticles effect on the rheological behavior of
polyacrylamide solution to enhance heavy oil recovery. Petrol. Sci. Technol. 31 (5),
500–508.
Maghzi, A., Mohebbi, A., Kharrat, R., Ghazanfari, M.H., 2011. Pore-scale monitoring of
wettability alteration by silica nanoparticles during polymer flooding to heavy oil in
a five-spot glass micromodel. Transport Porous Media 87 (3), 653–664.
Maurya, N.K., Kushwaha, P., Mandal, A., 2017. Studies on interfacial and rheological
properties of water soluble polymer grafted nanoparticle for application in enhanced
oil recovery. J. Taiwan Inst. Chem. Eng. 70, 319–330.
Minitab, L., 2020. Getting Started with Minitab Statistical Software. Software Manual.
Minitab.
Muggeridge, A., et al., 2014. Recovery rates, enhanced oil recovery and technological
limits. Phil. Trans. Math. Phys. Eng. Sci. 372 (2006), 20120320.
Orodu, K.B., Afolabi, R.O., Oluwasijuwomi, T.D., Orodu, O.D., 2019. Effect of aluminum
oxide nanoparticles on the rheology and stability of a biopolymer for enhanced oil
recovery. J. Mol. Liq. 288, 110864.
Orodu, O.D., Orodu, K.B., Afolabi, R.O., Dafe, E.A., 2018. Dataset on experimental
investigation of gum Arabic coated alumina nanoparticles for enhanced recovery of
nigerian medium crude oil. Data Brief 19, 475–480.
Ponnapati, R., et al., 2011. Polymer-functionalized nanoparticles for improving
waterflood sweep efficiency: characterization and transport properties. Ind. Eng.
Chem. Res. 50 (23), 13030–13036.
Pu, W.-F., et al., 2015. Water-soluble core–shell hyperbranched polymers for enhanced
oil recovery. Ind. Eng. Chem. Res. 54 (3), 798–807.
Rezaei, A., Abdi-Khangah, M., Mohebbi, A., Tatar, A., Mohammadi, A.H., 2016. Using
surface modified clay nanoparticles to improve rheological behavior of Hydrolized
Polyacrylamid (HPAM) solution for enhanced oil recovery with polymer flooding.
J. Mol. Liq. 222, 1148–1156.
Rueda, E., Akarri, S., Torsæter, O., Moreno, R.B., 2020. Experimental investigation of the
effect of adding nanoparticles to polymer flooding in water-wet micromodels.
Nanomaterials 10 (8), 1489.
Ruiz-Cañas, M.-C., Quintero-Perez, H.-I., Castro-Garcia, R.-H., Romero-Bohorquez, A.R.,
2020. Use of nanoparticles to improve thermochemical resistance of synthetic
21
Geoenergy Science and Engineering 226 (2023) 211753
polymer to enhanced oil recovery applications: a review. C.T. F Ciencia, Tecnol.,
Futuro 10 (2), 85–97.
Saboorian-Jooybari, H., Dejam, M., Chen, Z., 2016. Heavy oil polymer flooding from
laboratory core floods to pilot tests and field applications: half-century studies.
J. Petrol. Sci. Eng. 142, 85–100.
Saha, R., Uppaluri, R.V., Tiwari, P., 2018. Silica nanoparticle assisted polymer flooding
of heavy crude oil: emulsification, rheology, and wettability alteration
characteristics. Ind. Eng. Chem. Res. 57 (18), 6364–6376.
Salem, K.G., Gad, K., Abdulaziz, A.M., Aziz, A., Abdel Sattar A Dahab, A.S., 2018.
Prediction of hydraulic properties in carbonate reservoirs using artificial neural
network. In: Abu Dhabi International Petroleum Exhibition & Conference. OnePetro.
Shayan Nasr, M., Shayan Nasr, H., Karimian, M., Esmaeilnezhad, E., 2021. Application of
artificial intelligence to predict enhanced oil recovery using silica nanofluids. Nat.
Resour. Res. 30 (3), 2529–2542.
Sheng, J.J., Leonhardt, B., Azri, N., 2015. Status of polymer-flooding technology. J. Can.
Petrol. Technol. 54 (2), 116–126.
Snosy, M.F., Abu El Ela, M., El-Banbi, A., Sayyouh, H., 2020. Comprehensive
investigation of low-salinity waterflooding in sandstone reservoirs. J. Pet. Explor.
Prod. Technol. 10 (5), 2019–2034.
Snosy, M.F., Abu El Ela, M., El-Banbi, A., Sayyouh, H., 2021. Comprehensive
investigation of low salinity waterflooding in carbonate reservoirs. J. Pet. Explor.
Prod. Technol. 1–24.
Soliman, A.A., El-Hoshoudy, A.N., Attia, A.M., 2020. Assessment of xanthan gum and
xanthan-g-silica derivatives as chemical flooding agents and rock wettability
modifiers. Oil & Gas Science and Technology–Revue d’IFP Energies nouvelles 75, 12.
Sowunmi, A.O., Efeovbokhan, V.E., Orodu, O.D., Oni, B.A., 2022.
Polyelectrolyte–nanocomposite for enhanced oil recovery: influence of nanoparticle
on rheology, oil recovery and formation damage. J. Pet. Explor. Prod. Technol. 12
(2), 493–506.
Standnes, D.C., Skjevrak, I., 2014. Literature review of implemented polymer field
projects. J. Petrol. Sci. Eng. 122, 761–775.
Tavakkoli, O., Kamyab, H., Shariati, M., Mohamed, A.M., Junin, R., 2022. Effect of
nanoparticles on the performance of polymer/surfactant flooding for enhanced oil
recovery: a review. Fuel 312, 122867.
Ye, Z., et al., 2013. Synthesis and performance of an acrylamide copolymer containing
nano-SiO2 as enhanced oil recovery chemical. J. Chem. 2013.
Yousefvand, H., Jafari, A., 2015. Enhanced oil recovery using polymer/nanosilica.
Procedia Materials Science 11, 565–570.
Zahiri, M.G., Esmaeilnezhad, E., Choi, H.J., 2022. Effect of polymer–graphene-quantum-
dot solution on enhanced oil recovery performance. J. Mol. Liq. 349, 118092.
Zheng, C., Cheng, Y., Wei, Q., Li, X., Zhang, Z., 2017. Suspension of surface-modified
nano-SiO2 in partially hydrolyzed aqueous solution of polyacrylamide for enhanced
oil recovery. Colloids Surf. A Physicochem. Eng. Asp. 524, 169–177.
Zhu, D., Wei, L., Wang, B., Feng, Y., 2014. Aqueous hybrids of silica nanoparticles and
hydrophobically associating hydrolyzed polyacrylamide used for EOR in high-
temperature and high-salinity reservoirs. Energies 7 (6), 3858–3871.