Case Studies in Thermal Engineering 43 (2023) 102645

Contents lists available at ScienceDirect

Case Studies in Thermal Engineering

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

ANFIS based effectiveness and number of transfer units

predictions of MWCNT/water nanofluids flow in a double pipe
U-bend heat exchanger
L. Syam Sundar a, *, B. Deepanraj a, Hiren K. Mewada b
a
Department of Mechanical Engineering, College of Engineering, Prince Mohammad Bin Fahd University, Al-khobar, 31952, Saudi Arabia
b
Department of Electrical Engineering, College of Engineering, Prince Mohammad Bin Fahd University, Al-khobar, 31952, Saudi Arabia

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

Handling Editor: Huihe Qiu In this study, the heat transfer coefficient, Nusselt number, effectiveness and number of transfer
Keywords:
units of water mixed multi-walled carbon nanotubes nanofluids passes through a tube-in-tube
Nanofluid heat exchanger was experimentally investigated. Investigations were performed in the oper­
Effectiveness ating conditions of Reynolds number ranging from 3500 to 12000 and volume concentrations
Number of transfer units ranging from 0% to 0.3%, respectively. The obtained four parameters were predicted using
ANFIS modeling adaptive neuro fuzzy inference system (ANFIS). The Reynolds number and particle volume
Equations loadings are the input data in artificial neural network analysis and heat transfer coefficient,
Nusselt number, effectiveness and number of transfer units is output or target. The Nusselt
number, heat transfer coefficient, effectiveness, and number of transfer units was enhanced to
31.3%, 44.17%, 2.51% and 2.76% at φ = 0.3% and at a Re of 10005, against base fluid.
Implementation of ANFIS with various quantities of neurons in the mid layer provides 1–10− 6
with the correlation coefficient (R2) of 0.9978, and 0.9998 and root mean square error of
0.0018581, and 0.0014159 for heat transfer coefficient and Nusselt number, respectively. The
above developed structure has been successful in predicting 96% of variation in all the
parameters.

1. Introduction
The heat exchangers are commonly used in various industries for exchange of heat from one fluid to other fluid. The single phase
fluids such as water, ethylene glycol and propylene glycol and engine oils are commonly used fluids in the heat exchangers. The double
pipe (tube-in-tube) heat exchangers (DPHE) are commonly used in industries because of less size, more surface area and simple in
design. The effectiveness and heat transfer improvement of these heat exchangers are little bit difficult because of the use of the low
thermal conductivity single phase fluids. With the utilization of high thermal conductivity fluids called as nanofluids, the heat transfer
rate (hnf ) and effectiveness (ε) improvement may possible. Choi [1] invented the nanofluids due to the development in nano­
techanology and then prepared water based nanofluids and obtained higher thermal conductivity values.
By operating these nanofluids in DPHE is providing higher heat transfer coefficients. Several available published works related to
the utilization of various nanofluids in tube-in-tube heat exchangers are presented here. Bahmani et al. [2] preformed experiments for
the estimation of Nusselt number (Nu) and thermal efficiency (ηth ) of double-walled carbon nanotubes/water nanofluids flow in a

* Corresponding author.
E-mail address: sslingala@gmail.com (L.S. Sundar).

https://doi.org/10.1016/j.csite.2022.102645
Received 20 September 2022; Received in revised form 8 December 2022; Accepted 9 December 2022
Available online 26 December 2022
2214-157X/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
L.S. Sundar et al. Case Studies in Thermal Engineering 43 (2023) 102645

tube-in-tube heat exchanger and observed Nu and ηth enhancements of 32.8% and 31% over base fluid, respectively. Moradi et al. [3]
determined the hnf of MWCNT/water nanofluids in a counter flow pipe-in-pipe heat exchanger by using porous media in the Reynolds
number (Re) ranging from 4600 to 7600 and in the particle mass loadings ranging from 0.04% to 0.25%. They have observed hnf
improvement of 35% at mass loadings of 0.04% over base fluid. Aghayari et al. [4] evaluated hnf and Nu of 20 nm Al2O3/water
nanofluid flow in a DPHE and observed hnf and Nu increase of 19% and 24% at φ = 0.3%, respectively. Goodarzi et al. [5] performed
hnf evaluations by using nitrogen-doped graphene/water nanofluids by using 0.025 wt% of Triton X-100 surfactant in the weight
loadings from 0.01% to 0.06% and in the Reynolds number range from 5000 to 15000 in a DPHE. They observed hnf rise of 15.86% at
0.06 wt% against base fluid. Zamzamian et al. [6] observed hnf enhancement range from 2% to 50% for Al2O3/EG and CuO/EG
nanofluids in a DPHE under turbulent flow.
The heat transfer experiments have been conducted by Sarafraz and Hormozi [7] using Ag/50:50% W/EG nanofluids flow in a
DPHE in the φ over 0.1%–1% and they observed hnf rise up to 67% at φ = 1.0% vol. loadings. Duangthongsuk and Wongwises [8]
observed convective hnf rise of 6%–11% at 2.0% for 21 nm of TiO2/water nanofluid flow in a DPHE. Kumar et al. [9] estimated the
Nusselt number of hot Fe3O4 nanofluid flows through a DPHE over the Re from 15000 to 30000 with φ range from 0.005% to 0.06%.
They have found that Nusselt number raise of 14.7% at φ = 0.06% of nanofluids at a Reynolds number of 30000, compared to base
fluid. Sarafraz et al. [10] analyzed thermal conductivity and thermal performance of CNT/water nanofluids flow in a DPHE in the Re
range from 900 to 10500 and they have observed thermal conductivity rise of 56% and thermal performance raise of 44% at 0.3% mass
loadings. Wu et al. [11] determined hnf characteristics of Al2O3/water nanofluids in the weight loadings from 0.78 wt% to 7.04 wt% for
both laminar flow and turbulent flow in a DPHE and they obtained hnf rise of 3.43% at 7.04 wt% of nanofluid.
Shirvan et al. [12] analyzed Nusselt number of Al2O3/water nanofluid loaded in a DPHE by using the response surface methodology
and two phase mixture model in the Re range from 50 to 250 and φ range from 0.01% to 0.05% and they observed Nusselt number rise
of 57.70% at a Re of 150 and at φ = 0.03%. Arani and Amani [13] noticed hnf rise of 75% at a Re of 51000, while TiO2/water nanofluids
loaded in a DPHE. Poongavanam et al. [14] performed Nu experiments for solar glycol based activated carbon nanofluid passes
through DPHE and observed Nu rise of 57.06% at φ = 0.6% of nanofluid over base fluid. Hosseinian et al. [15] noticed hnf rise of 100%
at 0.04 wt% of MWCNT/water nanofluid in a DPHE. By utilizing the Al2O3/water nanofluids in a car radiator Peyghambarzadeh et al.
[16] found hnf rise of 45%.
In an experimental study, Ho and Lin [17] revealed the effect of fluid inlet temperature on the turbulent hnf of Al2O3/water
nanofluids and found over with water data the hnf rise of 44%. According to Reddy and Rao [18], hnf and Δp of DPHE is improved by
10.73% and 8.73%, respectively, at φ = 0.02% of water mixed TiO2 nanofluids. Through counter flow DPHE Zarringhalam et al. [19]
discovered 92.57% augment in hnf of φ = 2% of water/CuO nanofluids. Hemmat Esfe et al. [20] also noticed hnf rise by 25% at φ =
0.4% of MWCNT/water nanofluids in a tube-in-tube heat exchanger. In a tube-in-tube and laminar flow Chun et al. [21]’s in­
vestigations are giving hnf increase when they consider Al2O3/oil nanofluids. They concluded that, the thermal boundary layer almost
approaching to the tube wall and movement of nanoparticles are cause of hnf rise. With the use of water diluted TiO2 nanofluids in
DPHE Ebrahimnia-Bajestana et al. [22] have seen hnf rise of 21%.
Huminic and Huminic [23] analyzed hnf for flow of water based CuO and TiO2 nanofluids in a DPHE numerically. They observed
that, at φ = 2% of CuO nanofluid, the hnf higher of 14% than the base fluid. From the numerical analysis, Demir et al. [24] observed
higher hnf values for the flow of TiO2 and Al2O3 nanofluid in a DPHE. Qi et al. [25] performed the hnf experiments for water based TiO2
nanofluids flow in a shell and tube heat exchanger and noticed hnf increased of 10.8%, 13.4% and 14.8% at mass fractions of 0.1%,
0.3% and 0.5% at a Re of 12000. Bahmani et al. [26] used finite volume approach for the evaluation of hnf of water dispersed Al2O3
nanofluid flow in a DPHE and they observed enhanced hnf values. El-Maghlany [27] studied Cu/water nanofluid in a rotating DPHE. It
was inferred that at a φ = 3%, the effectiveness is enhanced by 16.5%. Pattanayak et al. [28] experimentally noticed an increase of hnf
for flow of water based Al2O3, CuO, TiO2, and ZnO nanofluids in a DPHE. Based on the investigated data of water mixed Al2O3 and Cu
nanofluids,
Jassim and Ahmed [29] have observed 13% and 23% of hnf increase. With the utilization of 0.06% vol. of water dispersed Fe3O4
nanofluids in a DPHE with core-rod wire coil insert of p/d = 1, Sundar et al. [30] noticed hnf increase of 37.90% at a Re of 29000. By
using the water diluted Al2O3 nanofluids in a DPHE and with helical tape inserts, Prasad et al. [31] noticed Nusselt number rise of
32.91% at 0.03% vol. of nanofluid with a helical tape of p/d = 5 at a Re of 30000. Moreover, Prasad et al. [32] is also noticed Nusselt
number rise of 21.73% at 0.3% of Cu/water nanofluid at a Re of 30000 over water. Isfahani and Afrand [33] used pseudo 2D Lattice
Boltzmann method to study the effective viscous drag force in a porous media by using Al2O3/water nanofluid and observed a good
agreement between the numerical and experimental results. Niknejadi et al. [34] observed the best overall hydrothermal performance
results of Fe3O4/water nanofluids at φ = 2%, at a Re of 2000 with a twisted tape pitch of 10 mm. Azimy et al. [35] used ultrasonic
waves for the MWCNT/water nanofluid before they use in the heat exchangers and they observed heat transfer coefficient
enhancement of 300%. Miandoab et al. [36] numerically studied the heat transfer and pressure drop of water-based Graphene-Oxide
nanofluid flow in a horizontal tube with twisted tape inserts. They observed an optimal input parameters of Re of 19,471, twist pitch of
0.0376, and the volume fraction of 0.0383, the Nusselt number and friction factor optimal values of 263.57 and 0.0725, respectively.
Above works experimentally and numerically reveals that the hnf is enhanced with the utilization of nanofluids in a heat exchanger.
The predictions of various parameters involved for the evaluation of nanofluids flow in a DPHE is an important task. The adaptive
neuro fuzzy inference system (ANFIS) modeling approaches or Artificial Neural Network (ANN) provide better validation method for
the measured thermal conductivity, particularly nanotubes applications and experiments [37]. Ahmad et al. [38] used bright one and
two soliton solutions for a nonlocal nonlinear integrable KdV equation using an improved Hirota bilinear method (HBM) and the

2
L.S. Sundar et al. Case Studies in Thermal Engineering 43 (2023) 102645

Visualization was seen using the MATLAB. Khaliq et al. [39] used novel exact travelling wave solutions for the (2 + 1)-dimensional
Boussinesq equation simulated using the Maple-18.
Few reports related to the ANFIS approach for nanofluids in tube-in-tube heat exchangers are only available and some of them are
discussed here. Maddah and Ghasemi [40] have obtained hnf increase of 30% by using 0.03 wt% of water based iron oxide nanofluids
flow a DPHE and they also proposed ANN approach for the evaluation of error and found that the ANN structure predicts 72%–97%
accuracy of hnf data. Mmohammadiun et al. [41] used ant colony optimization-back propagation (ACOR-BP) artificial neural networks
(ANN) model for the optimal design of CuO nanofluids flow in a double pipe heat exchanger with twisted tapes and they observed
exergy is lower than the base fluid. The ANN approach was used by Hemmat Esfe [42] for Ag/water nanofluids in a DPHE and they
noticed Nusselt number regression coefficient and pressure drop of 99.76% and 99.54%, respectively. Nasirzadehroshenin et al. [43]
used experimental exergy efficiency data of water based CNTs-10 wt% CuO nanofluid in a tube-in-tube heat exchanger for predicting
the modifying response surface methodology-central composite design (RSM-CCD) model and obtained optimal exergy efficiency of
the system is 98.4%.
Based on the aforementioned research, there is a substantial body of literature on the study of heat transmission in nanofluids, and
the results that have been presented therein have generated some debate. While some studies claim that the inclusion of nanoparticles
can significantly increase the pressure drop and pumping power, which results in the tradeoff scenario, others contend that the
opposite is true. As a result, a fair evaluation of the thermal performance of nanofluid is necessary in light of the pressure drop and
improved heat transfer.
The effectiveness and number of transfer units of water mixed MWCNT nanofluids flow in a DPHE with ANFIS predictions are not
available in the literature. The use of nanofluids in a pipe-in-pipe u-bend exchange comes under passive hnf augment method. Hence
the paper aims to determine effectiveness and number of transfer units experimentally at dissimilar Reynolds number and particle
loadings and the obtained values are predicted with ANFIS model to estimate the correlation coefficients. In addition to that the ANFIS
predictions are also extended to predict the heat transfer coefficient, and Nusselt number. Among all the nanofluids, the MWCNT
nanoparticles are offering higher thermal conductivity and Young’s modulus and also the cost is very less. Based on the ANFIS model a
polynomial regression equations were developed for all the estimated parameters with correction coefficient (R2) is equal to 0.9999.
Additionally, prolonged steadiness of the nanofluids is promise and outcomes are more trustworthy when utilizing the water diluted
COOH-CNT balanced by Nanosperse AQ.

2. Experimental study
2.1. Sample preparation
The MWCNT nanoparticles were used in this study were procured from the Ultrananotech Private Limited, India with a purity of
95%, outer diameter of 8–15 nm, and length of 30–50 μm. Prior to preparing the water based MWCNT nanofluids, the procured
MWCNT was treated with strong acid solution. The carboxyl groups on the exterior of MWCNT results the steadiness of the particles in
the base fluid. The MWCNT was disseminated in strong acid solution of 3:1 mol ratio of HNO3 and HCl and stirred for 72 h at a
temperature of 40 ◦ C using a magnetic stirrer. Later, the MWCNT was rinsed numerous times with purified water and dried it for 24 h at
a temperature of 80 ◦ C. This process involves the development of carboxyl (–COOH) bond over the surface of MWCNT.
The morphology was done on MWCNT with X-ray diffraction (Siemens D-500, 45 kV) instrument and Scanning Electronic Mi­
croscopy (Hitachi SU-70 SEM) instrument and the results were given in Fig. 1(a) and 1(b). The XRD spectrum was perfectly matches
with [44] and no different characteristics peaks were noted related to used acids for acid treatment of the MWCNT nanoparticles. The
B(2θ) positions of the planes (002), (100) and (004) was noted as 26◦ , 45◦ and 55.4◦ , respectively.
In order to prepare the SEM analysis (Fig. 1(b)), a small quality of acid treated MWCNT was diluted in pure water and then
sonicated with ultrasonicator about 30 min, later the MWCNT/water solution was transferred to the single crystal silicon substrate and
then dried it at 80 ◦ C. Later the silicon substrate was fixed to the SEM holder and used in the instrument. The tube like structure was
identified through the SEM analysis.
Total 10 L of MWCNT/water nanofluids were prepared by adding 10.54, 21.04, and 63.29 g of dry MWCNT nanoparticles into the

Fig. 1. MWCNT nanoparticles: (a) XRD patterns, and (b) SEM image.

3
L.S. Sundar et al. Case Studies in Thermal Engineering 43 (2023) 102645

water to match particle volume loadings of 0.05%, 0.1% and 0.3%, respectively. The Nanosperse AQ procured from NanoLab, Inc.,
USA was used as a stabilizing agent. For each 1 g of MWCNT, one drop of Nanosperse AQ surfactant was used. Quantity of 10.54 g of
MWCNT nanoparticles were mixed in 10 L of distilled water by adding the required quantity of surfactant in a 20 L capacity of tank and
then agitated slowly with a mechanical agitator at a rpm of 200 for 4 h to obtain 0.05% vol. of nanofluid. The same technique was
adopted for the development of other volumes of nanofluids. The prepared MWCNT/water nanofluids were stable for five months
without particle agglomeration, which is enough for investigations. Saleh and Sundar [45] estimated the thermophsical properties of
MWCNT/nanofluids and those are used in this study and they are fitted with a multi-linear regression method.

Fig. 2. (a) Schematic diagram of an experimental setup, and (b) test section details.

4
L.S. Sundar et al. Case Studies in Thermal Engineering 43 (2023) 102645

(a) The thermal conductivity data of [45] is fitted into the below Eq. (1).
knf = 1.13104 − 0.01098T + 0.22487φ (1)

(b) The viscosity data of [45] is fitted into the below Eq. (2).
μnf = 0.57259 + 0.0016T + 0.2568φ (2)

(c) The density data of [45] is fitted into the below Eq. (3).
ρnf = 1012.68 − 0.44961T + 29.3469φ (3)

(d) The specific heat is given by Ref. [45] is fitted into the below Eq. (4).
Cp,nf = 4170.922 + 0.18552T − 19.4819φ (4)

The above proposed equations were valid in the range of 20o C < T < 60o C; and 0 < φ < 0.3%.

2.2. Experimental apparatus and method

The line diagram of the setup is provided in Fig. 2 (a). It is a tube-in-tube heat exchanger type, and the annulus tube consists of
galvanized iron material with Do = 0.056m and Di = 0.062m. The inside tube consists of copper material with tube, do = 0.019m and
di = 0.022m. The inside tube length 5 m and the bend is provided at a tube length 2.2 m with a radius of 0.16 m. The other instruments
like cooler, tanks, pumps, data logger, personal computer, flow adjustment valve and manometer were used. The test tube specifi­
cations were provided in Fig. 2(b).
The hot-water passes circulate in the annulus section, however, the MWCNT nanofluids passes through the inner tube. The counter
flow direction was maintained between the hot water and nanofluids. The capacity of the hot water tank is 20 L, in which the 15 L of
water was filled and four 2000 W, and 230 V heaters copper heating material painted with nickel (Industrial heat products, India) are
immersed in it to heat the water. After reaching the hot water at a temperature of 80 ◦ C, they are allowed into the annulus tube at a
fixed volume rate of 3 lit/min with help of pump (Samudra Pumps (India) Pvt Ltd). The capacity of the nanofluids tank is 20 L, in which
10 L of nanofluids was used. The nanofluids are allowed inside the inner tube with help of pump at volume flow rates varies from 3 to 8
lit/min. The raise in outlet temperature (J-type thermocouples, Sontex Electrical Heaters, India, accuracy of ±0.01 ◦ C) of the nano­
fluids due to the heat transfer between hot and nanofluids are bring back to an atmospheric temperature (25 ◦ C) with help of Julabo
temperature controlled cooler bath (JULABO GmbH, Germany, accuracy of ±0.01 ◦ C). The hot water and nanofluids flow rates are
monitored with flow meters. The pressure across the inner tube was measured by using the u-tube manometer. The hot water and
nanofluids inlet and outlet temperatures at the double pipe were determined by eight J-type thermocouples, those are connected to
data logger and then it is connected to computer. Once, the system reaches to steady state (nearly equal to 2 h), the hot water and
nanofluids temperatures were recorded and used for further analysis. The uncertainties of various instruments and parameter used in
the experimental analysis are mentioned in appendix in supplementary file.

3. Data deduction
3.1. Nusselt number
Eqns. (5) and (6) predicts the heat drop and heat gain of both the fluids. Average between them is used for calculations [46–48].
( )
Qh = ṁh cp,h Th,i − Th,o (5)
( )
Qnf = ṁnf cp,nf Tnf ,o − Tnf ,i (6)

Qh + Qnf
Qavg = (7)
2
The nanofluids heat transfer coefficient (hnf ) is calculated from the below equation [47].
Qavg
h= (8)
A(ΔT)LMTD

[Th,i − Tc,o ]− [Th,o − Tc,i ]

Where, LMTD = [
Th,i − Tc,o
] ; A = π di L
ln Th,o − Tc,i

The Nusselt number is calculated from the below equation:

hnf di,nf
Nunf = (9)
knf

Eq. (10) is used to verify the Nusselt number (Dittus and Boelter [46] equation).

5
L.S. Sundar et al. Case Studies in Thermal Engineering 43 (2023) 102645

Nuh = 0.023Re0.8 0.4

h Pr h (10)

Nuh kh
hh = × 0.023 × (Reh )0.8 × (Prh )0.4 (11)
Di,h

Where, the hydraulic diameter, Dh = Di − di (12)

3.2. Effectiveness – Number of transfer units

Effectiveness-number of transfer units is determined from the below equation [48].
q
ε= (13)
qmax

Where, q = Ch (Th,i − Th,o ) = Ch (Tc,o − Tc,i )

( )
qmax = Cmin Th,i − Tc,i

Case-1: If Ch > Cc then Case-2: If Ch > Cc then

( ) ( )
Ch = mcp h = Cmax Ch = mcp h = Cmin
( ) ( )
Cc = mcp c = Cmin Cc = mcp c = Cmax

UA
NTU = (14)
Cmin

4. ANFIS modeling
ANFIS is among the most modern computing strategies advised due to its tools for dealing with complicated stochastic datasets and
non-linear modeling. The ANFIS fuzzy model is utilized to build rule systems by entering the data from the prediction model to produce
more effective membership functions. ANFIS is an adaptive network that performs as fuzzy modeling, so ANFIS approaches are used to
modify the membership function variables. The primary objective of this study is an integrated approach of fuzzy inference systems

Fig. 3. The ANFIS flow chart for prediction.

6
L.S. Sundar et al. Case Studies in Thermal Engineering 43 (2023) 102645

and neural networks, both prominent artificial intelligence approaches. The output of the ANFIS is calculated in the forward pass, and
an essential back propagation algorithm approach is employed to change the proposal parameters based on the prediction error. The
proposed ANFIS architecture for the predictions is mentioned in Fig. 3.
ANFIS is a five-layer structure, including the input and output layers. In layer-1, each node indicates how well an input “u” satisfies
the linguistic idea calculated using membership function as follows [40].
( )
u − σi
Ο1i = exp − (15)
2μi

Where, σi and μi are adaptive parameters of the membership function, also called non-linear parameters. The second layer produces the
firing strength using non-adaptive T-norm using declared fuzzy-set rules. It is modeled as [40]:
( )
Ο2i Ο2i − Wi (16)

The normalization of all firing strengths is performed by nonadaptive layer 3 and results in normalized firing strengths. It is
expressed as [40]:

Fig. 4. (a) Validation of experimental Nusselt number, (b) Heat transfer coefficient of nanofluids, (c) Nusselt number of nanofluids, (d) effectiveness of nanofluids, and
(e) number of transfer units.

7
L.S. Sundar et al. Case Studies in Thermal Engineering 43 (2023) 102645

Wi
Ο3i = ∑n (17)
i=1 Wi

where n is the number of nodes in the third layer. The node in layer four is associated with a malleable processing unit that carries out a
linear function denoted by Ref. [40]:

Ο4i = Ο3i (αi u1 + βi u2 + γi ) (18)

Where, αi , βi and γi determine the response of linear function and are referred to as the system’s linear parameters. The last layer-5
computes the final prediction as [41]:
∑ 4
Ο
P = ∑n i (19)
i=1 Wi

It is more robust due to its hybrid nature of fuzziness and neural network. This study uses a non-linear technique with two inputs:
Reynolds number (Re) and particle volume concentrations (φ). The membership function is identified using fuzzy C-means clustering.
Further Sugeno rules were used in determining the membership function. For different concentration and Raynold numbers, four
properties, i.e., heat transfer coefficient, Nusselt number, ε and NTU are used in ANFIS model development [42]:
n ( )2
1 ∑
MSE = pexp (i) − pANN(i) (20)
N i=1

√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
√ ∑ n ( )2
√ pexp (i) − pANN(i)
√
√
R = √1 − i=1 ∑ n ( )2 (21)
√ 1
N
pexp (i)
i=1

Where, pexp , pANN , and N stands for the experimental value, the value projected by the ANN, and the total number of data points.

5. Results and discussion

5.1. Nusselt number
The nanofluids volume flow rates were varies from 3 to 8 lit/min, whereas, the hot water volume flow (v̇) rate was stable to 3 lit/
min. The Reynolds number (Re = 4ṁ /πdi μ) of water and MWCNT/water nanofluids was calculated and those are: Re for water varies
from 4177 to 11139; the Re for 0.05% vol. varies from 4093 to 10915; the Re for 0.1% vol. varies from 4004 to 10686; and the Re for
0.3% vol. varies from 3751 to 10005, respectively.
Initially base fluid (cold fluid) was considered in inner tube of the heat exchanger and hot water is as usual flows in annulus tube.
The experimentally obtained data is validated with literature data in Fig. 4(a). As seen it from the data, the measured Nusselt number is
deviated ±2.8% with literature data [49,50].
( f)
(Re − 1000)Pr
(i) Gnielinski [49] : Nu = 2
( )0.5 ( 2/3 ) (22)
1.07 + 12.7 2f Pr − 1

2
f = (1.58 ln(Re) − 3.82)−

0.5 < Pr < 2000; 2300 < Re < 106

(ii) Notter and Sleicher [50] : Nu = 5 + 0.016Rea Prb (23)

0.24 0.6 Pr
a = 0.88 − ; b = 0.33 + 0.5e−
(4 + Pr)

104 < Re < 106 ; 0 < Pr < 104

The MWCNT/water nanofluids were loaded one after one in inner tube of heat exchanger. The same boundary conditions were
fixed for both cold nanofluids and hot-water. The hot water v̇ and temperature was maintained as 3 lit/min and 80 ◦ C, however, the v̇ of
nanofluids was varies from 3 lit/min to 8 lit/min at a fixed inlet temperature of 25 ◦ C. Using Eqns. (1) and (2), the hot water heat loss
and the nanofluids heat gain was calculated and the average of these two was used in Eq. (7) to evaluate the nanofluids hnf . Fig. 4(b)
shows the heat transfer coefficient of nanofluids at dissimilar Re and φ data. The hnf is raised to an augment of Re and φ data. Moreover,
at φ = 0.05%, the hnf is raised from 7.81% to 22% over the Re range from 4093 to 10915; however, at φ = 0.1%, the hnf is increased
16.51%–30.11% over the Re range from 4004 to 10686; similarly, at φ = 0.3%, the hnf is higher from 25.71% to 44.21% over the Re
range from 3751 to 10005, according to the water data.
The figure also shows the comparison of present MWCNT/water nanofluid with Sarafraz and Hormozi [7] of 50:50% W/EG-silver

8
L.S. Sundar et al. Case Studies in Thermal Engineering 43 (2023) 102645

nanofluids at 1.0% vol. and Duangthongsuk and Wongwises [8] of TiO2/water nanofluid at 0.2% vol., respectively. One can observe,
the present data showing higher hnf values than the literature values. The hnf data 0.3% vol. is predicting 27.54% higher than the
Sarafraz and Hormozi [7] at 1.0% vol. and 52.92% higher than of Duangthongsuk and Wongwises [8] at 0.2% vol. The same
tube-in-tube heat exchanger for the present case and [7,8] are used for the hnf study, maybe tube geometry is different. Particle thermal
conductivity, operating parameters, and dimensions of the test section matters on the hnf augmentation.
The founded Nusselt numbers of MWCNT nanofluids were mentioned in Fig. 4(c) over different Reynolds numbers. Over the base
fluid data, the Re ranging from 4093 to 10915, the Nu is increased from 7.84% to 18.65% at 0.05%; the Re ranging from 4004 to 10686,
the Nu is augmented from 10.14% to 23.61% at 0.1%; the Re ranging from 3751 to 10005, the Nu is raised from 14.52% to 31.23% at
0.3%, respectively.
The figure also shows the validation of current Nu of MWCNT/water along Sarafraz et al. [10] data of COOH-CNT/water nanofluid
at 0.3% wt. Clearly indicating the figure the present data is higher than the [10] data. Even though the nanoparticles are same, but the
present study MWCNT at 0.3% vol. predicting 36.73% higher Nusselt number over Sarafraz et al. [10] data at 0.3% wt. Particle volume
loadings were considered for the present analysis, whereas [10] used particle weight loadings. While for the preparation of volume
concentrations, the large quantity of CNT nanoparticles are required in the base fluid compared to weight concentrations at the same
amount of base fluid. Once the CNT is more in base fluid it’s knf is more, hence it leads to higher hnf .
Thermal conductivity is the likely cause of this rise. There are a several mechanisms and those explains about the rise of thermal
conductivity of a nanofluid, and those are development of liquid layer on the outer surface of the particles, Brownian motion, particle
classification, the transmission of phonon projectiles through the particles, and an increase in thermal conductivity of fluids with an
increase in the amount of nanoparticles in the pipe wall. In the thermal boundary layer close to the tube wall, a rise in thermal
conductivity may result in an increase in the heat transfer coefficient. One of the elements affecting the temperature is the nanofluid’s
increased thermal conductivity, which also affects the heat transfer coefficient and Nusselt number.
According to experimental findings, as the temperature rises, the impacts of nanoparticles on thermal conductivity get stronger.
The main mechanism for the nanofluid’s thermal conductivity is thought to be the nanoparticles’ random mobility. The fluid tem­
perature affects this pseudo-Brownian motion. Therefore, at larger temperatures, the rise in thermal conductivity is greater for smaller
particles than for bigger particles. Because Brownian motion is less significant at low temperatures, there is less of a difference in the
increase in thermal conductivity between smaller and larger particles.
Thermal boundary layer disturbance may occur due to the addition of nanoparticles in water, it enhanced the thermal conductivity,
and it causes the augmentation in heat transfer coefficient. There as it is understood that, the heat transfer coefficient is increased due
to the augment in thermal conductivity and Reynolds number. As it is understood that, the Cp is the amount of heat needed for one
degree raise in temperature of 1 unit mass of substance. So, based on Cp definition lower Cp provides a higher ΔT for same heat flow.
Sarafraz and Hormozi [10] developed the Nusselt number correlation based on the experimental data of biological silver/WEG50
nanofluids and the equation is given below.
( √̅̅̅ )
Nu = 0.00546 1 + 8.14 φ Re0.9324 .Pr0.38 .Pe0.0038 (24)

The present experimental data is fitted into single multi-linear regression equation by considering the Re and φ for the estimation of
Nusselt number, which is given below. The equation predicts the values with a standard deviation of 6.5%:
Nu = − 0.513203 + 0.00831788Re + 61.323136φ (25)

ε = 0.44334 + 0.00000284516Re + 0.0496885φ (26)

NTU = 0.429207 + 0.0000116743Re + 0.0687022φ (27)

3751 ≤ Re ≤ 11139.62; 0 ≤ φ ≤ 0.3%

5.2. Effectiveness –number of transfer units

The calculated effectiveness of MWCNT nanofluids is provided in Fig. 4(d) along with water and at various Reynolds number. The
effectiveness is increased with increase of particle volume concentrations and Reynolds number. Over the base fluid data, the Re
ranging from 4093 to 10915, the ε is raised from 0.291% to 0.878% at 0.05%; the Re ranging from 4004 to 10686, the ε is augmented
from 1.42% to 1.89% at 0.1%; the Re ranging from 3751 to 10005, the ε is raised from 2.19% to 2.49% at 0.3%, respectively.
The effectiveness of the present MWCNT/water nanofluids is compared with Prasad et al. [31] of Cu/water nanofluid in Fig. 4(d). It
is observed that, the present study MWCNT/water nanofluid offers 77.7% higher effectiveness than Cu/water nanofluids at 0.3% vol.
of nanofluid and at Reynolds number of 10005.09. Hence, under the same particle loadings (0.3%) and for same base fluid (water) the
utilization of MWCNT/water nanofluid offering more effectiveness in double pipe heat exchanger than the Cu/water nanofluid.
Fig. 4(e) indicates the number of transfer units of MWCNT nanofluids. As the increase of Reynolds number and particle volume
loadings, the number of transfer units is increased. Over the base fluid data, the Re ranging from 4093 to 10915, the NTU is raised from
0.591% to 0.692% at 0.05%; the Re ranging from 4004 to 10686, the NTU is augmented from 2.02% to 2.031% at 0.1%; the Re ranging
from 3751 to 10005, the NTU is raised from 2.97% to 2.76% at 0.3%, respectively.
The number of transfer units of the present MWCNT/water nanofluids is compared with Prasad et al. [31] of Cu/water nanofluid in
Fig. 4(e). It is observed that, the present study MWCNT/water nanofluid offers 15.15% higher effectiveness than Cu/water nanofluids

9
L.S. Sundar et al. Case Studies in Thermal Engineering 43 (2023) 102645

at 0.3% vol. of nanofluid and at Reynolds number of 10005.09. Hence, under the same particle loadings (0.3%) and for same base fluid
(water) the utilization of MWCNT/water nanofluid offering more number of transfer units in double pipe heat exchanger than the
Cu/water nanofluid. The similar results of an enhanced efficiency were observed by Kaya et al. [51] for conducting the experiments for
the evaluation of energetic and exergetic characteristics of CuO/methanol nanofluid in the concentrated air collector. They observed
average efficiency for methanol and CuO/methanol nanofluid is 65% and 64%, respectively. Yıldız et al. [52] have explained that, the
nanofluids thermophysical properties are also an important parameters for obtaining the higher performance of the thermal and
refrigeration systems. The uncertainty analysis is performed by using Beckwith et al. [56] and its values are provided in the sup­
plementary infromation.

Fig. 5. ANFIS based heat transfer coefficient predictions for all the data, the training data, and the testing data.

10
L.S. Sundar et al. Case Studies in Thermal Engineering 43 (2023) 102645

5.3. ANFIS modeling

The best ANFIS model can be evaluated with two quantitative parameters, i.e., root mean square error (RMSE) and R2 value. The
best ANFIS model results in a smaller RMSE and maximum R2 value. In the ANFIS system, this study uses 80% of data for training and
20% for testing from the obtained experiment datasets. An ANFIS model is trained for each parameter and the corresponding results for
all the parameters were expressed in Figs. 5–8.
Fig. 5 indicates all the predicted values, their deviation from the actual values, MSE and RMSE analysis for heat transfer coefficient.
The heat transfer coefficient value is considerable, ranging from 4000 to 13000. The heat transfer coefficient RMSE value is 25.6261,
which is equivalent to 1.98% an average error concerning its maximum value. For all the data points, the MSE is 0.00034524, RMSE is
0.018581, the mean error is 0.00029223, standard deviation is 0.018978 and the correlation coefficient is 0.99786, respectively.
Similarly, for the test data, the MSE is 0.00054871, RMSE is 0.02342, and the mean error is 0.0018798, standard deviation is
0.026705, and the correlation coefficient is 0.99728, respectively. Moreover, for the test data, the MSE is 0.0002917, RMSE is

Fig. 6. ANFIS based Nusselt number predictions for all the data, the training data, and the testing data.

11
L.S. Sundar et al. Case Studies in Thermal Engineering 43 (2023) 102645

Fig. 7. ANFIS based effectiveness predictions for all the data, the training data, and the testing data.

0.017079, and the mean error is − 0.00012556, standard deviation is 0.017547, and the correlation coefficient is 0.9982, respectively.
Fig. 6 is plotted to analyze the predicted Nusselt number values. The Nusselt number range obtained from the experimental study is
varies over 20 to 120. For all the data points, the MSE is 3.1503, RMSE is 1.7749, the mean error is 0.21756, standard deviation is
1.7994, and the correlation coefficient is 0.99617, respectively. Similarly, for the test data, the MSE is 6.4628, RMSE is 2.5422, and the
mean error is 0.96807, standard deviation is 2.6281, and the correlation coefficient is 0.99338, respectively. Moreover, for the test
data, the MSE is 2.2785, RMSE is 1.5095, and the mean error is 0.020054, standard deviation is 1.5507, and the correlation coefficient
is 0.99726, respectively.
Fig. 7 is plotted to analyze the predicted effectiveness values. The effectiveness range obtained from the experimental study is
varies over 0.46 to 0.58. For all the data points, the MSE is 3.501e− 06, RMSE is 0.0017467, the mean error is 0.0004827, standard
deviation is 0.0017147, and the correlation coefficient is 0.99824, respectively. Similarly, for the test data, the MSE is 7.5873e− 06,
RMSE is 0.0027545, the mean error is 0.0023658, standard deviation is 0.0015774, and the correlation coefficient is 0.99669,
respectively. Moreover, for the test data, the MSE is 1.8973e− 06, RMSE is 0.0013628, the mean error is 1.2584e− 05, standard deviation
is 0.0014001, and the correlation coefficient is 0.99893, respectively.
Fig. 8 is plotted to analyze the predicted number of transfer unit values. The number of transfer units obtained from the experi­
mental study is varies over 0.45 to 0.49. For all the data points, the MSE is 2.0549e− 06, RMSE is 0.0014335, the mean error is

12
L.S. Sundar et al. Case Studies in Thermal Engineering 43 (2023) 102645

Fig. 8. ANFIS based number for transfer units predictions for all the data, the training data, and the testing data.

− 0.00036572, standard deviation is 0.0014159, and the correlation coefficient is 0.98687, respectively. Similarly, for the test data, the
MSE is 4.4404e− 06, RMSE is 0.0021072, the mean error is 0.0016794, standard deviation is 0.0014231, and the correlation coefficient
is 0.99524, respectively. Moreover, for the test data, the MSE is 1.4272e− 06, RMSE is 0.0011947, the mean error is − 2.0021e− 05,
standard deviation is 0.0012272, and the correlation coefficient is 0.98494, respectively.
Karaağaç et al. [53] used Al2O3-paraffin wax nano phase change material in a concentrated photovoltaic-thermal solar dryer and
studied first and second law of thermodynamics experimentally and as well as machine learning algorithms and they observed the
drying rate of 0.436 g matter/g dry matter. min and also they observed the used machine learning algorithms are predicting high
accuracy with the experimental data. Ağbulut et al. [54] used deep learning, artificial neural network, Kernel nearest neighbor, and
support vector machine for the data of CI engine fuelled with Al2O3, CuO, and TiO2 nanoparticles to predict the emission and per­
formance and they observed the CO reduction of 8.3%, 11.6% and 15.5% for TiO2, Al2O3, and CuO compared to neat diesel fuel. Based
on the Taguchi’s grey relational and response surface methodology, Kaladgi et al. [55] designed and optimized the polyethylene
glycol/ZnO nanofluids properties.
The polynomial regression equation gives accurate results when compared to regular equations because the polynomial equation
involves nth order of the both the input parameters of Reynolds number and particle volume loadings. Hence, the proposed regression
model can be expressed using a polynomial equation. A polynomial equation for each property is given in Eqns. (28)-(31). These

13
L.S. Sundar et al. Case Studies in Thermal Engineering 43 (2023) 102645

coefficients are calculated with a 95% coefficient bound.

h = 3999 − 1.05e− 4 Re − 0.7985φ + 1.98e4 Re2 + 5.177Reφ + 0.000235φ2 − 8.534Re2 φ − 7.56e− 5 Reφ2 − 1.12e− 8 φ3 (28)

Nu = 38.74 − 100.6Re − 0.00795φ + 181.7Re2 + 0.04187Reφ + 2.21e− 06

φ2 − 0.06897Re2 φ − 7.18e− 07
Reφ2 − 1.05e− 10
φ3 (29)

ε = 0.2758 + 1.492Re + 6.82e− 5 φ − 0.1884Re2 + 3.76e− 7 Reφ + 6.30e− 9 φ2 − 1.27e− 5 Re2 φ + 7.40e− 11 Reφ2 + 2.12e− 13 φ3 (30)

NTU = 0.4186 + 0.1282Re + 9.92e− 6 φ − 0.2531Re2 + 2.16e− 6 Reφ + 6.00e− 10

φ2 − 5.12e− 7 Re2 φ + 2.71e− 11
Reφ2 + 1.12e− 14
φ3 (31)

5.4. Future works

The heat transfer, effectiveness and number of transfer units of the use of MWCNT/water nanofluids in a tube-in-tube heat
exchanger was analyzed experimentally and found that, an enhanced heat transfer coefficient, effectiveness, and number of transfer
units. The obtained data is also predicted with ANFIS model and found that the experimental values are very good agreement with the
predicted values.
In future works, the analyses like effectiveness and number of transfer units may carry forward for the use of different type of mono
and hybrid nanofluids flow in tube-in-tube heat exchangers and predict the obtained data with various machine learning algorithms.
The effectiveness and number of transfer units analyses may also further carry forwards for the flow of mono and hybrid nanofluids in a
tube-in-tube heat exchangers and with different kind of turbulent promoters such as twisted tape inserts, longitudinal strip inserts, wire
coil inserts, and helical coil inserts. The obtained experimental data is also predicted with various ANN and machine learning
algorithms.

6. Conclusion
In the present study, counter flow tube-in-tube heat exchanger circulating with MWCNT/water nanofluids effectiveness and
number of transfer units were investigated experimentally. Based on the experimental investigations it is concluding that, with the
utilization of nanofluids, the effectiveness and number of transfer units are significantly improved. In addition to that, the heat transfer
coefficient and Nusselt number is also improved. Addition of nanoparticles volume loadings of 0.05%, 0.1% and 0.3%, the heat transfer
coefficients were improved by 22%, 30.11%, and 44.21%, whereas the Nusselt number is augmented by 18.65%, 23.61%, and 31.23%,
respectively, at a Reynolds number of 10005. Simultaneously, the effectiveness and number of transfer units at 0.3% vol. loadings is
enhanced by 2.49% and 2.75% at Re equal to 10005 over the base fluid.
The obtained parameters are predicted with ANFIS model and it is noted that all the measured parameter correlation coefficient is
R2 < 0.996 with the room mean square error of ±2.5% deviation. From the ANFIS model, a polynomial regression equation was
developed for all the measured parameter. Hence the designed ANFIS architecture model predicts very accurate values.

CRediT authorship contribution statement

L. Syam Sundar: Conceptualization, Analysis, Investigation, Writing – original draft. B. Deepanraj: Methodology, Data curation,
Analysis, Validation. Hiren K. Mewada: Supervision, Formal analysis, Visualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to
influence the work reported in this paper.

Data availability

No data was used for the research described in the article.

Nomenclature

A Area, m2
Cp Specific heat, J/kg K
di Inner diameter of the tube, m
do Outer diameter of the tube, m
Di Annulus section inner diameter, m
Do Annulus section outer diameter, m
h Heat transfer coefficient, W/m2 K
k Thermal conductivity, W/m K
L Length of the tube, m
ṁ Mass flow rate, kg/sec
Nu Nusselt number, h D/k

14
L.S. Sundar et al. Case Studies in Thermal Engineering 43 (2023) 102645

Pr Prandtl number, (μ Cp )⁄ k
Qh Heat supplied, W
Qa Heat absorbed, W
Re Reynolds number, 4 ṁ/πDμ
T Temperature, o C
U Overall heat transfer coefficient, W/m k
v̇ Volume flow rate, lit/min

Greek symbols
ΔT Logarithmic mean temperature difference
φ Volume concentration of nanoparticles, %
μ Dynamic viscosity, kg/m2 sec
ρ Density, kg/m3

Subscripts
c Cold fluid
Exp Experimental
h Hot fluid
i Inlet
o Outlet
Reg Regression

Appendix A. Supplementary data

Supplementary data related to this article can be found at https://doi.org/10.1016/j.csite.2022.102645.

References
[1] S.U.S. Choi, Enhancing thermal conductivity of fluids with nanoparticles, in: D.A. Siginer, H.P. Wang (Eds.), Developments and Applications of Non-newtonian
Flows, ASME, New York, 1995, pp. 99–105. FED 231/MD 66.
[2] M.H. Bahmani, G. Sheikhzadeh, M. Zarringhalam, O.A. Akbari, A.A.A.A. Alrashed, G.A.S. Shabani, M. Goodarzi, Investigation of turbulent heat transfer and
nanofluid flow in a double pipe heat exchanger, Adv. Powder Technol. 29 (2018) 273–282.
[3] A. Moradi, D. Toghraie, A.H.M. Isfahani, A. Hosseinian, An experimental study on MWCNT–water nanofluids flow and heat transfer in double-pipe heat
exchanger using porous media, J. Therm. Anal. Calorim. 137 (2019) 1797–1807.
[4] R. Aghayari, H. Maddah, M. Zarei, M. Dehghani, S.G.K. Mahalle, Heat Transfer of Nanofluid in a Double Pipe Heat Exchanger, International Scholarly Research,
2014, https://doi.org/10.1155/2014/736424. Article ID 736424, 7 pages.
[5] M. Goodarzi, A.Sh Kherbeet, M. Afrand, E. Sadeghinezhad, M. Mehrali, P. Zahedi, S. Wongwises, M. Dahari, Investigation of heat transfer performance and
friction factor of a counter-flow double-pipe heat exchanger using nitrogen-doped, graphene-based nanofluids, Int. Commun. Heat Mass Tran. 76 (2016) 16–23.
[6] A. Zamzamian, S.N. Oskouie, A. Doosthoseini, A. Joneidi, M. Pazouki, Experimental investigation of forced convective heat transfer coefficient in nanofluids of
Al2O3/EG and CuO/EG in a double pipe and plate heat exchangers under turbulent flow, Exp. Therm. Fluid Sci. 35 (2011) 495–502.
[7] M.M. Sarafraz, F. Hormozi, Intensification of forced convection heat transfer using biological nanofluid in a double-pipe heat exchanger, Exp. Therm. Fluid Sci.
66 (2015) 279–289.
[8] W. Duangthongsuk, S. Wongwises, Heat transfer enhancement and pressure drop characteristics of TiO2–water nanofluid in a double-tube counter flow heat
exchanger, Int. J. Heat Mass Tran. 52 (2009) 2059–2067.
[9] N.T.R. Kumar, P. Bhramara, B.M. Addis, L.S. Sundar, M.K. Singh, A.C.M. Sousa, Heat transfer, friction factor and effectiveness analysis of Fe3O4/water nanofluid
flow in a double pipe heat exchanger with return bend, Int. Commun. Heat Mass Tran. 81 (2017) 155–163.
[10] M.M. Sarafraz, F. Hormozi, V. Nikkhah, Thermal performance of a counter-current double pipe heat exchanger working with COOH-CNT/water nanofluids, Exp.
Therm. Fluid Sci. 78 (2016) 41–49.
[11] Z. Wu, L. Wang, B. Sundén, Pressure drop and convective heat transfer of water and nanofluids in a double-pipe helical heat exchanger, Appl. Therm. Eng. 60
(2013) 266–274.
[12] K.M. Shirvan, M. Mamourian, S. Mirzakhanlari, R. Ellahi, Numerical investigation of heat exchanger effectiveness in a double pipe heat exchanger filled with
nanofluid: a sensitivity analysis by response surface methodology, Powder Technol. 15 (2017) 99–111.
[13] A.A.A. Arani, J. Amani, Experimental study on the effect of TiO2–water nanofluid on heat transfer and pressure drop, Exp. Therm. Fluid Sci. 42 (2012) 107–115.
[14] G.K. Poongavanam, B. Kumar, S. Duraisamy, K. Panchabikesan, V. Ramalingam, Heat transfer and pressure drop performance of solar glycol/activated carbon
based nanofluids in shot peened double pipe heat exchanger, Renew. Energy 140 (2019) 580–591.
[15] A. Hosseinian, A.H.M. Isfahani, E. Shirani, Experimental investigation of surface vibration effects on increasing the stability and heat transfer coefficient of
MWCNTs-water nanofluid in a flexible double pipe heat exchanger, Exp. Therm. Fluid Sci. 90 (2018) 275–285.
[16] S.M. Peyghambarzadeh, S.H. Hashemabadi, M. Seifi Jamnani, S.M. Hoseini, Improving the cooling performance of automobile radiator with Al2O3/water
nanofluid, Appl. Therm. Eng. 31 (2011) 1833–1838.
[17] C.J. Ho, Y.J. Lin, Turbulent forced convection effectiveness of alumina–water nanofluid in a circular tube with elevated inlet fluid temperatures: an
experimental study, Int. Commun. Heat Mass Tran. 57 (2014) 247–253.
[18] M.C.S. Reddy, V.V. Rao, Experimental investigation of heat transfer coefficient and friction factor of ethylene glycol water based TiO2 nanofluid in double pipe
heat exchanger with and without helical coil inserts, Int. Commun. Heat Mass Tran. 50 (2014) 68–76.
[19] 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.
[20] M. Hemmat Esfe, S. Saedodin, O. Mahian, S. Wongwises, Heat transfer characteristics and pressure drop of COOH-functionalized DWCNTs/Water nanofluid in
turbulent flow at low concentrations, Int. J. Heat Mass Tran. 73 (2014) 186–194.
[21] B.H. Chun, H.U. Kang, S.H. Kim, Effect of alumina nanoparticles in the fluid on heat transfer in double-pipe heat exchanger system, Kor. J. Chem. Eng. 25 (2008)
966–971.

15
L.S. Sundar et al. Case Studies in Thermal Engineering 43 (2023) 102645

[22] E. Ebrahimnia-Bajestana, M. Charjouei Moghadam, H. Niazmand, W. Daungthongsuk, S. Wongwises, Experimental and numerical investigation of nanofluids
heat transfer characteristics for application in solar heat exchangers, Int. J. Heat Mass Tran. 92 (2016) 1041–1052.
[23] G. Huminic, A. Huminic, Heat transfer characteristics in double tube helical heat exchangers using nanofluids, Int. J. Heat Mass Tran. 54 (2011) 4280–4287.
[24] H. Demir, A.S. Dalkilic, N.A. Kürekci, W. Duangthongsuk, S. Wongwises, Numerical investigation on the single phase forced convection heat transfer
characteristics of TiO2 nanofluids in a double-tube counter flow heat exchanger, Int. Commun. Heat Mass Tran. 38 (2011) 218–228.
[25] C. Qi, T. Luo, M. Liu, F. Fan, Y. Yan, Experimental study on the flow and heat transfer characteristics of nanofluids in double-tube heat exchangers based on
thermal efficiency assessment, Energy Convers. Manag. 197 (2019), 111877.
[26] M.H. Bahmani, O.A. Akbari, M. Zarringhalam, G.A.S.S. Shabani, M. Goodarzi, Forced convection in a double tube heat exchanger using nanofluids with constant
and variable thermophysical properties, Int. J. Numer. Methods Heat Fluid Flow 30 (2020) 3247–3265.
[27] W.M. El-Maghlany, A.A. Hanafy, A.A. Hassan, M.A. El-Magid, Experimental study of Cu–water nanofluid heat transfer and pressure drop in a horizontal double-
tube heat exchanger, Exp. Therm. Fluid Sci. 78 (2016) 100–111.
[28] B. Pattanayak, A. Mund, J.S. Jayakumar, K. Parashar, S.K.S. Parashar, Estimation of Nusselt number and effectiveness of double-pipe heat exchanger with
Al2O3–, CuO–, TiO2–, and ZnO–water based nanofluids, Heat Transfer 49 (2020) 2228–2247.
[29] E.I. Jassim, F. Ahmed, Experimental assessment of Al2O3 and Cu nanofluids on the performance and heat leak of double pipe heat exchanger, Heat Mass Tran. 56
(2020) 1845–1858.
[30] L.S. Sundar, N.T.R. Kumar, B.M. Addis, P. Bhramara, M.K. Singh, A.C.M. Sousa, Heat transfer and effectiveness experimentally-based analysis of wire coil with
core-rod inserted in Fe3O4/water nanofluid flow in a double pipe U-bend heat exchanger, Int. J. Heat Mass Tran. 134 (2019) 405–419.
[31] P.V.D. Prasad, A.V.S.S.K.S. Gupta, M. Sreeramulu, L.S. Sundar, M.K. Singh, A.C.M. Sousa, Experimental study of heat transfer and friction factor of Al2O3
nanofluid in U-tube heat exchanger with helical tape inserts, Exp. Therm. Fluid Sci. 62 (2015) 141–150.
[32] P.V.D. Prasad, L.S. Sundar, S. Mesfin, A.C.M. Sousa, Effectiveness and thermal performance Analysis of copper nanofluids flow in a U-bend double-pipe heat
exchanger, Heat Tran. Res. 52 (2021) 31–59.
[33] A.H.M. Isfahani, M. Afrand, Experiment and Lattice Boltzmann numerical study on nanofluids flow in a micromodel as porous medium, Phys. E Low-dimens.
Syst. Nanostruct. 94 (2017) 15–21.
[34] M. Niknejadi, M. Afrand, A. Karimipour, A. Shahsavar, A.H.M. Isfahani, Experimental investigation of the hydrothermal aspects of water–Fe3O4 nanofluid inside
a twisted tube, J. Therm. Anal. Calorim. 143 (2021) 801–810.
[35] H. Azimy, A.H.M. Isfahani, M. Farahnakian, Investigation of the effect of ultrasonic waves on heat transfer and nanofluid stability of MWCNTs in sono heat
exchanger: an experimental study, Heat Mass Tran. 58 (2022) 467–479.
[36] A.R. Miandoab, S.A. Bagherzadeh, A.H.M. Isfahani, Numerical study of the effects of twisted-tape inserts on heat transfer parameters and pressure drop across a
tube carrying Graphene Oxide nanofluid: an optimization by implementation of Artificial Neural Network and Genetic Algorithm, Eng. Anal. Bound. Elem. 140
(2022) 1–11.
[37] L.S. Sundar, S. Sambasivam, H.K. Mewada, ANFIS modelling with fuzzy C-mean clustering of experimentally evaluated thermophysical properties of zirconia-
water nanofluids, J. Mol. Liq. 364 (2022) 119987.
[38] S. Ahmad, S. Saifullah, A. Khan, M. Inc, New local and nonlocal soliton solutions of a nonlocal reverse space-time mKdV equation using improved Hirota bilinear
method, Phys. Lett. 450 (2022), 128393.
[39] S. Khaliq, A. Ullah, S. Ahmad, A. Akgül, A. Yusuf, T.A. Sulaiman, Some novel analytical solutions of a new extented (2 + 1)-dimensional Boussinesq equation
using a novel method, J. Ocean Engineering and Science (2022), https://doi.org/10.1016/j.joes.2022.04.010.
[40] H. Maddah, N. Ghasemi, Experimental evaluation of heat transfer efficiency of nanofluid in a double pipe heat exchanger and prediction of experimental results
using artificial neural networks, Heat Mass Tran. 53 (2017) 3459–3472.
[41] M. Mmohammadiun, F. Dashtestani, M. Alizadeh, Exergy prediction model of a double pipe heat exchanger using metal oxide nanofluids and twisted tape based
on the artificial neural network approach and experimental results, J. Heat Tran. 138 (2016) 1–10, 011801.
[42] M. Hemmat Esfe, Designing a neural network for predicting the heat transfer and pressure drop characteristics of Ag/water nanofluids in a heat exchanger, Appl.
Therm. Eng. 126 (2017) 559–565.
[43] F. Nasirzadehroshenin, H. Maddah, H. Sakhaeinia, A. Pourmozafari, Investigation of exergy of double-pipe heat exchanger using synthesized hybrid nanofluid
developed by modeling, Int. J. Thermophys. 40 (2019) 87, https://doi.org/10.1007/s10765-019-2551-z.
[44] S.D. Kim, J.W. Kim, J.S. Im, Y.H. Kim, Y.S. Lee, A comparative study on properties of multi-walled carbon nanotubes (MWCNTs) modified with acids and
oxyfluorination, J. Fluor. Chem. 128 (2007) 60–64.
[45] B. Saleh, L. Syam Sundar, Thermal performance, embodied energy and environmental CO2 emissions analyses double pipe U-bend heat exchanger working with
MWCNT/water nanofluid, Int. J. Thermal Science 169 (2022), 107094.
[46] F.W. Dittus, L.M.K. Boelter, Heat Transfer in Automobile Radiators of the Tubular Type, University California Publication in Engineering, 1930.
[47] B. Raei, F. Shahraki, M. Jamialahmadi, S.M. Peyghambarzadeh, Experimental study on the heat transfer and flow properties of γ-Al2O3/water nanofluid in a
double-tube heat exchanger, J. Therm. Anal. Calorim. 127 (2017) 2561–2575.
[48] C.P. Kothandaraman, S. Subramanyan, Heat and Mass Transfer Data Book, New Age International Limited Publisher, 2007.
[49] V. Gnielinski, New equations for heat and mass transfer in turbulent pipe and channel flow, Int. Chem. Eng. 16 (1976) 359–368.
[50] R.H. Notter, C.A. Sleicher, A solution to the turbulent Graetz problem—III Fully developed and entry region heat transfer rates, Chem. Eng. Sci. 27 (1972)
2073–2093.
[51] M. Kaya, A.E. Gürel, Ü. Ağbulut, İ. Ceylan, S. Çelik, A. Ergün, B. Acar, Performance analysis of using CuO-Methanol nanofluid in a hybrid system with
concentrated air collector and vacuum tube heat pipe, Energy Convers. Manag. 199 (2019), 111936.
[52] G. Yıldız, Ü. Ağbulut, A.E. Gürel, A review of stability, thermophysical properties and impact of using nanofluids on the performance of refrigeration systems,
Int. J. Refrigeration 129 (2021) 342–364.
[53] M.O. Karaağaç, A. Ergün, Ü. Ağbulut, A.E. Gürel, İ. Ceylan, Experimental analysis of CPV/T solar dryer with nano-enhanced PCM and prediction of drying
parameters using ANN and SVM algorithms, Sol. Energy 218 (2021) 57–67.
[54] Ü. Ağbulut, A.E. Gürel, S. Sarıdemir, Experimental investigation and prediction of performance and emission responses of a CI engine fuelled with different
metal-oxide based nanoparticles–diesel blends using different machine learning algorithms, Energy 215 (2021), 119076.
[55] A.R. Kaladgi, A. Afzal, A.M. Manokar, D. Thakur, U. Agbulut, S. Alshahrani, C.A. Saleel, R. Subbiah, Integrated Taguchi-GRA-RSM optimization and ANN
modelling of thermal performance of zinc oxide nanofluids in an automobile radiator, Case Stud. Therm. Eng. 26 (2021), 101068.
[56] T.G. Beckwith, R.D. Marangoni, J.H. Lienhard, Mechanical Measurements, fifth ed., Addison–Wesley Publishing Company, New York, 1990, pp. 45–112.

16