International Communications in Heat and Mass Transfer 127 (2021) 105543

Contents lists available at ScienceDirect

International Communications in Heat and Mass Transfer

journal homepage: www.elsevier.com/locate/ichmt

Nanofluid research and applications: A review

Junhao Li a, Xilong Zhang a, *, Bin Xu b, Mingyu Yuan b
a
School of Mechanical and Automotive Engineering, Qingdao University of Technology, Qingdao 266520, China
b
Qingdao WuShun AutoMobile Mould Parts Co. Ltd, 1689 Taifa Road, Huangdao District, Qingdao 266400, China.

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

Keywords: Nanofluids have been widely used in various fields because of their unique thermal conductivity. This paper
Nanofluid reviews and evaluates the thermal conductivity of nanofluids according to the recent literatures on the thermal
Review conductivity of nanofluids. The results showed that in recent years, with the support of multiple disciplines,
Influencing factors
many new nanofluids have been prepared, such as cellulose nanofluids, graphene nanofluids, etc. These nano­
Thermal conductivity prediction model
Nanofluid applications
fluids have even better thermal conductivity. In terms of the prediction of thermal conductivity of nanofluids,
neural network prediction has high accuracy and low time cost, so it is believed that it will become the main
prediction method in the future. Although the experiments and studies of nanofluids have been carried out, there
is no comprehensive review on the thermal conductivity of nanofluids from various aspects. In order to solve this
problem, the research on nanofluids in recent years is reviewed from three aspects: influencing factors, pre­
diction models and applications. The prediction models of thermal conductivity of nanofluids are introduced
emphatically and each prediction model is evaluated. The neural network prediction models in recent years are
discussed in detail. The purpose of this paper is to give a complete and detailed description of the thermal
conductivity of nanofluids, as a reference for researchers to understand and study the thermal conductivity of
nanofluids, and to update the latest research results in a timely manner. It is hoped to point out the direction for
the research of nanofluids in the future.

1. Properties and characteristics of nanofluids alcohols and other liquids. Then Xuan et al. [2] studied the thermal
conductivity of pure metal Cu nanofluids. They found that the heat
Traditional heat transfer fluid media such as water, air, oil, Freon transfer coefficient of the nanofluid increased compared to the base
and so on. They are widely used in chemical production, transportation, liquid pure water. Murshed et al. [3] measured the convective heat
air conditioning, machinery manufacturing and microelectronics tech­ transfer coefficient of the oxide TiO2 nanofluid. They found that the
nology. However, due to the development of industrial technology, convective heat transfer coefficient increased by 30–33% compared
these traditional heat transfer media can no longer meet the ultra-high with the base liquid. Since then, more and more researchers have
heat transfer rate. In order to improve the thermal conductivity of studied the thermal conductivity of nanofluids and also found that the
liquid. They added solid particles of the size of microns and millimeters thermal conductivity of nanofluids is higher than that of base fluids
into the liquid, which significantly improved thermal conductivity, but [4–9].
also caused problems such as easy precipitation and easy plugging. With In recent years, more kinds of nanofluids have been prepared and
the development of technology, people can make solid particles on the more experimental studies have been carried out. Deng et al. [10]
nanometer scale. studied the thermal conductivity of cellulose β-H2O nanofluids. It was
In 1995, Choi et al. [1] from Argonne National Laboratory in the found that the thermal conductivity of the nanofluid was about 36%
United States first proposed the concept of nanofluids. Nanofluid is a higher than that of the base fluid due to the advantages of the cellulose
suspension formed by the uniform distribution of extremely small as a directional thermal conductivity along the chain direction. Dadwal
diameter particles in the base liquid. These particles refer to nano­ et al. [11] coated the surface of nanoparticles with surfactant. The
particles with diameters between 1 nm and 100 nm, which can be thermal conductivity of iron ore nanofluids was measured. They found
polymers, metals, metal oxides, etc. Base liquid refers to water, oil, that the surfactant coated nanofluid had a much higher thermal

* Corresponding author.
E-mail address: [email protected] (X. Zhang).

https://doi.org/10.1016/j.icheatmasstransfer.2021.105543

Available online 26 August 2021

0735-1933/© 2021 Elsevier Ltd. All rights reserved.
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

conductivity than the base fluid, as well as a higher thermal conductivity However, Chopkar et al. [40], Esfe et al. [41], Arani et al. [42] and
than the non-surfactant coated nanofluid. Nese et al. [75] and Wang Esfahani et al. [43] found through experiments that the thermal con­
et al. [12] studied the effect of surfactants on the thermal conductivity of ductivity of nanofluids is inversely proportional to particle size. Yoo
nanofluids. It is found that appropriate surfactants can enhance the et al. [44] studied the thermal conductivity of TiO2, Al2O3, Fe nanofluids
stability of nanofluids and improve the thermal conductivity. However, and found that the main factors affecting the thermal conductivity of
not all surfactants enhance stability. Nanofluids containing cetyl­ nanofluids were the surface area and volume of nanoparticles. Murshed
trimethylammonium bromide, for example, are very unstable. Huminic et al. [3] found that the size and shape of particles also have an impact
et al. [13], Hosseinzadeh et al. [14–17] and Murshed et al. [18] prepared on the thermal conductivity.Later, Wen et al. [45] and Heris [46] et al.
mixed nanofluids and found that the thermal conductivity of mixed also found the same conclusion. Timofeeva et al. [47] believed that the
nanofluids was higher than that of pure nanofluids. Hamze et al. [19] shape of nanoparticles would have an impact on the viscosity of the
studied graphene, a popular carbon nanomaterial, and developed a nanofluid, thus influencing the thermal conductivity. Kim et al. [48]
novel graphene nanofluid. They found that this nanofluid has high proved that the particle shape has an effect on the suspension stability of
controllability, which makes a great contribution to the field of thermal nanofluids, and the suspension stability of nanoparticles is an important
applications of nanofluids. Finally, in order to make nanofluids can be factor to improve the thermal conductivity. Mirmohammadi et al. [49]
used in medicine. Hosseinzadeh et al. [20–22] began to investigate the believed that the surface area ratio of nanofluids is related to the thermal
effects of microorganisms on the properties of nanofluids. conductivity. PBM et al. [50] studied the thermal conductivity of five
With the prominence of the thermal conductivity advantages of different shapes of nanofluids and concluded that the thermal conduc­
nanofluids, nanofluids are widely used in automotive, electronics, solar tivity of cubic nanofluids was higher than that of rod-shaped and
energy, etc. [23,24]. Except for applications in conventional fields. spherical nanofluids, but the stability of cubic nanofluids was poor.
Nanofluids also have important applications in medicine [25,26]. In the
medical field, it is used to transport drugs and destroy cancer cells [27]. 2.2. Effect of volume fraction on thermal conductivity
In the process of oil exploitation, the efficiency of oil exploitation can be
improved [28,29]. Bianco et al. [7] found that the effect of heat transfer enhancement
With wider applications, a number of factors have been found to increased as the volume fraction of nanoparticles increased. Murshed
affect the thermal conductivity of nanofluids, such as particle size and et al. [3] also found that the thermal conductivity of nanofluids
shape, particle concentration, particle aggregation, fluid stability, tem­ increased significantly with the increase of the volume fraction of TiO2
perature and magnetic field, etc. So, they began to experiment and study nanoparticles. Bianco et al. [51] simulated the thermal conductivity of
the influencing factors. And researchers began to explore the mechanism Al2O3-H2O nanofluids through single-phase and two-phase models, and
of abnormal thermal conductivity of nanofluids and put forward related found that the thermal conductivity increased with the increase of
prediction models. From the beginning of considering particle Brownian particle volume concentration. Sebm et al. [52] studied the hydrody­
motion, interface effect, particle cluster and other dynamic mechanisms namics and thermal characteristics of forced convective flow of nano­
and proposing relevant models, to using neural network to predict fluids through numerical simulation. Experimental results showed that
thermal conductivity [30–37], we have learned more and more about the addition of nanoparticles to water-based liquids can significantly
the characteristics of nanofluids, which also provides a larger applica­ increase the thermal conductivity of the liquid, and that the thermal
tion platform for nanofluids in the future. conductivity increases with increasing volume concentration of nano­
particles. Heris et al. [53] measured the thermal conductivity of the
2. Experimental study and analysis of nanofluids CuO-H2O and Al2O3-H2O nanofluids and found that the thermal con­
ductivity increased as the volume fraction increased. And it is speculated
The abnormal increase of thermal conductivity of nanofluids has that each nanofluid system has an optimal concentration to maximize
attracted the attention of researchers. As a result, scientists have thermal conductivity. Namburu et al. [54] prepared two kinds of
embarked on a number of experiments in the hope of exploring the nanofluids, Al2O3-H2O and CuO-H2O. By changed particle concentra­
pattern of anomalous increases in the thermal conductivity of nano­ tion, they found that thermal conductivity further increased with the
fluids. All experiments investigate the properties of the nanoparticles, increase of particle concentration. And the thermal conductivity of
the properties of the base solution, the distribution or aggregation state Al2O3-H2O is greater than that of CuO-H2O. However, this increase does
of the nanoparticles in the base solution and the factors affecting the not always exist. As the thermal conductivity increases with volume
distribution or aggregation state of the nanoparticles in the base solu­ fraction to a certain value, the thermal conductivity begins to decrease
tion. Recent nanofluid experiments are summarized below in seven until it reaches that of the base liquid [55].
areas and the factors that affect the thermal conductivity of nanofluids
are summarized.
2.3. The effect of temperature on thermal conductivity

2.1. Influence of size and shape of nanoparticles on thermal conductivity The effect of temperature on the thermal conductivity of nanofluids
has been observed by many researchers. Mirmohammadi et al. [49]
Particle size is the main factor that directly affects the thermal con­ found that thermal conductivity increased with temperature. Using
ductivity of nanofluids and is also the standard to distinguish nanofluids sensors, Kakavandi et al. [56] also observed that the thermal conduc­
from other solid-liquid mixtures. Particle size has certain influence on tivity increased with the increase of temperature. Haifeng et al. [57]
the stability of liquid. proposed that increasing the temperature within a certain range would
Beck et al. [38] found that the thermal conductivity of nanofluids increase the interfacial thermal resistance and lead to a decrease in
decreased with the decrease of the size of nanoparticles, which was due thermal conductivity. However, at higher temperatures, the thermal
to the phonon scattering at the solid-liquid interface. Liu et al. [39] conductivity of convection induced by Brownian motion of nano­
found that the interfacial thermal resistance would be affected with the particles is much greater than that of base liquid. SayantanMukherjee
increase of particle size. Because the area of the solid-liquid interface et al. [58] reviewed recent studies on temperature and thermal con­
decreases with the increase of particle size, the thermal conductivity ductivity of nanofluids and reached a general conclusion that Brownian
increases. motion of particles is easily affected by temperature changes and

2
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

thermal conductivity increases with temperature increases. However, improve the thermal conductivity, ensure the stability of the nanofluid,
the results showed that the volume fraction also has a great effect on the and make the agglomeration of nanoparticles more uniform. Some
thermal conductivity [59]. methods have been adopted in the preparation of nanofluids: the use of
dispersants can reduce the surface tension of particles and prevent
2.4. Effect of aggregates on thermal conductivity excessive aggregation of particles, thus improving stability; Ultrasonic
oscillation is also a method to enhance stability, but the oscillation time
Sedighi et al. [60] found that particle aggregation has a positive ef­ is too long, but will make the liquid temperature rise, promote particle
fect on improving heat transfer efficiency. Wei et al. [61] and Sub­ agglomeration. [78].
ramaniam [62] proposed that the aggregation morphology of The ideal state of the nanofluid is that the nanoparticles are uni­
nanoparticles may affect the fractal dimension and thus have a certain formly distributed in the base solution, but because the nanoparticles
influence on the thermal conductivity of nanofluids. Suleimanov et al. have a lot of surface energy, they are extremely easy to form polymers.
[63] showed through experiments that the enhancement of thermal When particles are adsorbed together to form aggregates, the stability of
conductivity of nanofluids was inversely proportional to particle ag­ nanofluids will be destroyed and the characteristics of nanofluids will be
gregation rate. The radius of nanoparticle clusters has a strong influence lost [79].
on thermal conductivity, but the enhancement of thermal conductivity Li et al. [80] showed experimentally that non-ionic dispersants can
will weaken when precipitation begins to appear. Song et al. [64] pro­ significantly improve the stability and fluidity of BN-C2H4O nanofluids
posed a prediction model regarding the aggregation of nanoparticles. It and can maintain enhanced thermal conductivity. However, the stability
was showed that the aggregation of nanoparticles increases the thermal of anionic and cationic dispersants deteriorates rapidly, thus reducing
conductivity. The aggregation effect is stronger at smaller particle sizes the thermal conductivity.
and lower particle concentrations. The main reason for the enhancement Li et al. [68], Wang et al. [69] and Cacua et al. [72] studied the in­
of thermal conductivity is that the agglomeration of particles increases fluence of surfactant and pH value on the thermal conductivity of
the degree of freedom of phonons in the solid phase. Ding et al. [65] took nanofluids. The results showed that when the pH value was too high,
into account the effects of shear-induced and viscosity gradient-induced nanoparticles reached the equipotential point (IEP), which could
particle migration and established a model for predicting the concen­ improve the stability of the liquid and increase the thermal conductivity
tration of nanoparticles. The results showed that the concentration of with the use of surfactant.
nanoparticles in the pipeline is not homogeneous. The distance between
particles in different concentration regions creates a different equilib­ 2.7. Effect of magnetic field on thermal conductivity
rium between the particles, which affects the aggregation state of the
particles [66]. It is very important to study the heat transfer of nano­ Younes et al. [71] found through microscopy that the neutral of
fluids in the future. aggregates of Fe2O3 nanofluids formed in the presence of an external
Motevasel et al. [67] compared prediction models with and without magnetic field was enhanced. Jaykumar et al. [81] observed nanofluids
particle aggregation. The average deviation of the model without par­ with applied magnetic fields through optical microscopy. It is found that
ticle aggregation at low concentrations was found to be 2–6 times higher the long chain structure is formed at low concentration, while the
than that of the model with particle aggregation. This also suggests that segment structure is formed at high concentration, and it takes time to
particle aggregation at low concentrations is an important factor in the form this structure. The thermal conductivity of the nanofluid increases
thermal conductivity. with the increase of the time of external magnetic field. Altan et al. [82]
prepared two types of Fe3O4-H2O and Fe2O3-Heptane nanofluids and
2.5. The influence of pH on thermal conductivity applied a vertical magnetic field externally. The measurement results
showed that the thermal conductivity of the nanofluids increased by
Li et al. [68], Wang et al. [69], Wamkam et al. [70], Younes et al. 5.2% and 2.8%, respectively, at a magnetic field strength of 0.2 T.
[71] and Cacua et al. [72] studied the relationship between pH value Parekh and Lee [83] prepared magnetite and Mn-Zn-Fe oxide two
and thermal conductivity of nanofluids. The results showed that in a different magnetic nanofluids, and applied a transverse magnetic field
certain range, the thermal conductivity of the nanofluid increases with outside, and found that the maximum thermal coefficient of magnetite
the increase of pH value. However, pH value will have an impact on the nanofluids increased by 30%, but Mn-Zn-Fe oxide nanofluids did not
Zeta potential of nanoparticles. When the potential reaches the equi­ respond to the magnetic field. The reason for this may be that the
potential point (IEP), the repulsive force of nanoparticles is almost zero, magnetization of magnetite nanofluids is higher than that of Mn-Zn-Fe
and the particles will gather together, resulting in poor stability of the oxide nanofluids, and the magnetite nanofluids are more likely to
liquid and lower thermal conductivity. form chain-like structures under the action of magnetic field. Katiyar
et al. [84] prepared Fe, Ni, Co nanofluid and measured the thermal
2.6. Effect of additives and stability on thermal conductivity conductivity data of the nanofluid in an external magnetic field. The
experimental results showed that the thermal conductivity of Fe-oil
Leong et al. [73], Xia et al. [74], Nese et al. [75], Almitani et al. [76] magnetic nanofluid was increased by 284% compared with the base
prepared cationic surfactants: cetyltrimethylammonium bromide fluid at a magnetic field of 0.05 T and a particle concentration of 7.0 vol
(CTAB). Anionic surfactants: sodium dodecyl sulfate (SLS) and sodium %. In the experiment, they observed that the nanoparticles formed
dodecyl sulfate (SDS). Non-ionic surfactants: polysorbate 20 (PS20) and chains in the direction of the magnetic field, providing a channel for heat
polyvinylpyrrolidone (PVP). And they explored the relationship be­ transfer, so the magnetic nanoflow would increase thermal conductivity
tween these surfactants and the thermal conductivity of nanofluids. The in the presence of an external magnetic field.
experimental results showed that the nanofluids will agglomerate
themselves after a long time, which reduces the stability of the nano­ 3. Prediction model of thermal conductivity of nanofluids
fluids. The addition of surfactants can maintain the stability of the
nanofluids and increase the thermal conductivity. The thermal con­ The traditional heat transfer process of liquid part is mostly carried
ductivity increases with the increase of temperature and concentration out by convection, and the thermal conductivity of fluid depends on the
of the active agent. The anionic surfactant had the best effect and the thermal conductivity of liquid to a large extent. When people are looking
non-ionic surfactant had the worst effect. Tamjid et al. [77] found that as for heat transfer medium with high thermal conductivity, because it is
the volume fraction of particles increases, the interaction force between found that adding solid particles into liquid can improve the thermal
particles also increases. In order to improve the volume fraction, conductivity of liquid, more and more people begin to pay attention to

3
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

this field. this model. However, the disadvantage of this model is that it does not
Nanofluid, as a new heat transfer medium, has been widely used in a take the interaction Therefore, the prediction results of this model are
variety of fields in recent years. Nanofluids have attracted much atten­ not accurate in the case of large particle concentrations.
tion and a lot of research in recent years. As the addition of nanoparticles Bruggeman proposed a symmetric model (Eq. (5)). This model takes
into liquids significantly improves the thermal conductivity of liquids, it into account the intermolecular interactions. For high concentrations of
also shows the wide application prospect of nanofluids in the field of solid-liquid mixtures, the model has a higher prediction accuracy than
thermal dynamics. In order to make better use of nanofluids, it is the Maxwell model. While at low concentrations, the difference between
necessary to clearly study the mechanism of enhancing heat transfer of the two models is similar.
nanofluids. In recent years, more and more researchers have studied the Rayleigh’s model solves the problem that Maxwell’s model does not
mechanism of enhancing heat transfer of nanofluids. Li and Xuan [85] consider the interaction between molecules, and proposes a new pre­
made a qualitative analysis of the enhanced heat transfer mechanism of diction model under the premise of ideal state (Eq. (3)).
nanofluids, and divided the enhanced heat transfer mechanism into The above three models were generally calculated and applied by
static and dynamic ones. The static mechanism is mainly due to the researchers at the early stage of the study. With the improvement of the
nanoparticle itself. These include the size of the nanoparticles and the industrial level, nanoscale solid particles were produced, and people
macroscopic thermal conductivity effect where the liquid contacts the found that the prediction results of the models were no longer accurate.
surface of the particles to conduct heat transfer. The dynamic mecha­ Li [87] calculated the thermal conductivity of copper-water nanofluid
nism is the effect on the thermal conductivity is due to the extremely through three models and compared it with the experimental data, and
small size of the nanoparticles, which are subjected to intermolecular found that the experimental data was much larger than the data calcu­
forces, Brownian forces, electrostatic forces, etc. in the liquid. The lated by the models. Zhu [88] also conducted the same control experi­
micro-motions such as Brownian motion and thermal diffusion per­ ment with copper-ethylene glycol nanofluids, and found that the
formed by the nanoparticles enhance the thermal conductivity of the experimental results were also greatly different from the results calcu­
liquid. Therefore, in order to clarify the mechanism of heat transfer lated by the model. However, the validity of the above models for pre­
enhancement of nanofluids, we must explain it from two aspects. The dicting solid-liquid mixtures at the millimeter level has been
mechanism of heat transfer enhancement by nanofluids is reviewed from demonstrated. We can therefore know that it is not sufficient to explain
the following two aspects. the thermal conductivity of nanofluids by the thermal conductivity of
the base fluid, the thermal conductivity of the added particles and the
3.1. Static mechanism volume share of the particles.
Heat transfer in nanofluids is carried out on the surface of liquid and
Due to industrial technology, the solid particles added to the liquid in solid particles, so the surface area and surface shape of particles should
the early manufacture of solid mixtures were only in the micron and also be important factors affecting thermal conductivity. Hamiltion and
millimeter scales. Based on the conjecture principle that the thermal Crosser improved Maxwell’s model and proposed a new prediction
conductivity of solid is generally higher than that of liquid, the addition model (Eq. (4)) on the basis of considering the influence of particle
of solid changes the structure of liquid, so the thermal conductivity of surface shape on thermal conductivity. According to the Hamilon-
the solid-liquid mixture formed increases. Crosser (H–C) model, Li calculated the thermal conductivity of spher­
In the early stage of the study, people calculated the thermal con­ ical Al2O3-H2O nanofluids as 0.3–1.0. Compared with the results of
ductivity keff of the mixture from the relationship among the liquid Eastman experiment in 1997, it is found that under the condition of a
thermal conductivity kf, the thermal conductivity kp added to the solid certain volume share, changing the sphericity of solid particles has a
particles, and the volume share of the particles φ. The early prediction significant effect on the thermal conductivity. Although there is still a
models of the solid-liquid thermal conductivity were all defined based certain deviation in the comparison results, it further shows that the
on Eq. (1). On the basis of Eq. (1), Maxwell, Rayleigh, Jeffrey et al. put shape of solid particles has a great impact on the thermal conductivity of
forward the earliest prediction models of thermal conductivity (Eqs. (2), the mixture. However, the H–C model only further reveals that particle
(3), and (4)). shape has a strong influence on thermal conductivity, not that it does not
Maxwell proposed a thermal conductivity model (Eq. (2)) for pre­ take particle size into account. Thus, this model is only applicable to
dicting spherical particle mixtures. As the earliest prediction model, this mixtures of large particles. Therefore, it is still unable to accurately
model made a relatively accurate prediction at that time. Even in recent calculate the thermal conductivity of nanofluids.
studies, Adi et al. [86] used the Maxwell model to predict the thermal Sebm et al. [52] used TiO2-H2O nanofluid to conduct experiments.
conductivity of Al2O3-H2O nanofluids, and the results showed that the They measured the thermal conductivity of the nanofluid by changing
predicted results were basically consistent with the experimental results. the particle size and shape of the nanoparticles and compared it with the
And many researchers are now continuing to optimize on the basis of calculated data of the H–C model. They found that the conductivity of

Table 1
The classical thermal conductivity prediction models and their important parameters.
Number Prediction model of thermal conductivity The key parameters

1 kp ϕ(dt/dx)p + kf (1 − ϕ)(dt/dx)f
keff =
ϕ(dt/dx)p +( (1 − ϕ)(dt/dx)
) f
2 keff kp + 2kf − 2ϕ kf − kp keff、kp 、kf 、ϕ
= ( )
kf ⎡kp + 2kf + ϕ kf − kp ⎤
3 kp − kf
keff 、kp 、kf 、ϕ、β =
⎢
⎢
⎥
⎥ kp + 2kf
⎢ ⎥
keff ⎢
⎢1 + 3φβ ⎥
⎥
=⎢ ⎛ ⎞
kf ⎢
⎥
⎥
⎢ kp − k f ⎟ 10/3 ⎥
⎣ 1 − φβ − 0.525⎜ ⎝ ⎠βφ ⎦
4
kp + kf
(3 )
4 keff kp + (N − 1)kf − (N − 1)ϕ kf − kp 3
= ( ) keff 、kp 、kf 、ϕ、N =
k(
f kp )
+ (N − 1)kf(+ ϕ kf − k)p ψ
5 kp − keff kf − keff keff 、kp 、kf 、ϕ
ϕ + (1 − ϕ) =0
kp + 2keff kf + 2keff

4
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

the nanofluid is different between the rod-like nanoparticles and the interactions and collisions, fluid mixing fluctuations and turbulence
spherical nanoparticles; And it is found that the calculated result is enhancement, dispersion of nanoparticles.
smaller than the experimental result. Yoo et al. [89] also used the H–C Through the above research, we know that the influence of nano­
model to compare with the experimental results, which showed that the particle movement must be considered when analyzing the mechanism
H–C model could not predict the thermal conductivity of nanofluids, of heat transfer enhancement of nanofluids. It is summarized from four
and other factors such as the stability of the suspension of nanoparticles, aspects as follows.
the size of nanoparticles and the viscosity of the base liquid should be
taken into account. 3.2.1. Interface effect
The experimental results of Yoo et al. [89] showed that the above Yu et al. [92] found that in nanofluids, the liquid near the surface of
prediction models of thermal conductivity of solid-liquid mixture can no the nanoparticles forms a solid-like layered structure, which is also
longer satisfy the calculation of thermal conductivity of nanofluids nanoscale in thickness. Xue et al. [93] also observed that in nanofluids,
(Table 1). due to the interfacial effect, the surface of nanoparticles will form a
Because the traditional formulas for calculating the thermal con­ network of solid structures, which may play an important role in the
ductivity of solid-liquid mixtures only focus on the volume fraction of heat conduction process of nanofluids [90,94,95]. Therefore, the re­
solid particles and the shape of the solid, and assume that the solid and searchers proposed a new prediction model (Table 2).
liquid are heat transfer by diffusion. The traditional thermal conduc­ Yu et al. [95] modified Maxwell’s model and proposed a new model
tivity calculation models cannot better explain the mechanism of heat (Eq. (6)). This model takes the nanometer layer as a factor affecting the
conduction of nanofluids, which indicates that some other heat transfer thermal conductivity and can predict the thermal conductivity of the
mechanisms have been neglected. All the above research models study nanofluid with the thickness of the nanometer layer less than 10 nm. It is
the heat transfer mechanism of nanofluids from a macroscopic assumed that the nanometer layer around each particle can form a new
perspective. However, due to the small particle size of nanofluids, the equivalent particle with the particle to transform the Maxwell model. It
heat conduction mechanism of nanofluids is different from that of or­ can be seen from Fig. 1 that the improved model is more accurate than
dinary solid-liquid mixtures. Therefore, the above calculation models the Maxwell model, and it has a higher coincidence degree with the
cannot be simply applied to calculate the thermal conductivity of experimental data in the case of small particle size. However, in the case
nanofluids. Not only the thermal conductivity of liquid, the thermal of large particle size, the improved model is not much different from
conductivity of solid particles, the volume share and the shape and Maxwell model, so for the case of large particle size, this model is still
particle size of particles should be considered, but also the influence of unable to accurately predict the thermal conductivity of nanofluids.
the micro motion caused by the small size effect of nanoparticles on the Xie et al. [96] studied the relationship between the nanometer layer
enhancement of heat transfer from the micro point of view. and the thermal conductivity of nanofluids, and proposed a new pre­
diction model for the thermal conductivity of nanofluids (Eq. (7)) on the
basis of considering the interface effect. This model is based on the
3.2. The dynamic mechanism general solution of the heat conduction equation in spherical co­
ordinates and the new model of the microstructure of the mixture on the
Because the traditional macroscopic models cannot fully explain the equivalent spherical model. It is proved that the thickness, size, volume
heat transfer mechanism of nanofluid, people start to study it from other fraction, particle size, particle size and the thermal conductivity of the
aspects. Due to the small size effect of nanoparticles, particles in liquid fluid have an effect on the thermal conductivity of the nanofluid. The
are affected by intermolecular atomic force, Brownian force and elec­ advantages of this model apply not only when the nanometer layer is the
trostatic force, Brownian motion of nanoparticles, thermal diffusion, main factor. When the nanometer layer is no longer the dominant factor,
micro-motion and so on. Therefore, these micro-motions may play a the model’s prediction is still valid. It can be applied to the prediction of
more important role in enhancing heat transfer. effective thermal conductivity of various mixed systems. The thermal
In order to further explain the effect of micro-motion on the thermal conductivity of the nanofluid increases with the increase of the thickness
conductivity of nanofluids, Keblinski et al. [90] and Eastman et al. [91] of the nanolayer. As can be seen from Fig. 2, compared with the
proposed four possible reasons: Brownian motion of nanoparticles, experimental data of three nanofluids with different thermal conduc­
nano-layers of nanoparticles produced by interface effects, heat transfer tivity, the accuracy of the prediction model is still very high.
properties in nanoparticles, and the effect of nanoparticle aggregation. Xue et al. [93] proposed a new prediction model (Eq. (8)). In this
Li [87] also proposed four possible reasons for the increase in model, the influence of boundary effect and particle diameter is
effective thermal conductivity of nanofluids: Suspended nanoparticles considered, by considering nanoparticles and boundary layers as a
lead to increased surface area and fluid thermal conductivity, particle

Table 2
The prediction models of thermal conductivity considering interface effect and their important parameters.
Number Prediction model of thermal conductivity The key parameters
( ) [ ]
6 kpe + 2kf + 2 kpe − kf (1 − θ)3 ϕ 2(1 − γ) + (1 + θ)3 (1 + 2γ) γ
keff = kf h
( )
kpe + 2kf − kpe − kf (1 − θ)3 ϕ kpe = kp 、θ =
− (1 − γ) + (1 + θ)3 (1 + 2γ) rp
7 keff − kf 3Θ2 ϕ2t 4 ( )3
= 3Θϕt + ϕt = π rp + h n、Θ
( kf ) 1 − Θϕt 3 ( )
8 ϕ keff − kf ϕ keff − kpe 2kp + kf + 2∂ kf − kp ( r )3
p
1− + =0 kpe = kp ( ) 、∂ =
∂ ( 2keff +)kf [∂ 2keff + kpe ] ( ) [ ( ) ] 2kp + kf − ∂ kf − kp h+r
9 kp − k1 ϕ1 k1 2β31 − β32 + 1 + kp + 2k1 β31 ϕ1 β32 k1 − kf + kf θ
keff = β1 = 1 + 、β2 = 1 + θ
( ) 3 ( ) [ 3 3
k[p + 2k1 β1 − ] kp (− k1 ϕ1 β)1 +[ β2 − ( 1
] 2
( ) ) ]
10 kp − k1 ϕ1 k1 2β31 − β32 + 1 + kp + 2k1 β31 ϕ1 β32 k1 − kf + kf c、ρ、|V|
keff = ( ) 3 ( ) [ 3 ] + c(ρc)nf |V|ϕdp
kp + 2k1 β1 − kp − k1 ϕ1 β1 + β32 − 1 ( )
11 ( ) [ ] ( ) [ ( ) ] 6ϕ α、T、μeff
kp − k1 ϕ1 k1 2β31 − β32 + 1 + kp + 2k1 β31 ϕ1 β32 k1 − kf + kf π
keff = ( ) 3 ( ) [ 3 ] + ( )
kp + 2k1 β1 − kp − k1 ϕ1 β1 + β32 − 1 αeff ueff h
dp +
( ) k f kp T 4 π r(r + δ)k1 ∫ √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
12 keff kpe + kf (N − 1) + (N + 1) kpe − kf (1 + C)φ 1 π
= ( ) φ、N、kpe = k2pez sin2φ + k2pex cos2φ dφ
kf kpe + kf (N − 1) − kpe − kf (1 + C)φ π 0

5
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

Fig. 1. Comparisons of predicted data and experimental data considering interface effects are drawn from [69,71–76].

6
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

7
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

Fig. 1. (continued).

Fig. 1. (continued).

whole, the concept of complex nanoparticles is proposed. By comparing comparing the model’s prediction with the experimental data. However,
the calculated data of the model with the experimental data of thermal the effect caused by the micro motion is not only the nanometer layer.
conductivity of CuO-H2O and CuO-EG nanofluids, it is found that the Xue et al. [101] studied the interfacial effects using molecular dynamics
particle size and volume fraction of nanoparticles have an influence on simulations and showed that the liquid on the surface of the particles
the nanometer layer. It can be observed from Fig. 3 that this model is had no effect on the heat transfer characteristics perpendicular to and
more consistent with the experimental data than the previous model. parallel to the particles. This means that the heat transfer caused by the
Leong et al. [97] considered the influence of interface effect on the nano-layer generated by the interface effect may not be enough to
thermal conductivity of nanofluids and assumed that the temperature explain the thermal conductivity mechanism of the nano-fluid, so other
fields of particles, interface layer and liquid were continuous at the influencing factors need to be studied.
interface boundary, but the temperature gradient was discontinuous at
the interface boundary. It can be seen from Fig. 4 that the predicted data 3.2.2. Brownian motion
of this prediction model (Eq. (9)) are highly consistent with most As early as the 1990s, Xuan et al. [102] proposed the idea that the
experimental data, and the predicted results are better when the particle thermal conductivity of nanofluids should depend on the microscopic
size is larger and the volume fraction is lower. In addition, the model motion (Brownian motion and interparticle force). Brownian motion is
also proves that the thermal conductivity of the nanometer layer is about also the most widely studied influencing factor, which causes many
2–3 times that of the base liquid, and the thickness of the nanometer researchers to study it. Koo et al. [103] compared the effects of Brow­
layer is about 1 nm. nian motion, affinitive motion and osmotic motion on the effective
Sitprasert et al. [98] modified the model of Leong et al. [97] (Eq. thermal conductivity of nanofluids. It was found that Brownian motion
(10)) to take into account the thermal dispersion effect caused by the contributed more to the heat transfer of the nanofluid than thermo­
Brownian motion of nanoparticles. They believe that the dynamic model phoresis and osmotic motion. Patel et al. [104] found that at very low
of Leong et al. overpredicted the thermal conductivity of nanofluids, and particle concentrations, human thermal conductivity of nanofluids
that the increase in the thermal conductivity of the flowing nanofluids could still be improved by 9% compared with base fluids, but this result
may be more due to the interface effect. It can be seen from Fig. 5 that could not be predicted by previous formulas. Therefore, it is speculated
the modified model is more accurate than Leong’s model, and can reflect that the Brownian motion of nanometer particles may be an important
the change of thermal conductivity of nanofluids with volume fraction reason for the increase of thermal conductivity. Azizian et al. [105]
under different particle sizes. prepared nano-fluids with large particle size of alumina to eliminate the
Alipour et al. [99] also modified Leong’s model (Eq. (11)) based on influence of nano-layers and aggregates on thermal conductivity,
Leong’s model [97]. In the part of dynamic models, the influence of proving that Brownian motion is the main factor affecting thermal
Brownian motion was taken into account, and a variety of factors were conductivity.
considered to make the models more accurate. Xuan et al. [106] took the Brownian motion of nanoparticles into
Yang and Xu [100] modified the H–C model and proposed a new account on the basis of Maxwell’s model and modified the original
model to predict the thermal conductivity of carbon nanotubes consid­ model to obtain a new prediction model (Eq. (13)). This model simulates
ering the interface effect (Eq. (12)). This model has a good accuracy in the random movement and aggregation process of nanoparticles
predicting the thermal conductivity of carbon nanotubes. It is proved through Brownian motion theory and diffusion limited aggregation
that the thickness of the interfacial layer has little effect on the thermal model, and represents the aggregation structure through size. It can be
conductivity within the diameter of 25 nm - 100 nm. In addition, Fig. 7 seen from Fig. 8 that this model has a high degree of coincidence with
reflects that this model is closer to the experimental data than the other experimental data, and that the aggregation process is an important
models. influencing factor. Although this new model takes into account the in­
The above models mainly study the influence of the nanometer layer fluence factor of temperature, the data coincidence degree was not high
on the thermal conductivity, and the reliability of the model is proved by in the experiment with Das et al. [107].

8
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

Fig. 2. Considering the comparison between the Brownian motion prediction data and the experimental data, the figures are from [82,84–90].

9
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

Fig. 2. (continued).

10
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

Fig. 2. (continued).

Wang et al. [108] modified Maxwell model and obtained a new high concentrations of particles. The advantage is that the model can be
model (Eq. (14)). This model based on effective medium theory and further modified by changing the model constant C.
fractal theory, considering the size effect and surface adsorption of The model proposed by Bhattacharya et al. [110] (Eq. (16)) takes
nanoparticles on the impact of a fractal model was proposed to predict into account the conductivity of the base liquid and the conductivity of
the effective thermal conductivity of nano particle clusters, this model the nanoparticles. It can be seen from Fig. 11a and Fig. 11b that
an accurate simulation of the particle diameter of 50 nm CuO-H2O co­ compared with H–C model, the thermal conductivity of copper and
efficient of thermal conductivity of nanofluids, see from Fig. 9 that the aluminum nanofluids is predicted, and the accuracy of this model is
experimental data and predicted data with high contact ratio, and it is higher.
proved that the surface adsorption is the factor that affects the thermal Koo et al. [111,112] and Bao et al. [113] respectively modified
conductivity. However, the calculation of this formula is relatively Maxwell’s model and proposed two kinds of prediction models (Eqs.
complex, and it is only more accurate in the case of small particle size, so (17) and (18)). Eq. (17) considers including particle size, particle vol­
further improvement is needed. ume fraction and temperature dependence, as well as the influence of
Hemanth et al. [109] derived a new model from Stokes-Einstein fundamental fluid and particle on Brownian motion. It can be seen that
formula (Eq. (15)) on the basis of considering Brownian motion. This the first part of Eq. (17) is Maxwell model, and the second part is the
model took into account the influence factors of particle size, particle formula added by considering Brownian motion. Eq. (18) proposes a
concentration and temperature, and solves the problem that Maxwell new prediction model based on dynamics theory. It can be seen from
model cannot explain the thermal conductivity changes with tempera­ Figs. 12 and 13 that the accuracy of prediction of these two models is
ture. As can be seen from Fig. 10, this model has a high agreement with high.
the experimental results, indicating that the higher the temperature, the Dong et al. [114] proposed A new model (Eq. (19)), which was based
higher the thermal conductivity. However, this model is not suitable for on the generalized linear solution of temperature distribution. By

11
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

comparing with the thermal conductivity coefficient of CuO-H2O and applicable to the prediction of thermal conductivity of nanofluids at all
ZnO-TiO2 nanofluid, it was found that the prediction result was very concentrations.
good. If the accurate radius of Brownian motion could be obtained, the Deepak et al. [117] also proposed a prediction model (Eq. (21)) for
prediction accuracy of this model could be further improved. It can be the thermal conductivity of nanofluids based on the phenomenon of
seen from Fig. 14 that the difference between the experimental data and particle clusters based on PSD. This model has advantage by particle size
the predicted data is not large, and it reflects the fact that the thermal to display the particle clusters, compared with the fractal theory, the
conductivity increases with the increase of temperature. model is easier to calculate the data, and this model also takes into ac­
Brownian motion, as an important factor that has been verified to count the temperature, the particle property and time influence, see
affect the increase of thermal conductivity of nanofluids, has different Fig. 16 to see forecast data agrees well with the experimental data, but as
effects on the thermal conductivity under different heat transfer condi­ the particle size increases prediction results with the experimental data
tions. But other researchers said the opposite. Evans et al. [115] pro­ have certain deviation.
posed that the contribution of Brownian motion to the thermal Liu et al. [39] established a prediction model for the thermal con­
conductivity of nanofluids is very small and cannot be used to explain ductivity of nanofluids considering the cluster effect and interfacial
the mechanism of the enhancement of thermal conductivity of nano­ thermal resistance (Eq. (22)). It can be seen from Fig. 17 that the thermal
fluids. They also demonstrated this using molecular dynamics simula­ conductivity of nanofluid clusters increases with the increase of cluster
tions and effective dielectric mechanics. But since they confined their size, which is consistent with the model reaction that the thermal con­
discussion to stationary fluids, the generality of the results remains to be ductivity increases with the increase of cluster size and the increase of
verified (Table 3). cluster shape factors. In addition, it is also found that if the volume
fraction of nanoparticles is less than 0.1%, or the particle size is less than
3.2.3. Influences of nanoparticle clusters 10 nm and there is no cluster phenomenon, then the nanometer layer is
Some researchers hold opposite opinions on the theory that Brow­ the main factor affecting the thermal conductivity (Table 4).
nian operation can improve the thermal conductivity of nanofluids. Gao
et al. [116] found that Brownian motion of nanoparticles was not the 3.2.4. Heat transfer properties in nanoparticles
main cause of thermal conductivity enhancement. Instead, nanoparticle Li et al. [59], Murshed et al. [119], proposed that the enhancement of
aggregation appears to be a key factor [113,117]. Since then, many thermal conductivity of nanofluids is related to volume temperature,
researches have begun to study the thermal conductivity relationship and the thermal conductivity increases with the increase of temperature,
between the agglomerates and nanofluids, and it has been found that the but the dependence of thermal conductivity on temperature is lower
clustering of nanoparticles can significantly improve the thermal con­ than that of volume fraction. Because the relationship between thermal
ductivity of nanofluids. conductivity and temperature is not clear, they proposed that the in­
Zhou and Wu [118] characterized clusters in nanofluids from the crease in temperature may increase the Brownian velocity of nano­
perspective of particle size distribution (PSD) of nanofluids, and pro­ particles. Based on this theory, Kumear et al. [109] considered
posed a model to predict the thermal conductivity of nanofluids (Eq. temperature as a factor when establishing the prediction model with
(20)). This model is modified from Maxwell model on the basis of PSD, Brownian motion as the main influencing factor. Through the prediction
and the main modified parameter is the volume fraction ϕpm of model, it can be showed that the thermal conductivity increases with the
elementary particles. It can be seen from Fig. 15 that the prediction increase of temperature. Pasrija et al. [120] proposed that temperature
results of this model have a high similarity with the experimental results, affected the interface effect, resulting in the thermal conductivity of
with a deviation between − 5% and − 2%. However, this model is nanofluids changing with temperature. By modifying the model pro­
modified on the basis of Maxwell model, which still does not take into posed by Leong et al. [97], it was found that this model could predict the
account the influence of particle interaction, and this model is not thermal conductivity of nanofluids changing with temperature.
However, Ahmed et al. [121] found that the interface thermal
resistance increased with the increase of temperature, leading to a
Table 3
decrease in the thermal conductivity of nanofluids at high temperatures.
The thermal conductivity prediction models and their important parameters are
considered in Brownian motion. Some researchers also measured the thermal conductivity of nanofluids,
and the results showed that the thermal conductivity had nothing to do
Number Prediction model of thermal The key parameters
with temperature, which was in conflict with the above conclusion.
conductivity
Therefore, the mechanism between temperature and thermal conduc­
13 ( )
keff
=
kp + 2kf − kf − kp ϕ
( ) +
η、kB、rc
tivity needs to be further studied.
kf kp + 2kf + kf − kp ϕ
√̅̅̅̅̅̅̅̅̅̅̅̅
ρp ϕcp kB T 3.3. Other model
2kf 3πrc η
14 keff kc、n(R) =
kf
= ⎧ ⎡ ( ) ⎤2 ⎫ 3.3.1. The composite model
⎪ ln R/R ⎪
1 The above models only consider one of the influencing factors, so the
∫∞ [ ] ⎨ ⎬
(1 − ϕ) + 3ϕ 0 kc n(R)/ kc + 2kf dR √̅̅̅̅̅̅ exp − ⎣ √̅̅̅̅̅̅ ⎦
∫∞ [ ] r 2πlnσ ⎪ ⎩ 2π lnσ ⎪
⎭ application of the models is very small. Therefore, researchers began to
(1 − ϕ) + 3ϕ 0 kf n(R)/ kc + 2kf dR
15 keff
− 1 = c∣V∣
ϕrf ∣V∣、c、ϕ study the thermal conductivity prediction model that considers both
kf kf (1 − ϕ)rp factors.
16 keff = ϕkp + (1 − ϕ)kf kf、ϕ
17
(
kp + 2kf + 2 kp − kf ϕ
)
T、ρ、β Murshed et al. [122] took into account the influence of the nano­
keff = ) kf + 5 ×
meter layer and Brownian motion on the thermal conductivity and
(
kp + 2kf − kp − kf ϕ
proposed a prediction model that conforms to the thermal conductivity
√̅̅̅̅̅̅̅̅
kB T
104 βϕρcp f(T, ϕ)
ρD ( ) (Eq. (23)). This model is simplified to Maxwell model when nanometer
18 kp + 2kf + 2 kp − kf ϕ τ、∣V∣、Cf layer and Brownian motion are not considered. It can be seen from
keff = ( ) kf +
kp + 2kf − kp − kf ϕ
( ⃒ ⃒ ) Fig. 18 that compared with the Maxwell model and the model proposed
⃒ ⃒2
157.5ϕCf ⃒V⃒ τ by Prasher, the accuracy of this model is greatly improved.
19 β = 0.92、ϕ Prasher et al. [123] considered the Brownian motion and particle
keff =
( )
aggregation on the influence of the coefficient of thermal conductivity
(β + 1)kf + βkp + 2βϕ kp − kf
( ) kf (Eq. (24)) model. This model introduced a general relationship between
(β + 1)kf + βkp − βϕ kp − kf heat transfer coefficients and allowed to show visually the relationship

12
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

Fig. 3. The comparisons between the predicted data and the experimental data considering the clusters are showed in [15,93,94].

13
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

Table 4 predicting oxide nanofluids is poor.

The prediction models of thermal conductivity and their important parameters Komeilibirjandi et al. [129] established a neural network prediction
considering cluster effect. model for predicting thermal conductivity of CuO nanofluids by using
Number Prediction model of thermal conductivity The key parameters group method of data handling (GMDH) algorithm. GMDH algorithm is
20
( ) applied to deal with complex mathematical models, and the accuracy is
kpm kp + 2kf − 2ϕ kf − kp ϕ = ϕp − ϕc
kf
= ( )
kp + 2kf + ϕ k( f − kp )
improved by this algorithm. The model takes into account temperature,
21 kc + 2kf − 2ϕc kc − kf kc、ϕc = ϕ(1 + ξ)3 volume fraction and nanoparticle size. Fig. 23 is a comparison diagram
keff = ( ) kf
kc + 2kf + ϕc kc − kf ( ) of the predicted thermal conductivity data and the experimental data of
22 kp + (N − 1)kf + (N − 1) kp − kf ϕe r31
keff = ( ) kf ϕe = ϕ CuO nanofluids.
kp + (N − 1)kf − ϕe kp − kf r32
Jiang et al. [130] combined BP algorithm with A Genetic Algorithm
(GA) and A Mindevolutionary Algorithm (MEA), and the calculation
between thermal conductivity and temperature and particle size. But accuracy of a single BP algorithm could be improved through this
more information (e.g.: fractal size, polymerization and time constants) combination. The thermal conductivity of Cu/Al2O3-Ethylene glycol/
is needed to determine the corresponding constants. Fig. 19 shows that H2O mixed nanofluids was predicted by the neural network prediction
this model is more accurate than predicted by Maxwell’s model. models established by the two algorithms. The predicted data showed
Feng et al. [124] proposed a hybrid model (Eq. (25)) of an aggregate that the thermal conductivity of the nanofluid increases with the in­
of nanospheres and particles, in which they proposed that heat is crease of temperature, and the thermal conductivity and mixing ratio of
transferred between a combination of base liquid, solid particles and the base liquid increase non-linearly. In addition, we also found that
nanospheres. It can be seen from Fig. 20 that this model differs little compared with BP, the accuracy of GA-BPNN and MEA-BPNN improved
from the experimental data of Lee et al. [125] (Table 5). by 0.11% and 0.13%, respectively. It can be seen from Fig. 24 that MEA-
BPNN algorithm is more accurate among the three algorithms.
3.3.2. Neural network prediction Elsheikh et al. [131] used self-organizing maps (SOM) algorithm to
In recent years, neural network prediction technology has been design a neural network prediction model for thermal conductivity of
applied to various fields, applied to solve complex problems. Neural nanofluids. SOM algorithm can deal with complex input-output corre­
network is an alternative method based on numerical and analytical spondence and has high generalization ability. This neural network
methods without the need to deduce complex mathematical model prediction model takes the size, temperature and volume fraction of
formulas. Because the mathematical model will be very complex if the nanoparticles as inputs. Fig. 25a and b are graphs comparing the pre­
accurate prediction model of thermal conductivity of nanofluids is dicted and experimental data of thermal conductivity of Al2O3 and CuO
inferred, which needs to consider many factors, the process of using nanofluids. It can be seen from the figure that this model can better
neural network to replace the inference greatly shortens the research predict the thermal conductivity of oxide nanofluids.
time and can obtain more accurate prediction results [126]. Neural Sadeghzadeh et al. [132] also adopted the Self-Organizing Map
networks are made up of interconnected artificial neurons. Each of these (SOM) algorithm and Back Propagation-Levenberg-Marquardt (BP-LM)
neurons can receive one or more inputs, process them through different algorithm to establish a neural network prediction models for thermal
algorithms, and then output. Artificial neural network is a technology conductivity of nanofluids. These two models respectively predict the
used to establish the relationship between input and output, and an thermal conductivity of TiO2-Al2O3/H2O nanofluids. By comparing with
important part of this technology is different application algorithms the experimental data, the correlation coefficients are 0.938 and 0.98
[127]. At present, a large number of researchers have also begun to use respectively, which is a very impressive result. Fig. 26 also shows that
neural networks to predict the thermal conductivity of nanofluids. the correlation is very high.
Longo et al. [127] established a 3-input and 4-input neural network Ahmadi et al. [35] adopted three different neural network prediction
prediction model. The influence of particle size, volume temperature, models to predict the thermal conductivity of Al2O3/H2O nanofluids.
volume fraction and average particle size on thermal conductivity is Three different algorithms are adopted: Genetic Algorithm-Least Square
fully considered in this model. The elastic back propagation algorithm Support Vector Machine (GA-LSSVM), SOM and BP-LM. The correlation
(RProp) is used as the optimization algorithm, which is used to optimize coefficients of SOM and BP-LM algorithms are 0.888125 and 0.87575,
the input parameters. Only the sign of partial derivative is considered. respectively. It is also found that GA-LSSVM prediction model is more
These two models take into account the experimental variables such as suitable for predicting the thermal conductivity of Al2O3/H2O nano­
temperature, volume fraction of nanoparticles and thermal conductivity fluids. Fig. 27 shows the prediction of thermal conductivity of nanofluids
of nanoparticles, but the four-input model also takes into account the by GA-LSSVM algorithm at different temperatures.
average size of nanoparticle clusters. The neural network prediction The examples of using neural networks to establish models to predict
models highlight the influence of Brownian motion and cluster effect on the thermal conductivity of nanofluids in recent years are listed above.
thermal conductivity, and has a more accurate prediction for oxide By comparing the experimental data with the predicted data, it is found
nanofluids. Fig. 21a and b show the comparison of the predicted data that the data predicted by the neural network are closer to the experi­
and experimental data of the 4-input models and the 3-input models mental data, and the correlation coefficient is generally above 0.95. In
respectively. addition, with the development of neural network technology, the al­
Wang et al. [128] Backpropagation algorithm (BP) was used to gorithm is also in continuous progress. GA-BPNN, MEA-BPNN and BP-
establish a neural network prediction model to predict the thermal LM algorithms are derived from BP algorithm. These composite algo­
conductivity of ethylene glycol nanofluids. BP algorithm is one of the rithms greatly shorten the calculation time and have higher calculation
most basic optimization algorithms, which has high universality and accuracy. SOM and Levenberg algorithms are also commonly used.
adaptability. This model takes into account the thermal conductivity of Another advantage of the neural network prediction models is that,
the base liquid and the nanoparticles, the size of the nanoparticles, the compared with other prediction methods, this method has a high ac­
thermal conductivity of the nanofluid, the volume concentration and the curacy in predicting the thermal conductivity of oxide nanofluids.
temperature. As can be seen from Fig. 22, the comparison between the Therefore, in recent years, neural network has been widely used to
predicted results and the experimental data shows that the relative de­ predict the thermal conductivity of oxide nanofluids [133–135]
viation between the similarity of the data and the experimental data of (Table 6).
more than 99% is within 5%. However, the accuracy of this model for

14
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

Fig. 4. The comparisons diagram between the predicted data of the composite model and the experimental data are from [98–100].

15
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

Table 5
Thermal conductivity prediction models and their important parameters.
Number Prediction model of thermal conductivity The key
parameters
{ ( )( ) ( ) [ ]}
23 ϕω kp − ωkf 2ε31 − ε3 + 1 + kp + 2εkf ε31 ϕε3 (ε − 1) + 1
[ (
3Λ2 9Λ3 kcp + 2kf
)]
Gt、Cp-cp、∣V∣
keff = ( ) ( ) ( 3 ) kb + ϕ2 ε6 kf 3Λ2 + + +… +
ε1 kp + 2ωkf − kp + 2εkf ϕ ε1 − ε + 1
3 3 4 16 2kcp + 3kf
{ [√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
( ) ]}
1 3kf T 1 − 1.5ε3 ϕ Gt
ρ Cp− cp Ls + ⃒⃒ ⃒⃒
2 cp 2πρcp ε3 α3p 6π⃒V⃒εαp Ls
[ ( )]
24 ( ) kp + 2kf + 2ϕ kp − kf Re、Pr
keff = 1 +ARem Pr0.333 ϕ ( ) kb
k p + 2k − ϕ k − k
[ f
( p f ]
) {( [ ]}
25 kpe + 2kf + 2ϕ(1 − A)3 kpe − kf 3
)
1 rp + h A、ϕ、ϕ1
keff = (1 − ϕe ) 3( ) + ϕe 1 − ϕ1 +A ln( ) − 1
kpe + 2kf − ϕ(1 − A) kpe − kf 2 A rp + h (1 − A)

4. Nanofluid applications fluid as coolant in micro-channel radiators. Wu et al. [138] studied

the heat storage capacity of Al2O3-H2O nanofluid in the cooling system,
As a new type of material, nanofluids have a wider application space and infrared imaging showed that nanofluid could improve the freezing
due to their unique characteristics. Compared with traditional fluids, process of the fluid and increase the cooling rate by 20.5%. The appli­
nanofluids have the following advantages: cation of nanofluid in refrigeration industry can improve the perfor­
mance of refrigeration system and save the running time of refrigeration
(1) It has a large specific surface area, which makes the thermal system. At the same time, it can also achieve the purpose of energy
contact area between the nanoparticles and the base liquid larger, saving.
and has a higher thermal conductivity. The engine cooling system has a direct impact on the power of the
(2) With stronger controllability, the control of the thermal con­ automobile engine. Failure to cooling in time will reduce the power of
ductivity of magnetic nanofluid is achieved by controlling factors the engine. In order to make the cooling system meet the function of
such as particle concentration and particle size and by controlling rapid cooling, the cooling system using nanofluid as coolant has been
the direction and strength of the magnetic field. widely studied.
(3) Small particle size, in the flow process will not produce excessive Ali et al. [139] used Toyota Yaris 2007 cooling system to explore the
pressure drop. influence of different concentrations of Al2O3-H2O nanofluid on the
automotive cooling system, and collected the data of the engine cooling
Due to these unique advantages, nanofluids are widely used in a system under three working conditions. Compared with pure water as
variety of fields, such as automotive engine cooling system, solar energy coolant, nano-fluid coolant improved the working efficiency of the
system, heat pipe, refrigeration, etc. The broad application of nanofluids radiator. And the cooling system works best at a concentration of 1%.
will be summarized below. Salamon et al. [140] found that 0.3 vol% of TiO2 nanofluid could in­
crease the efficiency of the cooling system by approximately 8.5%
4.1. Application of cooling and heat dissipation compared to the base liquid as a coolant.
Siraj et al. [141] measured the operation of TiO2-H2O nanofluid as
Cooling/radiator has always been the most widely used system in coolant on Fiat Doblo 1.3 MJTD ENG automotive cooling system. The
mechanical and electronic industry. With the progress of industrial experimental results showed that the cooling efficiency of the automo­
science and technology, the heat dissipation of large computers and tive cooling system increased by 47% with the concentration of 0.2%
large engineering machinery has become a problem. In order to make TiO2-H2O nanofluid compared with the pure water and the concentra­
the cooling efficiency of the cooling system meet the normal operating tion of 0.1% and 0.3% nanofluid Mohammadi et al. [142] investigated
temperature of the machinery, a new type of cooling system designed by the addition of ɤ-Al2O3 and CuO nanoparticles to engine oil. They found
using nanofluid as heat transfer working medium is proposed. that ɤ-Al2O3 and CuO with a volume fraction of 2% increased their
Kumar et al. [136] analyzed the performance of automotive radiators thermal conductivity by 5% and 8%, respectively. Vasheghani [143]
using metal oxide nanofluids as coolant. Three - dimensional thermal dispersed alumina nanoparticles with a volume fraction of 3% in engine
fluid analysis of the radiator is also carried out. Due to the increase of oil, and the thermal conductivity was increased by 75.23% after mea­
nano-thermal conductivity and thermal dispersion effect, the perfor­ surement. In the latest study, Davis et al. [144] explored the different
mance of the micro-channel radiator using nano-fluid as coolant is shapes and the concentration of nanofluid influence on the automobile
higher than that using water as coolant. Due to the characteristics of cooling system. The results showed that the spherical (oxide), cylinder
small particle size and low volume fraction of nanoparticles, the pres­ (carbon nanotubes), flake (graphene) ternary hybrid nanofluid in the
sure drop is not too large while improving the thermal conductivity, volume fraction of 1% when the cooling efficiency of cooling system
which ensures the performance of the radiator. Due to the increase of increased by about 20%, However, it is necessary to overcome the need
nano-thermal conductivity and thermal dispersion effect, the perfor­ for higher pumping power to address pressure drop, stability and other
mance of micro-channel radiator using nano-fluid as coolant is higher issues. In addition, in order to reduce the overall consumption of the
than that using water as coolant. And because of the characteristics of vehicle, the automotive thermal management system based on nanofluid
small particle size and low volume fraction of nanoparticles, the thermal has also been studied. With the development of the electric vehicle in­
conductivity is enhanced at the same time, and the pressure drop is not dustry, nanofluids are also applied in the battery thermal management
too large to ensure the performance of the radiator. Jang and Choi [137] system of electric vehicles [145].
analyzed a nanofluid microchannel heat sink using a computational In addition to the automotive industry applications in the micro­
model of the thermal conductivity of nanofluids that takes Brownian electronics industry nanoflow body fluids have a corresponding appli­
motion into account. The results showed that the cooling performance of cation space. As the heat transfer medium of heat pipe, it is applied in
the water-based nanofluid was 10% higher than that of pure water. The the cooling system of some electronic equipment and the effect is better
above studies highlight the potential of ultra-high utilization of nano- than the general one [146].

16
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

Fig. 5. Comparison diagram of neural network prediction models and experimental data [35,127–132].

17
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

Fig. 5. (continued).

18
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

Table 6 heat of solar energy can be directly converted into electric energy, and
Neural network prediction methods. solar thermoelectric devices are also widely used [155]. Soltani et al.
Reference Prediction algorithm The key parameters [156] developed a new type of solar thermoelectric decoration using a
different cooling technology. Compared to other cooling technologies,
Longo et al. [127] elastic back propagation Temperature, volume
algorithm (Rprop) fraction, thermal nanofluid cooling not only improves the thermal performance but also
conductivity of particles, improves the total power generation.
average size of nanoparticle
clusters
4.3. Other applications
Wang et al. [128] Backpropagation algorithm The thermal conductivity of
(BP) the base solution and
nanoparticles, the size, In addition to the above applications, nanofluids are also widely used
volume concentration and in medicine. In the treatment of cancer, nanofluids can be used as a
temperature of the means of transportation to deliver drugs to the site of the tumor for
nanoparticles
treatment, without damaging the healthy tissue [26]. Compared with
Komeilibirjandi roup method of data handling Base liquid thermal
et al. [129] algorithm (GMDH) conductivity, volume ordinary particles, nanoparticles have higher adsorption and are more
fraction, temperature, conducive to the repair of injured tissue under the action of alternating
particle size magnetic field [25]. Nanofluid has recently been used in cryosurgery
Jiang et al. [130] GA-BPNN and MEA-BPNN Base liquid temperature,
[157]. Nanofluids are also widely used in the petroleum industry.
algorithm base liquid mixing ratio
Elsheikh et al. self-organizing maps Nanoparticle size,
Nanofluids can reduce the filtration volume of the mud and maintain the
[131] algorithm (SOM) temperature, volume compressive strength of the core. Nanofluid improves the stability of the
fraction emulsion and ensures the stability of the wellbore. The use of nanofluids
Sadeghzadeh et al. self-organizing maps Concentration, temperature, improves the efficiency of oil recovery. [158].
[132] algorithm (SOM) and Back thermal conductivity and
Propagation-Levenberq- viscosity
Marquardt algorithm (BP-LM) 5. Conclusion
Ahmad et al. [35] GA-LSSVM, SOM and BP-LM Volume concentration,
temperature, nanoparticle In this paper, the related research on nanofluids in recent years is
size
reviewed, and the factors affecting the thermal conductivity of nano­
fluids, the prediction models of thermal conductivity of nanofluids and
4.2. Application of solar energy the application fields of nanofluids in recent years are summarized and
evaluated.
Among clean energy sources, solar energy is as one of the most The thermal conductivity of the nanofluid depends on the size of the
important ones. How to use solar energy effectively is an important area nanoparticles, the volume fraction of the particles, the shape of the
of research. Although applications such as solar water heaters and col­ particles and their aggregation. It also depends on external factors such
lectors have been widely used, the actual utilization rate is very low. For as temperature [159], pH value [160], magnetic field [161,162] and
the full application of solar energy, much depends on the efficiency of dispersant [163]. However, it can be seen from the summary that the
the collectors and the heat transfer of the heat transfer medium [147]. influence of thermal conductivity of nanofluids is the stability of the
Therefore, researchers began to study the heat transfer working medium fluid and the shape of the nanoparticle aggregation state.
and found that nanofluids may be the most suitable heat transfer The heat transfer mechanism of nanofluids is very complicated. From
working medium [148]. the classical two-phase thermal conductivity prediction models, re­
Fikri et al. [149] applied the nanofluid to the solar heat collector and searchers began to consider the Brownian motion of particles, interface
measured its temperature. The results showed that when the nanofluid effect, particle cluster influence and heat transfer influence respectively
with a volume concentration of 0.3% was irradiated by solar radiation of to derive the corresponding single factor prediction models. In recent
300 W⋅m− 2, it could reach a maximum temperature of 53.0 ◦ C, and years, there are also multi-factor composite prediction models. To sum
when it was irradiated by solar radiation of 700 W⋅m− 2, it could reach a up these models, most of them are modified on the basis of classical
maximum temperature of 66.9 ◦ C. Are higher than the heat transfer models (e.g.: Maxwell model and H–C model). They consider more
medium of pure water. Zheng et al. [150] used CuO-H2O nanofluid as A factors, generalized linear solutions based on particle size distribution,
heat transfer working medium and found that the efficiency of the solar temperature distribution and other theories, or put forward different
collector was improved by about 24%, and the improved efficiency composite models to further improve the prediction accuracy of the
varied with the size of the nanoparticles. Lin et al. [151] designed a new prediction models. From the prediction results and experimental data,
solar collector that consists of a vacuum tube, a composite parabolic the error is basically controlled within 5%. In recent years, due to the
concentrator and an open thermosiphon, and uses CuO-H2O nanofluid as rapid development of neural network, some researchers also use neural
a heat transfer working medium. Experimental data showed that the network modeling to predict thermal conductivity. The advantage of
efficiency of the collector is improved by 6.6% compared to pure water. these methods is that the influence of many factors can be considered at
In addition to the application in the solar collector, nanofluids are the same time, and with the optimization of the algorithm, the compu­
also used in the photoelectric system. In order to better convert solar tational speed and accuracy of the neural network have reached a very
energy into electric energy, the photoelectric system must keep the high level. Especially, it has a high accuracy in predicting the thermal
temperature of the battery at a certain temperature, because the con­ conductivity of metal oxide nanofluids. Because of the neural network
version efficiency will be too low if the temperature of the battery is too does not need to derive complex mathematical formulas, but can well
high [152]. Therefore, people began to add nanofluids to the traditional relate the input and output. I believe these will be the main method to
photoelectric system to form a hybrid system of photoelectric heat, predict the future thermal conductivity. However, there are still some
which absorbs waste heat to achieve the purpose of full utilization of deviations between the results predicted by the neural network and the
energy [153]. Al-Waeli et al. [154] used Si–C nanofluid as coolant to experimental data. This may be caused by the composite effect of the
absorb the heat of the battery, and the experimental results showed that influencing factors or the important heat transfer mechanism is not clear
the overall efficiency of the photoelectric thermal system was improved yet, or the algorithm is not up to a high level, so further research is
by 85%. needed in this area.
In recent years, according to the Pelther effect and Seebeck effect, the As we learn more about the thermal conductivity of nanofluids, they
will make significant contributions to areas such as cooling/cooling

19
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

systems, solar energy collection, medicine, and more. However, the c Clusters of body
application of nanofluids is not only that, because of its unique thermal
conductivity, it is suggested that in the future in the field of new energy Declaration of Competing interest
utilization and energy recovery, a large number of nanofluids should be
used as the carrier of thermal energy recovery [164,165]. Nanofluids Here within enclosed is our paper for consideration to be published
have also be suggested for aerospace applications, where they could play on “International Communications in Heat and Mass Transfer”. In this
a huge role in converting solar energy into other forms of energy paper, the thermal conductivity of nanofluids is reviewed and evaluated
[166–169]. I believe that the unique thermal properties of nanofluids according to the recent literatures on the thermal conductivity of
will play an important role in future research in the field of heat nanofluids. The results show that the research on this aspect has been
exchange. ongoing, and there are many related articles, and the research results
Finally, although the research on thermal conductivity of nanofluids have made great progress in recent years with the support of multidis­
has been steadily advanced in recent years, new theories and models ciplinary. But there is no coherent review of the thermal conductivity of
have been proposed. And apply the methods of other fields to the study nanofluids from all aspects. In order to solve this problem, the research
of nanofluids. But there are still some unresolved issues. Therefore, on nanofluids in recent years is reviewed from three aspects: influencing
large-scale theoretical research and experiments should be carried out in factors, prediction models and applications, and the prediction models
the future. It is hoped that theoretical and predictive models applicable and evaluation of thermal conductivity of nanofluids are emphatically
to any nanofluid can be obtained. introduced. The purpose of this paper is to form a complete and detailed
description of the thermal conductivity of nanofluids, which can be used
Nomenclature as a reference guide for researchers to understand and study the thermal
conductivity of nanofluids, and to update the latest research results in a
k Thermal conductivity [W⋅m− 2⋅K] timely manner. It is hoped to point out the direction for the future
k1 Thermal conductivity of the nano-layer [W⋅m− 2⋅K] research of nanofluids.
keff Thermal conductivity of nanofluids [W⋅m− 2⋅K] All authors have seen the manuscript and approved to submit to your
r The radius [mm] journal. Neither the entire paper nor any part of its content has been
r1，2 Radius of thermal resistance [mm] published or has been accepted elsewhere. It is not being submitted to
T Temperature [K] any other journal.
h Nano layer thickness [nm]
C The heat capacity [J⋅kg− 1⋅K]
Acknowledgement
|V| Averaged velocity [m⋅s− 1]

Financial support is provided by the National Natural Science

Dimensionless
Foundation of China (51806114, 51874187) and China Postdoctoral
Science Foundation (2020M682206).
kb Boltzmann constant
Pr Prandtl number
References
Re Reynolds number
N Empirical shape factor (H–C model) [1] S. Choi, J.A. Eastman, Enhancing thermal conductivity of fluids with
Gt The motion due to the DLVO potential nanoparticles, ASME Fed. 231 (1995) 99–105.
A Ratio(1-kf/kpe) [2] Y. Xuan, Q. Li, Investigation on convective heat transfer and flow features of
nanofluids, J. Heat Transf. 125 (1) (2003) 151–155.
c Modified model constants [3] S. Murshed, K.C. Leong, C. Yang, Enhanced thermal conductivity of TiO2—water
n particle number per volume based nanofluids, Int. J. Therm. Sci. 44 (4) (2005) 367–373.
n(R) The radius distribution function [4] H.A. Moghaddam, A. Ghafouri, R.F. Khouzestani, Viscosity and thermal
conductivity correlations for various nanofluids based on different temperature
and nanoparticle diameter, J. Braz. Soc. Mech. Sci. Eng. 43 (2021) 6.
Greek letters [5] A. Abdullah, I. Syakir, A. Yusairi, S.H. Sheikh, S.Z. Abidin, Study of thermal
conductivity and heat teansfer coefficient of NWCNT-OH/deionised water-
ethylene glycol based nanofluids, Def. S and T Technical. Bull. 12 (2) (2020)
η Liquid viscosity [g⋅cm− 1⋅s− 1]
226–237.
α Thermal diffusion rate [m2⋅s− 1] [6] J. Tielke, M. Maas, M. Castillo, K. Rezwan, M. Avila, Statistical analysis of
φ Angle between the axial direction of the CNT and the isotherm thermal conductivity experimentally measured in water-based nanofluids, Proc.
[◦ ] R. Soc. A Math. Phys. Eng. Sci. 477 (2021) 2250.
[7] A. Abutaleb, M. Imran, Thermal conductivity enhancement for CuO nanoflakes in
ρ Density [kg⋅m− 3] oil-based and oil blend-based nanofluids, J. Chin. Chem. Soc. Taipei (2021),
μeff Viscosity of slurry [g⋅cm− 1⋅s− 1] https://doi.org/10.1002/jccs.202100005.
τ The relaxation time [8] B.M. Sikindar, A.V. Sita Rama, M. Bhagvanth, Heat transfer enhancement and
pressure drop of Fe3O4 -water nanofluid in a double tube heat exchanger with
θ Ratio of nanosphere thickness to original particle radius internal longitudinal fins, Case Stud. Therm. Eng. 12 (2018) 600–607.
ψ Sphericity [9] P.B. Maheshwary, C.C. Handa, K.R. Nemade, S.R. Chaudhary, Role of
ζ The thickness ratio of interfacial layer to cluster radius nanoparticle shape in enhancing the thermal conductivity of nanofluids, Mater.
Today Proc. 28 (Pt 2) (2020) 873–878.
β Modified model constants [10] X. Deng, Z. Wu, G. Wang, Thermal conductivity and heat capacity of water/Iβ
ϕ Nanoparticle volume fraction cellulose nanofluids: a molecular dynamics study, Int. J. Mod. Phys. B 34 (2020)
ϕ1 total volume fraction of a quarter of the column in the 27.
[11] A. Dadwal, P.A. Joy, Particle size effect in different base fluids on the thermal
aggregation model
conductivity of fatty acid coated magnetite nanofluids, J. Mol. Liq. 303 (2020)
112650.
Subscripts [12] J. Wang, G.L. Li, T. Li, M. Zeng, B. Sunden, Effect of various surfactants on
stability and thermophysical properties of nanofluids, J. Therm. Anal. Calorim.
143 (2020) 4057–4070.
pe The equivalent grain [13] G. Huminic, A. Huminic, F. Dumitrache, C. Fleac, I. Morjan, Study of the thermal
f Based on the liquid conductivity of hybrid nanofluids: recent research and experimental study,
pex The X-axis direction of the nanoparticle Powder Technol. 367 (2020) 347–357.
[14] K. Hosseinzadeh, A. Asadi, A.R. Mogharrebi, M.E. Azari, D.D. Ganji, Investigation
pez The Z axis of the nanoparticles of mixture fluid suspended by hybrid nanoparticles over vertical cylinder by
p Hybrid particles considering shape factor effect, J. Therm. Anal. Calorim. 143 (2020) 19–20.

20
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

[15] S. Salehi, A. Nori, K. Hosseinzadeh, D.D. Ganji, Hydrothermal analysis of MHD [41] M.H. Esfe, S. Saedodin, S. Wongwises, D. Toghraie, An experimental study on the
squeezing mixture fluid suspended by hybrid nanoparticles between two parallel effect of diameter on thermal conductivity and dynamic viscosity of Fe/water
plates, Case Stud. Therm. Eng. 21 (2020) 100650. nanofluids, J. Therm. Anal. Calorim. 119 (3) (2015) 1817–1824.
[16] M. Gholinia, K. Hosseinzadeh, D.D. Ganji, Investigation of different base fluids [42] A. Arani, J. Amani, Experimental investigation of diameter effect on heat transfer
suspend by CNTs hybrid nanoparticle over a vertical circular cylinder with performance and pressure drop of TiO2–water nanofluid, Exp. Therm. Fluid Sci.
sinusoidal radius, Case Stud. Therm. Eng. 21 (2020) 100666. 44 (2013) 520–533.
[17] K. Hosseinzadeh, E. Montazer, M.B. Shafii, D.D. Ganji, Heat transfer hybrid [43] M.R. Esfahani, E.M. Languri, M.R. Nunna, Effect of particle size and viscosity on
nanofluid (1-Butanol/MoS2 –Fe3O4) through a wavy porous cavity and its thermal conductivity enhancement of graphene oxide nanofluid, Int. Commun.
optimization, Int. J. Num. Methods Heat Fluid Flow 31 (2021) 1547–1567. Heat Mass Transf. 76 (2016) 308–315.
[18] S. Murshed, G. Solomon, M. Sharifpur, M.H. Ahmadi, J.P. Meyer, Experimental [44] N.T. Khoa, S.W. Kim, D.W. Yoo, E.J. Kim, S.H. Hahn, Fabrication of Au/
investigation on stability, viscosity, and electrical conductivity of water-based Graphene-Wrapped ZnO-Nanoparticle-Assembled Hollow Spheres with Effective
hybrid nanofluid of MWCNT-Fe2O3, Nanomaterials 11 (2021) 136. Photoinduced Charge Transfer for Photocatalysis, Acs Applied Materials &
[19] S. Hamze, N. Berrada, D. Cabaleiro, A. Desforges, P. Estelle, Few-layer graphene- Interfaces 7 (6) (2015) 3524–3531.
based nanofluids with enhanced thermal conductivity, Nanomaterials 10 (2020) [45] D. Wen, Y. Ding, Formulation of nanofluids for natural convective heat transfer
1258. applications, Int. J. Heat Fluid Flow 26 (6) (2005) 855–864.
[20] K. Hosseinzadeh, S. Roghani, A.R. Mogharrebi, A. Asadi, M. Waqas, D.D. Ganji, [46] S.Z. Heris, M.N. Esfahany, S.G. Etemad, Experimental investigation of convective
Investigation of cross-fluid flow containing motile gyrotactic microorganisms and heat transfer of Al2O3/water nanofluid in circular tube – ScienceDirect, Int. J.
nanoparticles over a three-dimensional cylinder, Alex. Eng. J. 59 (2020) Heat Fluid Flow 28 (2) (2007) 203–210.
3297–3307. [47] E.V. Timofeeva, J.L. Routbort, D. Singh, Particle shape effect on thermophysical
[21] S. Salehi, M.R. Mardani, D.D. Ganji, K. Hosseinzadeh, Investigation of nano- properties of alumina nanofluids, J. Appl. Phys. 106 (1) (2009)
Bioconvective fluid motile microorganism and nanoparticle flow by considering 014304.01–014304.10.
MHD and thermal radiation, Inform. Med. Unlocked 21 (2020) 100462. [48] H.J. Kim, S.H. Lee, J.H. Lee, S.P. Jang, Effect of particle shape on suspension
[22] A.R. Mogharrebi, A.R.D. Ganji, K. Hosseinzadeh, S. Roghani, A. Asadi, stability and thermal conductivities of water-based bohemite alumina nanofluids,
A. Fazlollahtabar, Investigation of magnetohydrodynamic nanofluid flow contain Energy 90 (2) (2015) 1290–1297. OCT. PT.
motile oxytactic microorganisms over rotating cone, Int. J. Num. Methods Heat [49] S.A. Mirmohammadi, M. Behi, Y.X. Gan, L.M. Shen, Particle-shape-, temperature-,
Fluid Flow (2021), https://doi.org/10.1108/HFF-08-2020-0493 ahead-of-print and concentration-dependent thermal conductivity and viscosity of nanofluids,
(ahead-of-print). Phys. Rev. E 99 (4–1) (2019), 043109.
[23] M. Izadi, M. Assad, Use of nanofluids in solar energy systems, design and [50] P.B. Maheshwary, C.C. Handa, K.R. Nemade, S.R. Chaudhary, Role of
performance optimization of renewable, Energy Syst. (2021) 221–250. nanoparticle shape in enhancing the thermal conductivity of nanofluids –
[24] X. Li, H. Wang, B. Luo, The thermophysical properties and enhanced heat transfer ScienceDirect, Mater. Today Proc. 28 (2020) 873–878.
performance of SiC-MWCNTs hybrid nanofluids for car radiator system, Colloids [51] V. Bianco, O. Manca, S. Nardini, Numerical investigation on nanofluids turbulent
Surf. A Physicochem. Eng. Asp. 612 (2020) 125968. convection heat transfer inside a circular tube, Int. J. Therm. Sci. 50 (3) (2011)
[25] A.M. Grumezescu, E. Andronescu, A. Ficai, C.H. Yang, C. Bleotu, Magnetic 341–349.
nanofluid with antitumoral properties, Lett. Appl. NanoBioSci. 1 (3) (2012) [52] A. Sebm, A. Sjp, T. Cong, R. Gilles, G. Nicolas, Heat transfer enhancement by
56–60. using nanofluids in forced convection flows, Int. J. Heat Fluid Flow 26 (4) (2005)
[26] S. Masood, M. Farooq, S. Ahmad, Description of viscous dissipation in 530–546.
magnetohydrodynamic flow of nanofluid: applications of biomedical treatment, [53] S.Z. Heris, S.G. Etemad, M.N. Esfahany, Experimental investigation of oxide
Adv. Mech. Eng. 12 (6) (2020) 1–13. nanofluids laminar flow convective heat transfer, Int. Commun. Heat Mass Transf.
[27] M.V. Krishna, A.J. Chamkha, Hall and ion slip effects on unsteady MHD 33 (4) (2006) 529–535.
convective rotating flow of nanofluids—application in biomedical engineering, [54] P.K. Namburu, D.K. Das, K.A. Tanguturi, R.S. Vajjha, Numerical study of
J. Egypt. Math. Soc. 28 (2020) 1. turbulent flow and heat transfer characteristics of nanofluids considering variable
[28] M.S. Alnarabiji, M.M. Husein, Application of bare nanoparticle-based nanofluids properties, Int. J. Therm. Sci. 48 (2) (2009) 290–302.
in enhanced oil recovery, Fuel 267 (2020) 117262. [55] A. Ceylan, K. Jastrzembski, S.I. Shah, Enhanced solubility Ag-Cu nanoparticles
[29] I. Ngo, K. Sasaki, R. Nguele, Y. Sugai, Formation damage induced by water-based and their thermal transport properties, Metall. Mater. Trans. A 37 (2006)
alumina nanofluids during enhanced oil recovery: influence of Postflush salinity, 2033–2038.
ACS Omega 5 (2020) 27103–27112. [56] A. Kakavandi, M. Akbari, Experimental investigation of thermal conductivity of
[30] M. Ghazvini, H. Maddah, R. Peymanfar, M.H. Ahmadi, R. Kumar, Experimental nanofluids containing of hybrid nanoparticles suspended in binary base fluids and
evaluation and artificial neural network modeling of thermal conductivity of propose a new correlation, Int. J. Heat Mass Transf. 124 (2018) 742–751.
water based nanofluid containing magnetic copper nanoparticles, Phys. A Stat. [57] H.F. Jiang, Q.H. Xu, C. Huang, L. Shi, Effect of temperature on the effective
Mech. Appl. 551 (2020) 124127. thermal conductivity of n-tetradecane-based nanofluids containing copper
[31] M.H. Ahmadi, M. Ghazvini, H. Maddah, M. Kahani, S. Pourfarhang, nanoparticles, Particuology 22 (2015) 101–105.
A. Pourfarhang, S.Z. Heris, Prediction of the pressure drop for CuO/(ethylene [58] S. Mukherjee, P.C. Mishra, S.K.S. Parashar, P. Chaudhuri, Role of temperature on
glycol-water) nanofluid flows in the car radiator by means of Artificial Neural thermal conductivity of nanofluids: a brief literature review, Heat Mass Transf. 52
Networks analysis integrated with genetic algorithm, Phys. A: Stat. Mech. Appl. (11) (2016) 2575–2585.
546 (2020) 124008. [59] S.R. Krishnan, V. Namboothiri, EMD analysis on the impact of temperature,
[32] M.H. Ahmadi, B. Mohseni-Gharyehsafa, M. Ghazvini, M. Goodarzi, R.D. Jilte, volume fraction and molecular weight on the thermal conductivity of water-based
R. Kumar, Comparing various machine learning approaches in modeling the nanofluids, J. Therm. Anal. Calorim. 7 (2020).
dynamic viscosity of CuO/water nanofluid, J. Therm. Anal. Calorim. 139 (2019) [60] M. Sedighi, A. Mohebbi, Investigation of nanoparticle aggregation effect on
4. thermal properties of nanofluid by a combined equilibrium and non-equilibrium
[33] M.H. Ahmadi, M. Ghazvini, M.H. Ahmadi, Predicting the efficiency of CuO/water molecular dynamics simulation, J. Mol. Liq. 197 (2014) 14–22.
nanofluid in heat pipe heat exchanger using neural network, Int. Commun. Heat [61] W. Wei, J.C. Cai, X.Y. Hu, Q. Han, S. Liu, Y.F. Zhou, Fractal analysis of the effect
Mass Transf. 104 (2019) 33–40. of particle aggregation distribution on thermal conductivity of nanofluids, Phys.
[34] M.H. Ahmadi, B. Mohseni-Gharyehsafa, M. Farzaneh-Gord, R.D. Jilte, R. Kumar, Lett. A 380 (2016) 2953–2956.
K.W. Chau, Applicability of connectionist methods to predict dynamic viscosity of [62] C.G. Subramaniam, On the effective thermal conductivity of nanofluids with
silver/water nanofluid by using ANN-MLP, MARS and MPR algorithms, Eng. fractal aggregation, J. Heat Transf. 141 (4) (2019) 044501.1–044501.4.
Appl. Comput. Fluid Mech. 13 (2019) 220–228. [63] B.A. Suleimanov, H.F. Abbasov, Effect of copper nanoparticle aggregation on the
[35] M.H. Ahmadi, M.A. Nazari, R. Ghasempour, H. Madah, M.B. Shafii, M.A. Ahmadi, thermal conductivity of nanofluids, Russ. J. Phys. Chem. 90 (2) (2016) 420–428.
Thermal conductivity ratio prediction of Al2O3/water nanofluid by applying [64] D. Song, D.W. Jing, W.G. Ma, X. Zhang, Effect of particle aggregation on thermal
connectionist methods, Colloids Surf. A Physicochem. Eng. Asp. 541 (2018) conductivity of nanofluids: enhancement of phonon MFP, J. Appl. Phys. 125 (1)
154–164. (2019), 015103.
[36] H. Maddah, R. Aghayari, M.H. Ahmadi, M. Rahimzadeh, N. Ghasemi, Prediction [65] Y. Ding, D. Wen, Particle migration in a flow of nanoparticle suspensions, Powder
and modeling of MWCNT/carbon (60/40)/SAE 10 W 40/SAE 85 W 90(50/50) Technol. 149 (2–3) (2005) 84–92.
nanofluid viscosity using artificial neural network (ANN) and self-organizing map [66] P.E. Gharagozloo, K.E. Goodson, Aggregate fractal dimensions and thermal
(SOM), J. Therm. Anal. Calorim. 134 (2018) 2275–2286. conduction in nanofluids, J. Appl. Phys. 108 (2010) 0743091–074309.7.
[37] M. Fatemeh, A. Sadeghzadeh, H. Khadang, M.H. Maddah, H. Ahmadi, L.G. [67] M. Motevasel, A. Nazar, M. Jamialahmadi, The effect of nanoparticles
Chen Sakhaeinia, Modeling of heat transfer performance of carbon nanotube aggregation on the thermal conductivity of nanofluids at very low concentrations:
nanofluid in a tube with fixed wall temperature by using ANN–GA, Eur. Phys. J. experimental and theoretical evaluations, Heat Mass Transf. 54 (1) (2018)
Plus 135 (2020) 1–20. 125–133.
[38] M.P. Beck, Y.H. Yuan, P. Warrier, A.S. Teja, The effect of particle size on the [68] X.F. Li, D.S. Zhu, X.J. Wang, N. Wang, J.W. Gao, H. Li, Thermal conductivity
thermal conductivity of alumina nanofluids, J. Nanopart. Res. 11 (5) (2008) enhancement dependent pH and chemical surfactant for Cu-H2O nanofluids,
1129–1136. Thermochim. Acta 469 (1–2) (2008) 98–103.
[39] M. Liu, C. Ding, J. Wang, Modeling of thermal conductivity of nanofluids [69] X.J. Wang, X.F. Li, S. Yang, Influence of pH and SDBS on the stability and thermal
considering aggregation and interfacial thermal resistance, RSC Adv. 6 (5) (2016) conductivity of nanofluids, Energy Fuel 23 (5) (2009) 270–278.
3571–3577. [70] C.T. Wamkam, M.K. Opoku, H.P. Hong, P. Smith, Effects of pH on heat transfer
[40] M. Chopkar, S. Sudarshan, P.K. Das, I. Manna, Effect of particle size on thermal nanofluids containing ZrO2 and TiO2 nanoparticles, J. Appl. Phys. 109 (2) (2011)
conductivity of Nanofluid, Metall. Mater. Trans. A 39 (7) (2008) 1535–1542. 2252.

21
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

[71] H. Younes, G. Christensen, X.N. Luan, H.P. Hong, P. Smith, Effects of alignment, [103] J. Koo, C. Kleinstreuer, Impact analysis of nanoparticle motion mechanisms on
pH, surfactant, and solvent on heat transfer nanofluids containing Fe2O3 and CuO the thermal conductivity of nanofluids, Int. Commun. Heat Mass Transf. 32 (9)
nanoparticles, J. Appl. Phys. 111 (6) (2012) 2252. (2005) 1111–1118.
[72] K. Cacua, F. Ordonez, C. Zapata, B. Herrera, E. Pabon, R. Buitrago-Sierra, [104] H.E. Patel, S.K. Das, T.A. Sundararajan, A.S. Nair, T. Pradeep, Thermal
Surfactant concentration and pH effects on the zeta potential values of alumina conductivities of naked and monolayer protected metal nanoparticle based
nanofluids to inspect stability, Colloids Surf. A Physicochem. Eng. Asp. 583 nanofluids: manifestation of anomalous enhancement and chemical effects, Appl.
(2019) 123960. Phys. Lett. 83 (14) (2003) 2931–2933.
[73] K.Y. Leong, N.M. Leong, M. Sohaimi, N.H. Amer, The effect of surfactant on [105] R. Azizian, E. Doroodchi, B. Moghtaderi, Effect of nanoconvection caused by
stability and thermal conductivity of carbon nanotube based nanofluids, Therm. Brownian Motion on the enhancement of thermal conductivity in nanofluids, Ind.
Sci. 20 (2016) 429–436. Eng. Chem. Res. 51 (4) (2012) 1782–1789.
[74] G. Xia, H. Jiang, R. Liu, Y. Zhai, Effects of surfactant on the stability and thermal [106] Y. Xuan, L. Qiang, W. Hu, Aggregation structure and thermal conductivity of
conductivity of Al2O3/de-ionized water nanofluids, Int. J. Therm. Sci. 84 (2014) nanofluids, AICHE J. 49 (4) (2003) 1038–1043.
118–124. [107] S.K. Das, N. Putra, P. Thiesen, W. Roetzel, Temperature dependence of thermal
[75] Keklikcioglu Nese, Cakmak, The impact of surfactants on the stability and thermal conductivity enhancement for nanofluids, J. Heat Transf. 125 (4) (2003) 567.
conductivity of graphene oxide de-ionized water nanofluids, J. Therm. Anal. [108] B.X. Wang, L.P. Zhou, X.F. Peng, A fractal model for predicting the effective
Calorim. 139 (3) (2020) 1895–1902. thermal conductivity of liquid with suspension of nanoparticles, Int. J. Heat Mass
[76] K.H. Almitani, N.H. Abu-Hamdeh, S. Etedali, A. Abdollahi, A.S. Goldanlou, Transf. 46 (14) (2003) 2665–2672.
A. Golmohammadzadeh, Effects of surfactant on thermal conductivity of aqueous [109] D. Hemanth Kumar, H.E. Patel, V.R. Rajeev Kumar, T. Sundararajan, T. Pradeep,
silica nanofluids, J. Mol. Liq. 327 (2020) 114883. K.D. Sarit, Model for heat conduction in nanofluids, Phys. Rev. Lett. 93 (14)
[77] E. Tamjid, B.H. Guenther, Rheology and colloidal structure of silver nanoparticles (2004) 144301.
dispersed in diethylene glycol, Powder Technol. 197 (1–2) (2010) 49–53. [110] P. Bhattacharya, S.K. Saha, A. Yadav, P.E. Phelan, Brownian dynamics simulation
[78] G.D. Xia, R. Liu, M. Du, Analysis of thermal conductivity of nanofiltration fluids, to determine the effective thermal conductivity of nanofluids, J. Appl. Phys. 95
J. Beijing Univ. Technol. 42 (2016) 1252–1258 (in Chinese). (11) (2004) 6492–6494.
[79] S. Chakraborty, P.K. Panigrahi, Stability of nanofluid: a review, Appl. Therm. Eng. [111] J. Koo, C. Kleinstreuer, A new thermal conductivity model for nanofluids,
174 (2020) 115259. J. Nanopart. Res. 6 (2004) 577–588.
[80] Y. Li, Z. Qin, Z. Zou, L. Hou, Study on influence of dispersants on stability, [112] J. Koo, J.C. Kleinstreuer, Laminar nanofluid flow in microheat-sinks, Int. J. Heat
thermal conductivity and viscosity of BN/EG nanofluids, IOP Conf. Ser. Mater. Mass Transf. 48 (13) (2005) 2652–2661.
Sci. Eng. 774 (1) (2020), 012118. [113] Y. Bao, Thermal conductivity equations based on Brownian motion in suspensions
[81] J. Patel, K. Parekh, R.V. Upadhyay, Maneuvering thermal conductivity of of nanoparticles (nanofluids), J. Heat Transf. 130 (4) (2008) 156–161.
magnetic nanofluids by tunable magnetic fields, J. Appl. Phys. 117 (24) (2015) [114] S. Dong, X. Chen, An improved model for thermal conductivity of nanofluids with
243906. effects of particle size and Brownian motion, J. Therm. Anal. Calorim. 129 (2)
[82] C.L. Altan, A. Elkatmis, M. Yuksel, N. Aslan, S. Bucak, Enhancement of thermal (2017) 1–9.
conductivity upon application of magnetic field to Fe3O4 nanofluids, J. Appl. [115] W. Evans, Role of Brownian motion hydrodynamics on nanofluid thermal
Phys. 110 (9) (2011) 093917.1–093917.7. conductivity, Appl. Phys. Lett. 88 (9) (2006) 219–482.
[83] K. Parekh, H.S. Lee, Experimental investigation of thermal conductivity of [116] J.W. Gao, R.T. Zheng, H. Ohtani, D.S. Zhu, G. Chen, Experimental investigation of
magnetic nanofluids, AIP Conf. Proc. 1447 (100) (2012) 121901–134101. heat conduction mechanisms in nanofluids. Clue on clustering, Nano Lett. 9 (12)
[84] A. Katiyar, P. Dhar, T. Nandi, S.K. Das, Enhanced heat conduction characteristics (2009) 4128–4132.
of Fe, Ni and Co nanofluids influenced by magnetic field, Exp. Thermal Fluid Sci. [117] R. Deepak Selvakumar, S. Dhinakaran, A multi-level homogenization model for
78 (2016) 345–353. thermal conductivity of nanofluids based on particle size distribution (PSD)
[85] Q. Li, Y.M. Xuan, A study on the mechanism of the enhancement of thermal analysis, Powder Technol. 301 (C) (2016) 310–317.
conductivity by nano-fluids, J. Thermal Eng. 17 (2002) 63–73 (in Chinese). [118] D. Zhou, H. Wu, A thermal conductivity model of nanofluids based on particle size
[86] A.T. Utomo, H. Poth, P.T. Robbins, A.W. Pacek, Experimental and theoretical distribution analysis, Appl. Phys. Lett. 105 (2014), 083117.1–083117.4.
studies of thermal conductivity, viscosity and heat transfer coefficient of titania [119] S.M.S. Murshed, K.C. Leong, C. Yang, Thermophysical and electrokinetic
and alumina nanofluids, Int. J. Heat Mass Transf. 55 (2012) 25–26. properties of nanofluids-A critical review, Appl. Therm. Eng. 28 (17–18) (2008)
[87] Q. Li, Study on Heat Transfer Enhancement Mechanism of Nanofluids, Nanjing 2109–2125.
University of Science and Technology, 2004 (in Chinese). [120] R. Pasrija, S. Srivastava, The Interfacial Layer and the Thermal Conductivity of
[88] H.T. Zhu, Preparation, Stability and Thermal Conductivity of Nanofluids, Nanofluid, Heat Transfer—Asian Research 43 (3) (2014) 288–296.
Shandong University, 2005 (in Chinese). [121] Z. Ahmed, A. Bhargav, Thermal interfacial resistance and nanolayer effect on the
[89] D.H. Yoo, K.S. Hong, H.S. Yang, Study of thermal conductivity of nanofluids for thermal conductivity of Al2O3-CO2 nanofluid: A Molecular Dynamics approach,
the application of heat transfer fluids, Thermochim. Acta 455 (1–2) (2007) Journal of Enhanced Heat Transfer 28 (2021) 41–56.
66–69. [122] S.M.S. Murshed, K.C. Leong, C. Yang, A combined model for the effective thermal
[90] P. Keblinski, S.R. Phillpot, S. Choi, J.A. Eastman, Mechanisms of heat flow in conductivity of nanofluids, Appl. Therm. Eng. 29 (11–12) (2009) 2477–2483.
suspensions of nano-sized particles (nanofluids), Int. J. Heat Mass Transf. 45 (4) [123] R. Prasher, P. Bhattacharya, P.E. Phelan, Thermal conductivity of nanoscale
(2002) 855–863. colloidal solutions (nanofluids), Phys. Rev. Lett. 94 (2) (2005), 025901.
[91] J.A. Eastman, S.R. Phillpot, S.U.S. Choi, P. Keblinski, Thermal transport in [124] Y.J. Feng, B. Yu, P. Xu, M.Q. Zou, The effective thermal conductivity of nanofluids
nanofluids, Annu. Rev. Mater. Res. 34 (1) (2004) 219–246. based on the nanolayer and the aggregation of nanoparticles, J. Phys. D. Appl.
[92] C.J. Yu, A.G. Richter, A. Datta, M.K. Durbin, P. Dutta, Molecular layering in a Phys. 40 (10) (2007) 3164.
liquid on a solid substrate: an X-ray reflectivity study, Phys. B Condens. Matter [125] J.A. Eastman, S.U.S. Choi, S. Li, L.J. Tompson, S. Lee, Enhanced Thermal
283 (1) (2000) 27–31. Conductivity through the Development of Nanofluids, MRS Online Proceedings
[93] Q. Xue, W.M. Xu, A model of thermal conductivity of nanofluids with interfacial Library 457 (1999) 3–11.
shells, Mater. Chem. Phys. 90 (2–3) (2005) 298–301. [126] M. Ramezanizadeh, A. Puia, M.H. Ahmadi, G. Lorenzini, A review on the
[94] S. Lee, S.U.S. Choi, Measuring thermal conductivity of fluids containing oxide applications of intelligence methods in predicting thermal conductivity of
nanoparticles, J. Heat Transf. 121 (2) (1999) 280. nanofluids, J. Therm. Anal. Calorim. 138 (2019) 827–843.
[95] W. Yu, S.U.S. Choi, The role of interfacial layers in the enhanced thermal [127] G.A. Longo, et al., Application of Artificial Neural Network (ANN) for the
conductivity of nanofluids: a renovated Maxwell model, J. Nanopart. Res. 6 (4) prediction of thermal conductivity of oxide–water nanofluids, Nano Energy 1 (2)
(2004) 355–361. (2012) 290–296.
[96] H. Xie, M. Fujii, Z. Xing, Effect of interfacial nanolayer on the effective thermal [128] X.H. Wang, X.N. Yan, N. Gao, G.M. Chen, Prediction of thermal conductivity of
conductivity of nanoparticle-fluid mixture – ScienceDirect, Int. J. Heat Mass various nanofluids with ethylene glycol using artificial neural network, J. Therm.
Transf. 48 (14) (2005) 2926–2932. Sci. 3 (29) (2019) 1504–1512.
[97] K.C. Leong, C. Yang, S. Murshed, A model for the thermal conductivity of [129] A. Komeilibirjandi, A.H. Raffiee, M.A. Nazari, M.S. Shadloo, Thermal conductivity
nanofluids – the effect of interfacial layer, J. Nanopart. Res. 8 (2) (2006) prediction of nanofluids containing CuO nanoparticles by using correlation and
245–254. artificial neural network, J. Therm. Anal. Calorim. 139 (4) (2020) 2679–2689.
[98] C. Sitprasert, P. Dechaumphai, V. Juntasaro, A thermal conductivity model for [130] J. Wang, Y.L. Zhai, P.T. Yao, M.Y. Ma, H. Wang, Established prediction models of
nanofluids including effect of the temperature-dependent interfacial layer, thermal conductivity of hybrid nanofluids based on artificial neural network
J. Nanopart. Res. 11 (6) (2008) 1465–1476. (ANN) models in waste heat system – ScienceDirect, Int. Commun. Heat Mass
[99] R. Alipour, M. Ghoranneviss, M. Mirzaee, A. Jafari, The direct effect of interfacial Transf. 110 (2020) 104444.
nanolayers on thermal conductivity of nanofluids, Heat Mass Transf. 50 (12) [131] A.H. Elsheikh, S.W. Sharshir, A.S. Ismai, R. Sathyamurthy, H. Zhang, An artificial
(2014) 1727–1735. neural network-based approach for prediction the thermal conductivity of
[100] L. Yang, X. Xu, A renovated Hamilton–Crosser model for the effective thermal nanofluids, SN Appl. Sci. 2 (2020) 235.
conductivity of CNTs nanofluids, Int. Commun. Heat Mass Transf. 81 (2017) [132] M. Sadeghzadeh, H. Maddah, M.H. Ahmadi, A. Khadang, M. Ghazvini, A. Mosavi,
42–50. N. Nabipour, Prediction of thermo-physical properties of TiO2-Al2O3/water
[101] L. Xue, P. Keblinski, S.R. Phillpot, S.U.S. Choi, J.A. Eastman, Effect of liquid nanoparticles by using artificial neural network, Nanomaterials 10 (2020) 4.
layering at the liquid-solid interface on thermal transport, Int. J. Heat Mass [133] Y.C. Li, R. Kalbasi, A. Karimipour, M. Sharifpur, J. Meyer, Using of artificial
Transf. 47 (19–20) (2004) 4277–4284. neural networks (ANNs) to predict the rheological behavior of magnesium oxide-
[102] X. Wang, X. Xu, Thermal conductivity of nanoparticle – fluid mixture, water nanofluid in a different volume fraction of nanoparticles, temperatures, and
J. Thermophys. Heat Transf. 13 (13) (2012) 474–480.

22
J. Li et al. International Communications in Heat and Mass Transfer 127 (2021) 105543

shear rates, Math. Methods Appl. Sci. (2020), https://doi.org/10.1002/ [152] J.S. Wu, X.X. Zhang, J.C. Shen, Y.P. Wu, K. Connelly, T. Yang, L. Tang, M.X. Xiao,
mma.6418. Y.X. Wei, K. Jiang, A review of thermal absorbers and their integration methods
[134] L. Li, Y.L. Zhai, Y.J. Jin, J. Wang, H. Wang, M.Y. Ma, Stability, thermal for the combined solar photovoltaic/thermal (PV/T) modules, Renew. Sust.
performance and artificial neural network modeling of viscosity and thermal Energ. Rev. 75 (2017) 839–854.
conductivity of Al2Oe-ethylene glycol nanofluids, Powder Technol. 363 (2020) [153] Z. Said, S. Arora, E. Bellos, A review on performance and environmental effects of
360–368. conventional and nanofluid-based thermal photovoltaics, Renew. Sust. Energ.
[135] N.B. Zhao, Z.M. Li, Experiment and artificial neural network prediction of thermal Rev. 96 (2018) 302–316. OCT.
conductivity and viscosity for alumina-water Nanofluids, Materials 10 (2017) [154] A.H.A. Al-Waeli, M.T. Chaichan, K. Sopian, H.A. Kazem, H.B. Mahood, A.
552. A. Khadom, Modeling and experimental validation of a PVT system using
[136] A. Kumar, M.A. Hassan, P. Chand, Heat transport in nanofluid coolant car radiator nanofluid coolant and nano-PCM, Sol. Energy 177 (2019) 178–191. JAN.
with louvered fins, Powder Technol. 376 (2020) 631–642. [155] C. Erdem, P.M. Cuce, T. Guclu, A.B. Besir, On the use of nanofluids in solar energy
[137] S.P. Jang, S. Choi, Cooling performance of a microchannel heat sink with applications, J. Therm. Sci. 29 (2020) 1–22.
nanofluids, Korean J. Air-Condit. Refrig. Eng. 26 (17) (2006) 2457–2463. [156] S. Soltani, A. Kasaeian, H. Sarrafha, D.S. Wen, An experimental investigation of a
[138] S. Wu, D. Zhu, X. Li, H. Li, J. Lei, Thermal energy storage behavior of AleOe-HeO hybrid photovoltaic/thermoelectric system with nanofluid application, Sol.
nanofluids, Thermochim. Acta 483 (2009) 73–77. Energy 155 (2017) 1033–1043.
[139] M. Ali, A.M. El-Leathy, Z. Al-Sofyany, The effect of nanofluid concentration on the [157] K.V. Wong, O.D. Leon, Applications of nanofluids: current and future, Adv. Mech.
cooling system of vehicles radiator, Adv. Mech. Eng. 6 (2014) 962510. Eng. 2010 (2) (2010) 1652–1660.
[140] V. Salamon, D.S. Kumar, S. Thirumalini, Experimental investigation of heat [158] G. Jiang, X. Ni, L. Yang, W. Li, Y. Li, Z. Deng, Synthesis of superamphiphobic
transfer characteristics of automobile radiator using TiOe-nanofluid coolant, IOP nanofluid as a multi-functional additive in oil-based drilling fluid, especially the
Conf. Ser. Mater. Sci. Eng. 225 (1) (2017), 012101. stabilization performance on the water/oil interface, Colloids Surf. A
[141] S.A. Ahmed, M. Ozkaymak, A. Sozen, T. Menlik, A. Fahed, Improving car radiator Physicochem. Eng. Asp. 588 (2019) 124385.
performance by using TiO2-water nanofluid, Eng. Sci. Technol. Int. J. 21 (5) [159] S. Sarvar-Ardeh, R. Rafee, S. Rashidi, Hybrid nanofluids with temperature-
(2018) 996–1005. dependent properties for use in double-layered microchannel heat sink;
[142] S.K. Mohammadi, S. Etemad, J. Thibault, Measurement of thermal properties of hydrothermal investigation, J. Taiwan Inst. Chem. Eng. 124 (2021) 53–62.
suspensions of nanoparticles in engine oil, Nanotechnology (2009) 74–77, 2009: [160] H. Zhang, Q. Shan, Y. Zhai, X. Zhang, A. Zhang, The changes induced by pH in
Technical proceedings. TiO2/water nanofluids: stability, thermophysical properties and thermal
[143] M. Vasheghani, Enhancement of the thermal conductivity and viscosity of performance, Powder Technol. 377 (2021) 748–759.
aluminum component-engine oil nanofluids, Int. J. Nanomech. Sci. Technol. 3 (4) [161] X.L. Zhang, Y.L. Zhang, Heat transfer and flow characteristics of Fe3O4 -water
(2012) 333–340. nanofluids under magnetic excitation, Int. J. Therm. Sci. 163 (2021) 106826.
[144] D. Davis, S. Saleem, F. Shah, S. Singh, Development of ethylene glycol-Cr2AlC [162] X.L. Zhang, Y.L. Zhang, Experimental study on enhanced heat transfer and flow
nanofluid for thermal management in the automotive sector, Int. J. Appl. Ceram. performance of magnetic nanofluids under alternating magnetic field, Int. J.
Technol. 17 (2020) 1071–1078. Therm. Sci. 164 (2021) 106897.
[145] S. Wiriyasart, C. Hommalee, S. Sirikasemsuk, R. Prurapark, P. Naphon, Thermal [163] S. Chakraborty, I. Sarkar, D.K. Behera, S.K. Pal, Experimental investigation on the
management system with nanofluids for electric vehicle battery cooling modules, effect of dispersant addition on thermal and rheological characteristics of TiO2
Case Stud. Therm. Eng. 18 (2020) 100583. nanofluid, Powder Technology 307 (2016) 10–24.
[146] N. Putra, F.N. Iskandar Yanuar, Application of nanofluids to a heat pipe liquid- [164] E.O. Sheremet, V.A. Uvarov, N.Y. Nikulin, et al., Prospects for the use of
block and the thermoelectric cooling of electronic equipment, Exp. Thermal Fluid nanofluids in heating systems, J. Phys. Conf. Ser. 2021 (1926), 012033 (6pp).
Sci. 35 (7) (2011) 1274–1281. [165] F. Rubbi, K. Habib, R. Saidur, et al., Performance optimization of a hybrid PV/T
[147] A.K. Hussein, Applications of nanotechnology to improve the performance of solar system using Soybean oil/MXene nanofluids as a new class of heat transfer
solar collectors – recent advances and overview, Renew. Sust. Energ. Rev. 62 fluids, Sol. Energy 208 (2020) 124–138.
(2016) 767–792. [166] W. Han, B. Dai, J. Liu, et al., Ignition and combustion characteristics of heptane-
[148] A.H. Elsheikh, S.W. Sharshir, M.E. Mostafa, F.A. Essa, M.K.A. Ali, Applications of based nanofluid fuel droplets, Energy Fuel 33 (2019) 10282–10289.
nanofluids in solar energy: a review of recent advances, Renew. Sust. Energ. Rev. [167] S. Li, H. Du, Z. Zhuo, et al., Dispersion stability, physical properties, electrostatic
82 (2017) 3483–3502. breakup of surfactant-loaded Al/n-decane nanofluid fuel: nanoparticle size effect,
[149] M.A. Fikri, F.F. Asri, W.M. Faizal, H.K. Adli, R. Mamat, W.H. Azmi, G. Najafi, Energy Fuel 34 (2019) 1082–1092.
T. Yusaf, TiO2 -SiO2 nanofluid characterization: towards efficient with water/ [168] A. Nilankush, D. Kalidas, K.P. Kumar, On the heat transport mechanism and
ethylene glycol mixture for solar application, IOP Conf. Ser. Mater. Sci. Eng. 863 entropy generation in a nozzle of liquid rocket engine using ferrofluid: a
(2020), 012055. computational framework, J. Comput. Des. Eng. 4 (2019) 739–750.
[150] Z. Zheng, Q. He, S. Yin, Experimental investigation on efficiency of flat-plate solar [169] H.B. Kulkarni, M.M. Nadakatti, S.C. Kulkarni, et al., Investigations on effect of
collectors with Cu-HeO nanofluids, Taiyangneng Xuebao/Acta Energiae Solaris nanofluid based minimum quantity lubrication technique for surface milling of
Sin. 36 (3) (2015) 562–567. Al7075-T6 aerospace alloy, Mater. Today Proc. 27 (2019).
[151] L. Lin, Z.H. Liu, H.S. Xiao, Thermal performance of an open thermosyphon using
nanofluids for high-temperature evacuated tubular solar collectors: Part 1: indoor
experiment, Sol. Energy 85 (2) (2011) 379–387.

23