Physica A 534 (2019) 122142

Contents lists available at ScienceDirect

Physica A
journal homepage: www.elsevier.com/locate/physa

Designing an Artificial Neural Network (ANN) to predict the

viscosity of Silver/Ethylene glycol nanofluid at different
temperatures and volume fraction of nanoparticles
Davood Toghraie a , Nima Sina b , Niyusha Adavoodi Jolfaei c , Mehdi Hajian d ,
∗
Masoud Afrand e,f ,
a
Department of Mechanical Engineering, Khomeinishahr Branch, Islamic Azad University, Khomeinishahr, Iran
b
Department of Mechanical Engineering, Najafabad Branch, Islamic Azad University, Najafabad, Iran
c
Department of Pharmaceutical Sciences, KLE College of Pharmacy, 2nd block, Rajajinagar, Bengaluru, Karnatka, India
d
Department of Mechanical Engineering, Isfahan University of Technology, Isfahan, 84156-83111, Iran
e
Laboratory of Magnetism and Magnetic Materials, Advanced Institute of Materials Science, Ton Duc Thang University, Ho Chi
Minh City, Vietnam
f
Faculty of Applied Sciences, Ton Duc Thang University, Ho Chi Minh City, Vietnam

article info a b s t r a c t

Article history: In the current work, we investigate the dynamic viscosity of Ag/Ethylene glycol nanofluid
Received 6 April 2019 within the temperature range of 25–55 ◦ C with volume fraction of nanoparticles range
Received in revised form 17 May 2019 of 0.2%–2%. The experimental data includes 42 samples. At first, an Artificial Neural
Available online 7 August 2019
Network (ANN) is designed to predict the dynamic viscosity of this nanofluid and finally
Keywords: the results of ANN and correlation has been compared. The algorithm of generating the
Artificial Neural Network best architecture of ANN has been proposed and the best ANN has been used to predict
Viscosity the dynamic viscosity of Silver/Ethylene glycol nanofluid. It is found that the ANN can
Nanofluid predict the viscosity of Ag/Ethylene glycol nanofluid with good precision compared to
Silver the correlation method. Also, in the correlation method, MSE is 0.0012, SSE is 0.0512
Ethylene glycol and the maximum value of error is 0.0858.
Experimental data © 2019 Published by Elsevier B.V.

1. Introduction

Today, Artificial Neural Networks (ANNs) are used in predicting the behavior of complex systems. Also, ANNs have been
trained to predict the mechanical properties of nano-materials. Some researchers applied ANNs to predict the behavior
of nanofluids. ANN cannot be exactly categorized as an algorithm. ANN was published in 1940’s by Kenneth Levenberg
and Marquardt. ANN is an iterative algorithm in minimization. ANN is used in learning process [1–3].
In the recent years, the ability of ANN to estimate the viscosity of nanofluids has persuaded researchers to apply this
efficient tool in their research.
Torkar et al. [4] predicted a viscosity of Alumina–Paraffin suspensions with ANNs. They found a good coincidence be-
tween experimental and ANN data. Afrand et al. [5] developed an ANN approach to predict the viscosity of MWCNT/water
nanofluid. They showed that the ANN model is more accurate compared to other correlations. Using an ANN method,
Afrand et al. [6] predicted the viscosity of a hybrid nano-lubricant. They found that ANN is accurate compared to other

∗ Corresponding author.
E-mail address: [email protected] (M. Afrand).

https://doi.org/10.1016/j.physa.2019.122142
0378-4371/© 2019 Published by Elsevier B.V.
2 D. Toghraie, N. Sina, N.A. Jolfaei et al. / Physica A 534 (2019) 122142

Fig. 1. Viscosity of nanofluid versus volume fraction of silver nanoparticles at different temperatures [18].

empirical correlations. Esfe et al. [7] designed an ANN to predict viscosity of TiO2 /Water nanofluid. They obtained a
relationship to estimate the viscosity data. Their correlation was able to estimate the viscosity with a very good precision.
Bispo et al. [8] developed an ANN method to estimate the viscosity of nanofluids. They showed that their selected ANNs
and the neuron model can predict the viscosity. Longo et al. [9] modeled the viscosity of nanofluids with an ANN model.
Using experimental data and ANN, Vakili et al. [10] estimated the viscosity of Graphene/water nanofluid. Their results
show that the presented model is compatible with experimental work. Hashemi Fath et al. [11] showed that the outputs
of ANN model were in agreement with experimental data. Derakhshanfard and Mehralizadeh [12] applied an ANNs for
viscosity of oxides nanoparticles-crude oil-based nanofluids. By using of an experimental study and ANNs, Esfe and
Arani [13] predicted the viscosity of nanofluid. The results showed that the ANNs can estimate the dynamic viscosity.
Mohamed [14] estimated the properties of nanofluid with ANNs. Their represent an excellent accuracy and prove that
ANN model has a great ability to model thermal conductivity and viscosity of nanofluids. Amani et al. [15] modeled the
properties of nanofluid under using an ANN method. They obtained a good agreement between the experimental data
and ANN data. Aghayari et al. [16] modeled the conductivity of CuO/glycerol nanofluids at various conditions using ANN
method. Zhao and Li [17] predicted the properties of alumina–water nanofluids. They revealed that the ANN can predict
the properties of Al2 O3 -Water nanofluids.
In the present study, we design an ANN method and predict the viscosity of Ag/Ethylene glycol nanofluid at different
temperatures and volume fraction of nanoparticles. Viscosity of Ag/Ethylene glycol nanofluid has been studied within
the temperature range of 25 ◦ C–55 ◦ C and volume fraction range of 0.25%–2%. The algorithm of generating the best
architecture of ANN has been proposed and the best ANN has been used to predict the viscosity of Silver/Ethylene
glycol nanofluid. As the ANN should predict only one parameter (dynamic viscosity), only one neuron is required in the
output layer. In order to get the best neuron number, different neuron number in the hidden layer has been applied. To
the author’s knowledge, there is no comprehensive and thorough investigation to predict the dynamics viscosity of the
supposed nanofluid by using of ANN method. It is noteworthy that the dynamic viscosity of Ag/Ethylene glycol nanofluid
is measured in our previous investigation [18]. The purpose of this investigation is design an ANN method to predict the
viscosity of supposed nanofluid at different conditions.

2. Experimental data

Increasing viscosity can significantly increase the pumping power. Therefore, the design of fluid systems depends on
the viscosity of the fluids. The addition of nanoparticles to the base fluid significantly increases the viscosity of base fluid.
In laboratory work done in 2018 by authors [18], the dynamic viscosity of Ag/Ethylene glycol nanofluid was measured.
Fig. 1 shows the dynamic viscosity versus volume fraction of Silver nanoparticles at different temperatures [18] for
Ag/Ethylene glycol nanofluid. According to this figure, viscosity of Ag/Ethylene glycol nanofluid increases with increasing
the volume fraction of nanoparticles; whereas, the viscosity of Ag/Ethylene glycol nanofluid decreases with increasing
temperature.
Fig. 2 shows the viscosity of Ag/Ethylene glycol nanofluid versus volume fraction of nanoparticles [18]. This figure
shows the effect of temperature increase on viscosity reduction at different volume fraction of nanoparticles.
 D. Toghraie, N. Sina, N.A. Jolfaei et al. / Physica A 534 (2019) 122142 3

Fig. 2. Viscosity of nanofluid versus temperature at different volume fraction of silver nanoparticles [18].

3. Artificial neural network (ANN)

3.1. The architecture of ANN

An ANN is made up of simple parts which are called neurons. Neurons have activation function. The input signal is
weighted by a vector and then added by a bias. Then this signal goes to the activation function and the output of neuron
is reached. In the current work the learning algorithm is Levenberg–Marquardt which is called damped least-square.
In the hidden layer the activation function is tangent sigmoid which is presented in Eq. (1).
2
tan sig(x) = −1 (1)
1 + e−2x
The activation function of the output layer is purelin. In purelin, the output value is exactly the input value [19–22].
As the ANN should predict only one parameter (dynamic viscosity of Silver/Ethylene glycol nanofluid), only one neuron
is required in the output layer. To get the best neuron number, different neuron number in the hidden layer has been
applied. It should be considered that the processes of generating coefficients in the ANN are random. Therefore, the ANN
may show different output results. In order to reduce the risk of uncertainty, using an internal loop in the process of
selecting the best ANN can be very useful. The proposed algorithm is presented in Fig. 3.
In the presented algorithm, after reading the experimental data set and applying some ANN, settings such as defining
the activation function, the percentage of train and test dataset, the main loop initiates predicting the data using 6 neurons
to 31 neurons. For each neuron number the ANN simulates for 30 times (considering inner iteration number), then the
best network is selected based on the best performance. In the current work, the best neuron number is 17. The Best ANN
is depicted in Fig. 4
The results of the selected algorithm are presented in Table 1, in which the results are sorted based on the best
performance. In Eq. (2) the process of neuron is shown in the mathematical format.
( m
)
∑
yk = ϕ wkj xj + bk (2)
i=1

In Eq. (2), xj is considered as input value, yk is output value, wkj is weighting factor, bk is considered as bias and ϕ is
the activating function. The experimental data is used for training the ANN. In designing ANN, minimizing the error is
very important. Error can be defined as the absolute value of the difference between the experimental data and the ANN
output [23–27]. In some cases, the Mean Square Error (MSE) which is called the performance is considered as a criterion
in analyzing the ANN. The MSE is shown in Eq. (3)
N
1 ∑
MSE = (yexp − yANN )2 (3)
N
i=1
4 D. Toghraie, N. Sina, N.A. Jolfaei et al. / Physica A 534 (2019) 122142

Fig. 3. Flowchart of finding the best ANN.

 D. Toghraie, N. Sina, N.A. Jolfaei et al. / Physica A 534 (2019) 122142 5

Fig. 4. The best ANN.

Table 1
The results of the selected algorithm.
Neuron number MSE (All) MSE (Train) MSE (Test)
17 6.96517E−05 3.1E−13 0.000272127
16 7.70675E−05 6.22757E−13 0.000301101
21 8.02467E−05 9.78124E−13 0.000313522
24 8.13549E−05 4.12223E−13 0.000317852
12 8.49949E−05 5.25361E−13 0.000332073
20 6.96856E−05 8.7185E−13 0.000347738
19 7.03072E−05 3.43776E−13 0.00035084
18 7.25885E−05 5.92114E−13 0.000362224
13 7.34105E−05 5.7545E−13 0.000366326
14 7.48823E−05 3.08937E−13 0.00037367
27 7.709E−05 7.43559E−13 0.000384687
26 7.74752E−05 3.08155E−13 0.000386609
22 7.83258E−05 1.84791E−13 0.000390854
25 8.38542E−05 4.41437E−13 0.000418441
23 8.52098E−05 1.74497E−12 0.000425205
10 8.85572E−05 0.00029208 0.000441898
29 9.84091E−05 7.81431E−13 0.000491071
30 9.90913E−05 2.57822E−13 0.000494475
15 0.000111069 1.63089E−12 0.000554248
28 0.00011273 1.6286E−13 0.000562533
11 0.00012636 0.000212011 0.00063054
31 0.000128857 4.49954E−13 0.000643008
8 0.000257789 0.026915786 0.001285365
6 0.005714375 0.144032814 0.02850979
9 0.005723993 0.004985948 0.0285631
7 1734.885379 0.072896895 8657.249811

In which yexp is the experimental data and yANN is the output value of ANN. Also N is the number of datapoints.
An ANN with smallest MSE can be chosen as the best one. As the number of experimental data is limited, it would
be better to divide the dataset into two main categories including train and test. 75% of data set is used for train and 25
percent for the test.
The train output data is shown in Fig. 5. The train data and output values are compatible. Also, these points are located
near the compatibility line. The compatibility line is shown by the blue color. For the train data, it can be seen that MSE
is 3.1000e−13, SSE is 9.9201e−12 and maximum error is 1.5911e−06.
The test output data is shown in Fig. 6. ANN are judged by their performance and for calculating the performance, the
test dataset is used. These points are near the compatibility line and it can be concluded that the performance of this
ANN is suitable. For the test data, it can be seen that MSE is 2.7151e−04 SSE is 0.0030 and maximum error is 0.0314.
All ANN output is shown in Fig. 7. All data points are near the compatibility line and this proves that the designed
ANN could predict the dynamic viscosity of Silver/Ethylene glycol Nano fluid. For all ANN output, it can be seen that MSE
is 6.9456e−05 SSE is 0.0030 and maximum error is 0.0314.
Viscosity of Ag/Ethylene glycol nanofluid versus data number for the both of experimental data and ANN outputs is
shown in Fig. 8. The experimental data has been followed by the ANN accurately. Therefore, this ANN can be used to
predict the viscosity of Ag/Ethylene glycol nanofluid.
6 D. Toghraie, N. Sina, N.A. Jolfaei et al. / Physica A 534 (2019) 122142

Fig. 5. The train data.

Fig. 6. The test data.

Fig. 7. All data.

 D. Toghraie, N. Sina, N.A. Jolfaei et al. / Physica A 534 (2019) 122142 7

Fig. 8. Dynamic viscosity of Ag/Ethylene glycol nanofluid versus data number: Experimental and ANN outputs.

Fig. 9. Dynamic viscosity of Ag/Ethylene glycol nanofluid versus data number: Experimental data and correlation values.

Viscosity of Ag/Ethylene glycol nanofluid versus data number for the both of experimental and correlation values is
shown in Fig. 9. In this figure the correlation method can predict the experimental data but for some data points the
difference between experimental values and correlation method is not small. Although this method is acceptable but it
is not as good as the ANN method.
Error can be defined as the absolute value of the difference between the experimental data and the ANN output. Error
in dynamic viscosity of Ag/Ethylene glycol nanofluid versus data number for ANN values and experimental data are shown
in Fig. 10. As can be seen, the distribution of the error is centered around zero value which proves the model’s accuracy.
The error for correlation method is larger than the ANN method. There are some fluctuations for both methods but
the range of error for the ANN method is smaller than the correlation method.

3.2. Correlation method

Considering temperature as x1, volume fraction of nanoparticles as x2, and applying curve fitting, the output function
can be presented as follows:

f(x1, x2) = p0 + p1∗ x1 + p2∗ x2 + p3∗ x1∧ 2 + p4∗ x1∗ x2 + p02∗ x2∧ 2 + p6∗ x1∧ 3 + p7∗ x1∧ 2∗ x2 + p8∗ x1∗ x2∧ 2 + p9∗ x2∧ 3
(4)
In which, p0 to p9 are coefficients of the fitted surface. The coefficients of the fitted surface p0 to p9 are presented in
Table 2. For correlation values, it can be seen that MSE is 0.0012, SSE is 0.0512 and maximum error is 0.0858.
The fitted surface is presented in Fig. 11. The experimental data points are shown by dots. A 3-D surface on the
experimental data is fitted. This surface is obtained using the correlation method. It can be seen that the correlational
method can predict the experimental exactly. In the correlation method, MSE is 0.0012, SSE is 0.0512 and the maximum
value of error is 0.0858.
8 D. Toghraie, N. Sina, N.A. Jolfaei et al. / Physica A 534 (2019) 122142

Fig. 10. Error in dynamic viscosity of Ag/Ethylene glycol nanofluid versus data number: ANN and correlation.

Fig. 11. The surface of correlation method.

Table 2
Coefficients of the fitted surface.
p0 = 2.677 p1 = −0.1416 p2 = −0.1597 p3 = 0.003586 p4 = 0.03183
p5 = −0.00437 p6 = −2.867e−05 p7 = −0.0003494 p8 = -0.0008629 p9 = 0.0133

4. Conclusion

In this work, viscosity of Ag/Ethylene glycol nanofluid has been studied within the temperature range of 25–55 ◦ C
and volume fractions range of 0.25%–2%. The experimental data includes 42 samples. For these ranges using experimental
methods, the viscosity of Ag/Ethylene glycol nanofluid has been calculated in the laboratory [18]. The proposed algorithm
is used to generate the optimum ANN and predict the dynamic viscosity of Silver/Ethylene nanofluid based on the
temperature and volume fraction of nanoparticles as input values. In the next step, a correlation method has been used to
calculate the dynamic viscosity and then the ANN results and correlation results have been compared. We concluded that
the ANN can predict the viscosity of Ag/Ethylene nanofluid more accurate compared to the correlation method. Also, in
the correlation method, MSE is 0.0012, SSE is 0.0512 and the maximum value of error is 0.0858. Based on the presented
results, this network can predict the behavior of nanofluid. Therefore, by avoiding experiments, it can decrease the lab
costs in calculating the dynamic viscosity of nanofluid.
The results of this paper and previous studies [28–59] can be used for nanofluid investigation.

References

[1] Artificial Neural Networks as Models of Neural Information Processing | Frontiers Research Topic. Retrieved 2018, 02-20.
[2] Behnke Sven, Hierarchical Neural Networks for Image Interpretation, in: Lecture Notes in Computer Science, vol. 2766, 2003.
[3] C. Ferreira, Designing neural networks using gene expression programming, in: A. Abraham, B. de Baets, M. Köppen, B. Nickolay (Eds.), Applied
Soft Computing Technologies: The Challenge of Complexity, 2006, pp. 517–536.
 D. Toghraie, N. Sina, N.A. Jolfaei et al. / Physica A 534 (2019) 122142 9

[4] Torkar Drago, Novak Sasa, Novak Franc, Apparent viscosity prediction of alumina–paraffin suspensions using artificial neural networks, J. Mater
Process. Technol. 203 (2008) 208–215.
[5] Masoud Afrand, Afshin Ahmadi Nadooshan, Mohsen Hassani, Hooman Yarmand, M. Dahari, Predicting the viscosity of multi-walled carbon
nanotubes/water nanofluid by developing an optimal artificial neural network based on experimental data, Int. Commun. Heat Mass Transfer
77 (2016) 49–53.
[6] Masoud Afrand, Karim Nazari Najafabadi, Nima Sina, Mohammad Reza Safaei, A.Sh. Kherbeet, Somchai Wongwises, Mahidzal Dahari, Prediction
of dynamic viscosity of a hybrid nano-lubricant by an optimal artificial neural network, Int. Commun. Heat Mass Transfer 76 (2016) 209–214.
[7] Mohammad Hemmat Esfe, Mohammad Reza Hassani Ahangar, Mousa Rejvani, Davood Toghraie, Mohammad Hadi Hajmohammad, Designing an
artificial neural network to predict dynamic viscosity of aqueous nanofluid of TiO2 using experimental data, Int. Commun. Heat Mass Transfer
75 (2016) 192–196.
[8] Vitor Diego da Silva Bispo, Cláudia Miriam Scheid, Luís Américo Calçada, Luiz Augusto da Cruz Meleiro, Development of an ANN-based soft-sensor
to estimate the apparent viscosity of water-based drilling fluids, J. Pet. Sci. Eng. 150 (2017) 69–73.
[9] Giovanni A. Longo, Claudio Zilio, Ludovico Ortombina, Mauro Zigliotto, Application of Artificial Neural Network (ANN) for modeling oxide-based
nanofluids dynamic viscosity, Int. Commun. Heat Mass Transfer 83 (2017) 8–14.
[10] M. Vakili, S. Khosrojerdi, P. Aghajannezhad, M. Yahyaei, A hybrid artificial neural network-genetic algorithm modeling approach for viscosity
estimation of graphene nanoplatelets nanofluid using experimental data, Int. Commun. Heat Mass Transfer 82 (2017) 40–48.
[11] Aref Hashemi Fath, Abdolrasoul Pouranfard, Pouyan Foroughizadeh, Development of an artificial neural network model for prediction of bubble
point pressure of crude oils, Petroleum 4 (2018) 281–291.
[12] Fahimeh Derakhshanfard, Amir Mehralizadeh, Application of artificial neural networks for viscosity of crude oil-based nanofluids containing
oxides nanoparticles, J. Pet. Sci. Eng. 168 (2018) 263–272.
[13] Mohammad Hemmat Esfe, Ali Akbar Abbasian Arani, An experimental determination and accurate prediction of dynamic viscosity of MWCNT
(%40)-SiO2 (%60)/5W50 nano-lubricant, J. Molecular Liquids 259 (2018) 227–237.
[14] R.A. Mohamed, Application of artificial neural network model for prediction of thermo–physical properties of carbon nanotubes (CNTs)
containing nanofluid, J. Nanofluids 8 (1) (2019) 199–204, (6).
[15] Mohammad Amani, Pouria Amani, Alibakhsh Kasaeian, Omid Mahian, Ioan Pop, Somchai Wongwises, Modeling and optimization of thermal
conductivity and viscosity of MnFe2O4 nanofluid under magnetic field using an ANN, Sci. Rep. (2017) http://dx.doi.org/10.1038/s41598-017-
17444-5.
[16] Reza Aghayari, Heydar Maddah, Mohammad Hossein Ahmadi, Wei-Mon Yan, Nahid Ghasemi, Measurement and artificial neural network
modeling of electrical conductivity of cuo/glycerol nanofluids at various thermal and concentration conditions, Energies 11 (2018) 1190,
http://dx.doi.org/10.3390/en11051190.
[17] N. Zhao, Z. Li, Experiment and artificial neural network prediction of thermal conductivity and viscosity for alumina-water nanofluids, Materials
(Basel) 10 (5) (2017).
[18] A.D. Zadeh, D. Toghraie, Experimental investigation for developing a new model for the dynamic viscosity of silver/ethylene glycol nanofluid
at different temperatures and solid volume fractions, J. Therm. Anal. Calorim. 131 (2) (2018) 1449–1461.
[19] H. Eshgarf, N. Sina, M.H. Esfe, F. Izadi, M. Afrand, Prediction of rheological behavior of MWCNTs–SiO2/EG–water non-Newtonian hybrid nanofluid
by designing new correlations and optimal artificial neural networks, J. Therm. Anal. Calorim. 132 (2) (2018) 1029–1038.
[20] M. Afrand, D. Toghraie, N. Sina, Experimental study on thermal conductivity of water-based Fe3O4 nanofluid: development of a new correlation
and modeled by artificial neural network, Int. Commun. Heat Mass Transfer 75 (2016) 262–269.
[21] M. Vafaei, M. Afrand, N. Sina, R. Kalbasi, F. Sourani, H. Teimouri, Evaluation of thermal conductivity of MgO-MWCNTs/EG hybrid nanofluids
based on experimental data by selecting optimal artificial neural networks, Physica E 85 (2017) 90–96.
[22] M.H. Esfe, S. Saedodin, N. Sina, M. Afrand, S. Rostami, Designing an artificial neural network to predict thermal conductivity and dynamic
viscosity of ferromagnetic nanofluid, Int. Commun. Heat Mass Transfer 68 (2015) 50–57.
[23] M.H. Esfe, S. Saedodin, M. Bahiraei, D. Toghraie, O. Mahian, S. Wongwises, Thermal conductivity modeling of MgO/EG nanofluids using
experimental data and artificial neural network, J. Therm. Anal. Calorim. 118 (1) (2014) 287–294.
[24] M.H. Esfe, W.M. Yan, M. Afrand, M. Sarraf, D. Toghraie, M. Dahari, Estimation of thermal conductivity of Al2 O3 /water (40%)–ethylene glycol
(60%) by artificial neural network and correlation using experimental data, Int. Commun. Heat Mass Transfer 74 (2016) 125–128.
[25] M.H. Esfe, H. Rostamian, D. Toghraie, W.M. Yan, Using artificial neural network to predict thermal conductivity of ethylene glycol with alumina
nanoparticle, J. Therm. Anal. Calorim. 126 (2) (2016) 643–648.
[26] Mohammad Hemmat Esfe, Mohammad Reza Hassani Ahangar, Davood Toghraie, Mohammad Hadi Hajmohammad, Hadi Rostamian, Hossein
Tourang, Designing artificial neural network on thermal conductivity of Al2 O3 –water–EG (60%–40%) nanofluid using experimental data, J. Therm.
Anal. Calorim. 126 (2) (2016) 837–843.
[27] M.R. Faridzadeh, D. Toghraie, A. Niroomand, Analysis of laminar mixed convection in an inclined square lid-driven cavity with a nanofluid by
using an artificial neural network, Heat Transfer Res. 45 (4) (2014).
[28] Q. Gravndyan, O.A. Akbari, D. Toghraie, A. Marzban, R. Mashayekhi, R. Karimi, F. Pourfattah, The effect of aspect ratios of rib on the heat
transfer and laminar water/TiO 2 nanofluid flow in a two-dimensional rectangular microchannel, J. Molecular Liquids 236 (2017) 254–265.
[29] M.R. Shamsi, O.A. Akbari, A. Marzban, D. Toghraie, R. Mashayekhi, Increasing heat transfer of non-Newtonian nanofluid in rectangular
microchannel with triangular ribs, Physica E 93 (2017) 167–178.
[30] M. Zarringhalam, A. Karimipour, D. Toghraie, Experimental study of the effect of solid volume fraction and Reynolds number on heat transfer
coefficient and pressure drop of CuO–Water nanofluid, Exp. Therm. Fluid Sci. 76 (2016) 342–351.
[31] D. Toghraie, V.A. Chaharsoghi, M. Afr, Measurement of thermal conductivity of ZnO–TiO2/EG hybrid nanofluid, J. Therm. Anal. Calorim. (2016)
1–9, http://dx.doi.org/10.1007/s10973-016-5436-4.
[32] D. Toghraie, S.M.B. Alempour, M. Afr, Experimental determination of viscosity of Water based magnetite nanofluid for application in heating
and cooling systems, J. Magn. Magn. Mater. 417 (2016) 243–248.
[33] O.A. Akbari, A. Karimipour, D. Toghraie, A. Karimipour, Impact of ribs on flow parameters and laminar heat transfer of Water–Aluminum oxide
nanofluid with different nanoparticle volume fractions in a three-dimensional rectangular microchannel, Adv. Mech. Eng. 7 (2016) 1–11.
[34] O.A. Akbari, A. Karimipour, D. Toghraie, M.R. Safaei, M. Alipour, H. Goodarzi, M. Dahari, Investigation of Rib’s height effect on heat transfer and
flow parameters of laminar water- Al2O3 nanofluid in a two dimensional rib-microchannel, Appl. Math. Comput. 290 (2016) 135–153.
[35] O.A. Akbari, D. Toghraie, A. Karimipour, Numerical simulation of heat transfer and turbulent flow of Water nanofluids copper oxide in rectangular
microchannel with semi attached rib, Adv. Mech. Eng. 8 (2016) 1–25.
[36] S. Nazari, D. Toghraie, Numerical simulation of heat transfer and fluid flow of Water-CuO Nanofluid in a sinusoidal channel with a porous
medium. Physica E, 123; 87: 134-140.
[37] A. Aghanajaf, D. Toghraie, B. Mehmandoust, Numerical simulation of laminar forced convection of Water-CuO nanofluid inside a triangular
duct, Physica E 85 (2017) 103–108.
10 D. Toghraie, N. Sina, N.A. Jolfaei et al. / Physica A 534 (2019) 122142

[38] D. Toghraie, M.M.D. Abdollah, F. Pourfattah, O.A. Akbari, B. Ruhani, Numerical investigation of flow and heat transfer characteristics in
smooth, sinusoidal and zigzag-shaped microchannel with and without nanofluid, J. Therm. Anal. Calorim. 131 (2) (2018) 1757–1766, http:
//dx.doi.org/10.1007/s10973-017-6624-6.
[39] A. Arabpour, A. Karimipour, D. Toghraie, The study of heat transfer and laminar flow of kerosene/multi-walled carbon nanotubes (MWCNTs)
nanofluid in the microchannel heat sink with slip boundary condition, J. Therm. Anal. Calorim. 131 (2) (2018) 1553–1566, http://dx.doi.org/10.
1007/s10973-017-6649-x.
[40] M.A. Esfahani, D. Toghraie, Experimental investigation for developing a new model for the thermal conductivity of Silica/Water-Ethylene glycol
(40%–60%) nanofluid at different temperatures and solid volume fractions, J. Molecular Liquids 232 (2017) 105–112.
[41] M. Zadkhast, D. Toghraie, A. Karimipour, Developing a new correlation to estimate the thermal conductivity of mwcnt-cuo/water hybrid
nanofluid via an experimental investigation, J. Therm. Anal. Calorim. 129 (2) (2017) 859–867, http://dx.doi.org/10.1007/s10973-017-6213-8.
[42] M. Heydari, D. Toghraie, O.A. Akbari, The effect of semi-attached and offset mid-truncated ribs and Water/TiO2 nanofluid on flow and heat
transfer properties in a triangular microchannel, Therm. Sci. Eng. Prog. 2 (2017) 140–150.
[43] N.N. Esfahani, D. Toghraie, M. Afrand, A new correlation for predicting the thermal conductivity of ZnO–Ag (50%–50%)/water hybrid nanofluid:
an experimental study, Powder Technol. 323 (2018) 367–373.
[44] Mohammad Hemmat Esfe, Hadi Hajmohammad, Davood Toghraie, Hadi Rostamian, Omid Mahian, Somchai Wongwises, Multi-objective
optimization of nanofluid flow in double tube heat exchangers for applications in energy systems, Energy 137 (2017) 160–171.
[45] F. Pourfattah, M. Motamedian, G. Sheikhzadeh, D. Toghraie, O.A. Akbari, The numerical investigation of angle of attack of inclined rectangular
rib on the turbulent heat transfer of Water-Al2 O3 nanofluid in a tube, Int. J. Mech. Sci. 131 (2017) 1106–1116.
[46] M. Afrand, E. Abedini, H. Teimouri, How the dispersion of magnesium oxide nanoparticles effects on the viscosity of water-ethylene glycol
mixture: experimental evaluation and correlation development, Physica E 87 (2017) 273–280.
[47] M. Akbari, M. Afrand, A. Arshi, A. Karimipour, An experimental study on rheological behavior of ethylene glycol based nanofluid: proposing a
new correlation as a function of silica concentration and temperature, J. Molecular Liquids 233 (2017) 352–357.
[48] M.A. Ali Karimi, Numerical study on thermal performance of an air-cooled heat exchanger: Effects of hybrid nanofluid, pipe arrangement and
cross section, Energy Convers. Manage. 164 (2018) 615–628.
[49] A. Asadi, M. Asadi, A. Rezaniakolaei, L.A. Rosendahl, M. Afrand, S. Wongwises, Heat transfer efficiency of Al2O3-MWCNT/thermal oil hybrid
nanofluid as a cooling fluid in thermal and energy management applications: An experimental and theoretical investigation, Int. J. Heat Mass
Transfer 117 (2018) 474–486.
[50] R.A. Dehkordi, M.H. Esfe, M. Afrand, Effects of functionalized single walled carbon nanotubes on thermal performance of antifreeze: an
experimental study on thermal conductivity, Appl. Therm. Eng. 120 (2017) 358–366.
[51] M.H. Esfe, M. Bahiraei, M.H. Hajmohammad, M. Afrand, Rheological characteristics of MgO/oil nanolubricants: experimental study and neural
network modeling, Int. Commun. Heat Mass Transfer 86 (2017) 245–252.
[52] M.H. Esfe, M. Firouzi, M. Afrand, Experimental and theoretical investigation of thermal conductivity of ethylene glycol containing functionalized
single walled carbon nanotubes, Physica E 95 (2018) 71–77.
[53] M.H. Esfe, K. Motahari, E. Sanatizadeh, M. Afrand, H. Rostamian, M.R.H. Ahangar, Estimation of thermal conductivity of CNTs-water in low
temperature by artificial neural network and correlation, Int. Commun. Heat Mass Transfer 76 (2016) 376–381.
[54] H. Khodadadi, D. Toghraie, A. Karimipour, Effects of nanoparticles to present a statistical model for the viscosity of MgO-Water nanofluid,
Powder Technol. 342 (2019) 166–180.
[55] A.A. Nadooshan, H. Eshgarf, M. Afrand, Measuring the viscosity of Fe3O4-MWCNTs/EG hybrid nanofluid for evaluation of thermal efficiency:
Newtonian and non-Newtonian behavior, J. Molecular Liquids 253 (2018) 169–177.
[56] A. Akhgar, D. Toghraie, Sina N., Afrand M., Developing dissimilar artificial neural networks (ANNs) to prediction the thermal conductivity of
MWCNT-TiO2 /ethylene glycol hybrid nanofluid, Powder Technol. 335 (2019) 602–610.
[57] B. Ruhani, D. Toghraie, M. Hekmatifar, M. Hadian, Statistical investigation for developing a new model for rheological behavior of ZnO-Ag
(50%-50%)/Water hybrid Newtonian nanofluid using experimental data, Physica A 525 (2019) 741–751.
[58] B. Ruhani, P. Barnoon, D. Toghraie, Statistical investigation for developing a new model for rheological behavior of Silica-ethylene glycol/Water
hybrid Newtonian nanofluid using experimental data, Physica A 525 (2019) 616–627.
[59] D.T. Semiromi, A.R. Azimian, Molecular dynamics simulation of annular flow boiling with the modified Lennard–Jones potential function, Heat
Mass Transf. 48 (2012) 141–152.