Applied Thermal Engineering 163 (2019) 114435

Contents lists available at ScienceDirect

Applied Thermal Engineering

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

Heat transfer modeling of a parabolic trough solar collector with working T

fluid of Fe3O4 and CuO/Therminol 66 nanofluids under magnetic field
⁎
Mohammad Malekana, Ali Khosravib, , Sanna Syrib
a
Department of Engineering, Aarhus University, 8000 Aarhus C, Denmark
b
Department of Mechanical Engineering, School of Engineering, Aalto University, Finland

H I GH L IG H T S

• CFD simulation for heat transfer modeling of a parabolic trough solar collector with internal fins is carried out.
• FeIncreasing
3O /Therminol 66 and CuO/Therminol 66 under magnetic field are proposed as working fluids.
4

• The collectortheefficiency
volume fraction and decreasing the particle size enhance the collector efficiency.
• The collector efficiency with Fe O /Therminol 66 is better than CuO/Therminol 66.
3 4

• rises in the presence of the magnetic field.

A R T I C LE I N FO A B S T R A C T

Keywords: Solar energy is among the cleanest and most adaptable compared to other renewable energy sources. The major
CFD simulation challenge is how to get this energy in efficient way to make it available for industrial applications such as
Parabolic trough collector electricity generation. One of the most efficient techniques to harvest solar energy and transform it into electrical
Fe3O4 and CuO nanoparticles energy is parabolic trough solar collector (PTSC), which is a type of concentrating solar power generation
Heat transfer coefficient
systems. This system operates by concentrating solar irradiance onto a tubular receiver in which this centralized
Magnetic field
Smooth and finned tubes
energy is absorbed by a heat transfer fluid and transported to the power cycle. Improving the performance of the
PTSC can enhance efficiency as well as power generation of a PTS power plant. Hence, this issue has been
considered as one of the major challenges for scholars in this field. One promising solution is finding more
efficient heat transfer working fluids. Another suggestion is proposing a different geometry for the receiver. In
the current research, ferrofluids due to their heat transfer characteristics are proposed as working fluid for a
PTSC. Fe3O4/Therminol 66 and CuO/Therminol 66 nanofluids are examined under external magnetic field for
this target. Besides, to improve the heat transfer characteristics of the collector, the receiver is designed with
internal fins. This work is carried out using computational fluid dynamics (CFD). The assessments are done by
considering the different nanoparticle sizes on the friction factor, thermal efficiency, performance evaluation
criteria (PEC) and convective heat transfer.
The results depict that reducing the particle size and enhancing the nanoparticles volume fraction increase the
convective heat transfer coefficient, Nusselt number, PEC and the collector efficiency. In addition, the collector
efficiency rises in the attendance of the magnetic field and maximum efficiency of the collector was obtained for
4% Fe3O4/Therminol 66 working fluid.

1. Introduction is the most abundant and easily accessible. Although many engineering
systems have been designed to catch and absorb the SE (such as pho-
The use of fossil fuels due to their adverse effects on the environ- tovoltaic system and solar thermal facilities), it is crucial to make these
ment, global warming and climate change needs to be reduced [1]. alternative technologies more efficient [4]. Research into solar thermal
Hence researchers strive to attain independence from such sources of systems has provided us with high-efficiency solar collectors, which
energy and move towards alternative and renewable sources like geo- research is continuing aiming to further enhance efficiency.
thermal, wind and specifically solar energy (SE) [2,3]. Among them, SE Although solar photovoltaic system is recognized as one of the most

⁎
Corresponding author.
E-mail address: ali.khosravi@aalto.fi (A. Khosravi).

https://doi.org/10.1016/j.applthermaleng.2019.114435
Received 25 March 2019; Received in revised form 12 August 2019; Accepted 23 September 2019
Available online 23 September 2019
1359-4311/ © 2019 Elsevier Ltd. All rights reserved.
M. Malekan, et al. Applied Thermal Engineering 163 (2019) 114435

Nomenclature μ0 magnetic permeability constant in vacuum

μt turbulent viscosity
Aa collector aperture Nu Nusselt number
B magnetic field Nu0, receiver Nusselt number based on base fluid
bf base fluid P particle
C1ε, C2ε constants PEC performance evaluation criteria
CR collector concentration ratio Qs solar irradiance over the collector
D characteristic length Qu useful heat
ΔP pressure drop r interval between wire position and tube
ε dissipation rate Re Reynolds number
ηth thermal efficiency σk, σε turbulent Prandtl numbers for k and ε
f friction factor Tin, Tout inlet and outlet temperatures
f0 receiver friction factor based on base fluid u fluid velocity
ff ferrofluid
φ volume fraction Abbreviations
Gb direct beam solar irradiation
Gk turbulence kinetic energy generation CFD computational fluid dynamics
h convective heat transfer coefficient CHT convective heat transfer
Hx, Hy magnetic fields in x and y directions DPHX double pipe heat exchanger
I electric current HTC heat transfer coefficient
Ib beam radiation LCOE levelized cost of energy
Ig global radiation MF magnetic field
K thermal conductivity coefficient MWCNT multi-wall carbon nanotube
k turbulent kinetic energy PTSC parabolic trough solar collector
L length of tube PTC parabolic trough collector
ṁ mass flow rate RNG re-normalisation group
μ dynamic viscosity

commercialized technologies for harnessing SE, solar thermal has some heat pipe solar collector. They reported that enhancing the mass flow
striking advantages such as: designing a thermal energy storage system rate inside the collector increased its thermal efficiency. Moreover, the
(using molten salt) to produce the electricity during night, benefiting of HTC of solar collector enhanced when nanofluids were applied instead
a wide range of solar radiation, and durability against damage during of the base fluids. Sharafeldin and Grof [16] proposed WO3/water na-
intense radiation and operation under high temperature [5–7]. A nofluid for an evacuated tube solar system and obtained that the heat
parabolic trough solar collector (PTSC) is a type of solar thermal power gain and solar collector outlet temperature increased with 23% and
generation systems in which the SE is harvested using a parabolic 21%, respectively. Tafarroj et al. [17] analyzed the proficiency of a
mirror and concentrated onto a long metal tube [8,9]. In this system, PTSC for nanofluids through CFD simulation and artificial neural net-
the concentrated solar radiation onto the tube is received by the work. Nanosilica and multi-wall carbon nanotube (MWCNT) were
working fluid and then it is utilized to generate steam for the Rankine mixed together with ethylene glycol as the base fluid. Their results
cycle. A parabolic trough collector commonly uses thermic oils like depicted that the nanofluid containing 0.6% MWCNT had the maximum
Syltherm, Dowtherm, Therminol, and Sandotherm where Therminol temperature outlet.
VP1 and Syltherm 800 due to their reliability and accessibility are more Besides, in this area, molten salts with nanoparticles (nano-salts)
applicable [10]. have been evaluated extensively by many scholars. Yaxuan et al. [18]
Nanofluids are a unique class of industrial fluids that are composed performed an analysis to enhance the performance of bromide salt by
by addition of nanoparticles (for instance Al2O3, CuO, Fe3O4, and SiO2) nano-particle dispersion employed for high-temperature heat pipes in
into the common industrial fluids like water and oil [11]. Nanofluids concentrated solar systems. They reported that decomposing point and
possess exhibited high thermal conductivity and promising thermal heat of fusion were enhanced by 68.4 °C and 99.19%, respectively. Wei
characteristics compared to conventional industrial fluids. For PTSCs, et al. [19] ameliorated the thermal conductivity of liquid nitrate and
many investigations have proposed and evaluated different nanofluids carbonate salts doped with MgO particles. It was derived that the
as working fluid. Tagle-Salzar et al. [12] developed an experimental thermal diffusivity of nitrate salt and carbonate salt remarkably im-
study to assesses the thermal proficiency of the PTSC with Al2O3-water proved with the augmentation of MgO particles. In another study,
nanofluid. It was obtained that the quantity thermal enhanced when specific heat of silica nanofluid was improved by Shin and Banerjee
Al2O3 nanoparticles were added into base fluid. For one collector this [20].
increasing has been reported as 0.3% for heat gain and 0.03% for Ferrofluids are defined as liquids containing single-domain nano-
thermal efficiency. Khakrah et al. [13] implemented an exergy analysis particles (for instance iron nickel oxide, cobalt and their oxides, ferric
for a PTSC with working fluid of Al2O3/synthetic oil. It was found out oxide, etc.) with an average diameter of 15 nm or less. Ferrofluids are
that utilizing nanoparticles with 5% volume fraction enhances the re- simulated through ferrohydrodynamic governing equations under
lative exergy efficiency approximately around 19%. Korres et al. [14] magnetic field (MF). As we know, thermal conductivity of metals is
evaluated the thermal efficiency of a PTSC with Syltherm 800/CuO higher than liquids. Hence, it will be predicted ferrofluids that are a
(nanoparticle concentration: 5%) working fluid. They have shown that colloidal mixture of magnetic metal particles in a base fluid perform
the medium and uttermost heat transfer coefficient (HTC) improve- better than the conventional industrial fluids in thermal characteristics.
ments were 16.16% and 17.41%, respectively. In addition, they have Various numerical efforts have been made to assess the thermal per-
shown that the values of pressure drop as well as the pumping work formance of ferrofluids with/without the attendance of MF. Malekan
demand for nanofluid in all cases were too low. Shafiey and Zamani and Khosravi [21] assessed the result of MF upon the HTC of Fe3O4/
[15] examined MgO nanofluids to improve the thermal efficiency of a water nanofluid trough an intelligent model called adaptive neuro-

2
M. Malekan, et al. Applied Thermal Engineering 163 (2019) 114435

fuzzy inference system optimized with particle swarm optimization, Therminol 66 and Fe3O4/Therminol 66 under various MF for a PTSC.
and CFD simulation. It was derived that the presence of MF can enhance These investigations are undertaken while the central receiver is de-
the HTC of ferrofluid. Malekan et al. [22] enhanced the thermal effi- signed with internal fins. As a matter of fact, in the current study, for
ciency of a double pipe heat exchanger (DPHE) incorporated with a the first time, the distribution of the thermal and fluid flow char-
small scale CAES (compressed air energy storage) system by con- acteristics of the mentioned ferrofluids impressed by MF for the finned
sidering a ferrofluid as secondary fluid for the exchanger. Khosravi tube and smooth tube are compared and investigated. Besides, the
et al. [10] proposed Fe3O4-Therminol 66 (the examined volume fraction impress of the particle size of nanoparticles on the thermal demeanor of
(1–4%)) at presence of magnetic field (0–500 G) as working fluid for a the collector is determined. Moreover, this research illuminates the
PTSC. Their results have illustrated that using MF improves the thermal contribution of geometry changes against working fluid changes.
characteristics of the solar collector including thermal efficiency, HTC,
and output temperature. In another research work, the HTC of nano-
fluids under MF was evaluated by Jafari et al. [23]. They assessed the 2. Methodology
heat transfer of nanofluids under MF in a helical DPHX and for laminar
stream in which increasing the HTC at presence of MF was reported. 2.1. Parabolic trough solar system
Khosravi and Malekan [24] developed intelligent methods to predict
the HTC of Fe3O4/water ferrofluid in attendance of various MFs and Commonly a PTSC power plant that commonly uses U-curved mir-
operational conditions. Aminfar et al. [25,26] assessed the hydro-dy- rors in order to harvest solar radiation. This system is classified between
namic and hydro-thermal manner of ferrofluids by using the non-uni- the concentrating solar systems in which the direct normal solar irra-
form transverse/axial MF. diance is collected and transformed to the thermic energy. This con-
Based on the literature review, many articles have proposed nano- centrated energy is received by the working fluid and its evaporating
fluids (such as Al2O3/water, SiO2/water, TiO2/water, etc.) as working lead to electricity generation. In this research, in order to attain to
fluids for PTSCs. Extensive research works have been carried out on the solution for increasing the thermal efficiency of the solar collectors,
effect of MF on heat transfer of ferrofluids in diverse industrial pur- ferrofluids are proposed and evaluated using CFD simulation.
poses. It could be plainly viewed that in most of them Fe3O4 nano- Generally, a PTSC power plant is incorporated with a solar field (filled
particle has been appraised for ferrofluid suspension. In the case of with lots of solar collectors), power cycle (for instance a Rankine cycle),
PTSC, CuO has been recommended as nanoparticle suspended in the and in some cases, thermal energy storage system accompanied by fossil
base fluid. There is no study to figure out the thermal and fluid behavior fuel backup system. The solar field includes the parabolic, trough-
of CuO under MF and this study has been intrigued to fill this knowl- shaped collectors that navigate normal solar irradiance onto tubular
edge gap. Indeed, this study provides a comparison between the CuO/ receivers. Each collector contains the mirrors and its structure in order
to support the receivers, mirrors as well as sun tracker system; and each

Fig. 1. (a) Schematic of the PTC model, (b) CR distribution in the receiver, and (c) collector tube cross section, with and without fins, and (d) a typical grid
discretization of the problem.

3
M. Malekan, et al. Applied Thermal Engineering 163 (2019) 114435

receiver contains a metal pipe with a solar absorbing surface. Table 1

Physical attributes of all heat transfer fluids and nanoparticles [29,40].
2.2. Model definition Material ρ (kg/m3) Cp (J/kg. K) k (W/m. K) μ (kg/m. s)

The utilized model in this work includes a solar thermal collector Therminol 66 (bf) 899.5 2122 0.107 0.00106
and a receiver tube, as demonstrated in Fig. 1(a). A non-uniform MF is Fe3O4 (p) 5200 670 6 –
CuO (p) 6500 540 18 –
generated by embedding an electrical wire parallel to the receiver tube
Therminol 66 – φ = 2% 985.5 2092.96 0.113206 0.001113
axial direction, hereunder the tube. Further, Fig. 1(b) presents the ap- Fe3O4 (ff) φ = 4% 1071.5 2063.92 0.119657 0.001166
proximate model of the local concentration ratio (CR ) for the receiver Therminol 66 – φ = 2% 1011.51 2090.36 0.113431 0.001113
tube. This figure demonstrates that only the tube lower half receives the CuO (ff) φ = 4% 1123.52 2058.72 0.120125 0.001166
reflected solar radiation by the collector.
Area under curves of the simplified model and typical LCR profile
Table 2
from [27] show an error of less than 5%, which indicates the simplified
Mesh discretization effect on the thermal efficiency and y+ parameter of the
model can be considered as a good approximation of the typical LCR
solar collector.
profile. This simplified model was also proposed by Munoz et al. [28].
Two types of collector tubes are considered in this study, smooth and Mesh type Coarse Medium Fine Very fine
finned tubes as shown in Fig. 1(c). These two configurations are con-
# of cells 2,909,300 5,753,248 8,461,704 12,022,300
sidered to evaluate the effects of fins on the PTC performance with and y+ 2.1 1.6 1.25 1.15
without presence of MFs. In addition, a typical grid discretization of the ηth 0.7195 0.7258 0.7295 0.7305
receiver tube problem is shown in Fig. 1(d).

2.3. Governing equations under presence of magnetic fields where FK (x ) and FK (y ) are Kelvin force (owing to the magnetic gra-
∂H
dient). The components of Kelvin force are defined as μ0 M ∂x and
Numerical analysis is done by coupling the equations of energy and ∂H
μ0 M ∂y in x and y directions, respectively. These components are ob-
Navier-Stokes to ascertain the heat transfer specifications inward the tained as a result of the electric current through the wire. The position
tube. The influence of MF is accounted by calculating the components of this current-carrying wire and schematic of distribution of MFs are
of MF in the momentum equations. Continuity equation is defined as presented in Fig. 2. The MF of electric current are Hx and Hy , in the x
[29]: and y directions respectively, and are defined by the following equa-
∂u ∂v ∂w tions [30]:
+ + =0
∂x ∂y ∂z (1) I (x − a)
Hx (x , y ) =
in which, following assumption were used to utilized continuity equa- 2π (x − a)2 + (y − b)2 (6)
tion to numerically model the current nanofluid: steady, and in-
compressible and laminar nanofluid flow, thermo-physical properties to I (y − b)
Hy (x , y ) =
be constant, and viscosity loss to be negligible. Momentum equations 2π (x − a)2 + (y − b)2 (7)
are [29]:
The MF strength is calculated by [24,29]:
∂u ∂u ∂u ∂p ∂ 2u ∂ 2u ∂ 2u
ρ ⎛u
⎜ +v +w ⎞=− + μ ⎛ 2 + 2 + 2 ⎞ + FK (x )
⎟ ⎜ ⎟
H (x , y, z ) =
I 1
⎝ ∂x ∂y ∂z ⎠ ∂x ⎝ ∂x ∂y ∂z ⎠ (2) 2π (x − a)2 + (y − b)2 (8)
∂v ∂v ∂v ∂p ∂ 2v ∂ 2v ∂ 2v Also, M is defined as magnetization and is obtained by [24,29]:
ρ ⎛u
⎜ +v +w ⎞=− + μ ⎛ 2 + 2 + 2 ⎞ + FK (y )
⎟ ⎜ ⎟

⎝ ∂x ∂y ∂z ⎠ ∂y ⎝ ∂x ∂y ∂z ⎠ (3)
6mp ⎡ 1
M= coth (ξ ) − ⎤
∂w ∂w ∂w ∂p ∂ 2w ∂ 2w ∂ 2w ⎞ πdp3 ⎢ ξ⎥ (9)
ρ ⎛u
⎜ +v +w ⎞=− + μ⎛ 2 +
⎟
2
+ ⎜ ⎟
⎣ ⎦
⎝ ∂x ∂y ∂z ⎠ ∂x ⎝ ∂ x ∂ y ∂z 2 ⎠ (4)
where ξ is the Langevin parameter and can be obtained by the following
Energy equation is defined as [29]: formula [24,29]:
∂T ∂T ∂T ∂ 2T ∂ 2T ∂ 2T ⎞ μo mp H
(ρCp) ⎛u
⎜ +v + w ⎞ = k⎛ 2 + ⎟ ⎜
2
+ ⎟ ξ=
⎝ ∂x ∂y ∂z ⎠ ⎝ ∂ x ∂ y ∂z 2 ⎠ (5) KB T (10)

Fig. 2. Schematic of position of an electric wire accompany with the tube cross section which shows distribution of the MF.

4
M. Malekan, et al. Applied Thermal Engineering 163 (2019) 114435

Fig 3. Nusselt number variation for the current model and from the theories.

and the particle magnetic moment is [31]: 2.4. Thermal investigation

4μB πdp3
mp = The major goal of this investigation is to appraise the heat transfer
6 × 91.25 × 10−30 (11) from the absorber to the heat transfer fluid. It concerns to the con-
In order to quantify the MF intensity effects on the working mag- vection heat transfer coefficient, in which smaller magnitudes of this
netic nanofluid, a dimensionless magnetic number (Mn) can be utilized. variable causes a higher absorber temperature resulting in higher
This dimensionless value is proportional to the MF intensity and is thermal losses. PTSCs employ the direct beam solar radiation (Gb). The
defined as [29]: existing radiation upon the collector is defined by Qs = Aa ·Gb , and the
useful heat is calculated as Qu = ṁ ·cp·[Tout − Tin], in which where Aa, ṁ ,
μ0 χH 2h2 Tin, and Tout are the collector aperture, mass flow rate, inlet and outlet
Mn =
ρα 2 (12) temperatures, respectively. The convection HTC (h) can be calculated
using Nusselt number. On the plus side, Nusselt number pertains on the
where μ0, h, χ, ρ, and α are magnetic permeability in vacuum, tube
problem geometry and flow conditions [10]. Eqs. (16)–(18) indicate
hydraulic diameter, magnetic susceptibility and density of the nano-
corresponding formula for Nusselt number, Prandtl number, and fric-
fluid, and thermal diffusivity (1.599 × 10−7 m2/s).
tion factor [10]:
The k-ε RNG turbulence method is applied in this study [28,32]:
hDi
→ μ Nu =
∇ ·(ρff vff k ) = ∇ ·⎡ ⎛μ + t ⎞ ∇k⎤ + Gk − ρff ε
⎜ ⎟ k (16)
⎢⎝ σk ⎠ ⎥ (13)
⎣ ⎦
μCp
→ μ C2ε ρff ε 2 Pr =
∇ ·(ρff vff ε ) = ∇ ·⎡ ⎛μ + t ⎞ ∇ε⎤ + C1ε (ε / k ) Gk −
⎜ ⎟ k (17)
⎢⎝ σε ⎠ ⎥ k (14)
⎣ ⎦
where k is turbulent kinetic energy, ε is dissipation rate, Gk is the 2ΔP Di
f= ⎛ ⎞
generation of turbulence kinetic energy, σk and σε are the turbulent ρff u2 ⎝ L ⎠ (18)
Prandtl numbers for k and ε, respectively, C1ε and C2ε are two constants,
and μt is the turbulent viscosity. The constants for the k-ε RNG model where ΔP , L and u are the pressure loss, length of the receiver tube and
are defined by [33]: fluid velocity, respectively. In turbulent flow regime (Re > 2300), the
Nusselt number is determined based on Colburn correlation [34] as:
k2
μt = ρff Cμ ε
, Cμ = 0.0845
Nu = 0.023·Re 0.8 ·Pr 0.4 (19)
C1ε = 1.42, C2ε = 1.68
σk = 1, σε = 1.3 Another relationship that can be used to calculate the Nusselt
Gk = 2μt Eij Eij, Eij =
1
2 ( ∂ui
∂xj
+
∂uj
∂xi ) (15)
number for nanofluid was proposed by Leinhard and Leinhard [35] as
follows:

5
M. Malekan, et al. Applied Thermal Engineering 163 (2019) 114435

Fig. 4. Local convective HTC (h): (a) comparison between two particle sizes, (b) particle size of 10 nm, and (c) particle size of 20 nm.

Fig. 5. Results for the base fluid for Tin = 503 K: (a) outlet temperature (K) for Re = 31,000, and velocities (m/s) for: (b) Re = 31,000 and (c) Re = 62,000.

6
M. Malekan, et al. Applied Thermal Engineering 163 (2019) 114435

Fig. 6. Outlet velocity (m/s) assuming Re = 31,000, Tin = 503 °K, φ = 4% and particle size of 10 nm, for: (a) CuO with B = 0, (b) Fe3O4 with B = 0, (c) CuO with
B = 250G, and (d) Fe3O4 with B = 250G.

Nu =
( )·Re·Pr
fth
8
with fth =
184
for turbulent flow kff =
⎛ kp + (n − 1) kbf − (n − 1) φ (kbf − kp) ⎞
⎜ ⎟
k
fth Re 0.2 kp + (n − 1) kbf + φ (kbf − k p)
1 + 12.8· ·(Pr 0.68 − 1) ⎝ ⎠ bf (22d)
8

(20) where φ and n are the volumetric fraction of the nanoparticles and
shape factor, respectively. Eq. (22d) was developed by Hamilton &
Petukhov relation [36] is a widely used theoretical way to calculate Crossor [39]. Therminol 66 and two different nanoparticles (Fe3O4 and
the friction factor using Reynolds number, as: CuO) are used to form different ferrofluids for this study, with the
f = (0.79·lnRe − 1.64)−2 (21) different φ values. This volumetric fraction typically varies between 0.1
and 4%. Table 1 epitomizes the physical attributes of the whole heat
transfer fluids and nanoparticles.
2.5. Ferrofluid physical properties
3. Results and discussion
Ferrofluid physical properties are calculating by considering the
ANSYS® Fluent® version 19.1 is adopted to solve the governing
attributes of based fluid (bf) and nanoparticles simultaneously and are
equations of a three-dimensional steady-state model. As it was ex-
presented by [24,37,38]:
plained in the previous section, the turbulent model adopted here is k-ε
ρff = (1 − φ) ρbf + φρp (22a) RNG model in order to model the PTSC problem [28]. The solution of
momentum and energy equations are obtained using the second order
Cp, ff = (1 − φ) Cp, bf + φCp, p upwind differencing scheme. A convergence limit of 10−3 and 10−6 are
(22b)
considered for momentum, mass and energy equations. A user-defined
function code is written to include the ferrofluid properties and MF in
μff = (1 + 2.5φ) μbf (22c)
all the simulations.

7
M. Malekan, et al. Applied Thermal Engineering 163 (2019) 114435

Fig. 7. Outlet temperature (K) for ferrofluids with Re = 31,000, Tin = 503 °K, φ = 4% and particle size of 10 nm, for: (a) CuO with B = 0, (b) Fe3O4 with B = 0, (c)
CuO with B = 250G, and (d) Fe3O4 with B = 250G.

The flowing fluid possesses uniform velocity and temperature at the u∗·y
y+≡
tube inlet as u = u 0 , Tf = T0 = 230 °C . As it was stated before, an v (24)
outer surface of the tube receives a uniform heat flux due to the solar
reflection from the collector (taking into account the mirror efficiency), in which u∗, y, and v are defined as friction velocity at the nearest wall,
while top surface receives direct solar radiation for the sun. Therefore, distance to the nearest wall, and local kinematic viscosity of the fluid,
top and bottom half tube surfaces are subjected to qup ̇ = Ig and respectively. A value close to 1 can be considered as a good output for
̇
qdown = Ib CR , respectively, in which Ig is global radiation with the in- the y+, and hence, the discretized model can capture near wall
tensity of 680 W/m2, Ib is beam radiation with the intensity of 630 W/ boundary layer adequately [42–44]. A mesh sensitivity analysis was
m2, and CR is about 15.46. Also, outlet boundary condition is assumed made for the receiver tube and thermal efficiency of solar collector as
to be zero pressure gradient. In addition, it is assumed that nano- well as y+ of receiver tube were calculated for different meshes.
particles have spherical shape with a diameter of 10–20 nm, and hence Table 2 gives the results of the mesh sensitivity analysis, in which it was
the shape factor (n) would be equal to 3 [29]. The Biot–Savart law is performed for the smooth absorber and for inlet temperature equal to
applied to compute the MF produced by current-carrying wire as: 230 °C. Finally and according to these results, a mesh with around
twelve million cells (very fine case) was selected as a good discretiza-
μ0 I tion model for current problem. A schematic of the discretized model
B=
2πr (23) with the very fine mesh case was shown in Fig. 1(d).
where r is the distance between MF calculation point and the wire lo- Variation of Nusselt number obtained from the current simulation
cation. for smooth tube with all the working fluids with nanoparticle size of
The discretized model for the current work is composed by hex- 10 nm, and those obtained using either Eqs. (19) and (20) are used to
ahedral cells in all regions (fluid and receiver tube). The boundary layer validate the numerical model. Equation (19) is used to validate the base
between the receiver tube inner wall and the working fluid was defined fluid, while Eq. (20) is used to validate the ferrofluids, according to
in such a way to have a good output for the y+ parameter, which can [27,45]. Fig. 3 presents the outputs of the validity study, in which the
be calculated as [41]: current CFD outputs are in accordance with the results from theory,

8
M. Malekan, et al. Applied Thermal Engineering 163 (2019) 114435

Fig. 8. Effects of MF on the Nusselt number: (a) comparison between two particle sizes, (b) particle size of 10 nm, and (c) particle size of 20 nm.

leading to have less than 6% error between them. In all the graphs in which is mainly due to have more nanoparticles in the ferrofluid, and
this section, the curve names consist of the following parts: CuO or hence, higher thermal conductivity for the ferrofluid. Furthermore, it
Fe3O4 which refers to the type of nanoparticles, 2% or 4% refers to the was observed that the applied MF enhances the cooling performance of
volumetric fraction of the particles, B refers to the intensity of the MF, the ferrofluid. Furthermore, the 4% Fe3O4/Therminol 66 ferrofluid with
and 10 nm or 20 nm refers to the size of nanoparticles. When needed, a 10 nm particles size and with the MF showed the highest convective
magnification was added to some figures to better understand the HTC. The HTC of ferrofluid and particle size have an inverse relation
comparison between different cases. (one decreases, the other increases).
The gradient of velocity close to the walls is enhanced mainly be-
cause of the presence of non-uniform transverse MF. This results in
3.1. Simulation results for smooth tube
increasing the HTC value, as it was shown in Fig. 4. Fig. 5 shows
contour of temperature and velocity for the base fluid at the outlet,
Ferrofluids with different volume fractions (2% and 4%) are used as
considering the inlet temperature of 503 K and Re of 31,000 and
the working fluid for the PTSC. As above mentioned, the MF in-
62,000, respectively. In addition, Fig. 6 illustrate outlet velocity for
vestigated in the research, was generated through a wire placed near to
Tin = 503 K, B = 0 G, and 250 G , and Reff = 31, 000 , and for particle
the receiver tube. MF effects on the local HTC for nanoparticles of
size of 10 nm for both CuO and Fe3O4 ferrofluids. This figure demon-
10 nm and 20 nm are shown in Fig. 4. The figure reveals that the HTC
strates the enhancement in velocity in the case of ferrofluid with the
has a tendency to increase with a growth in Reynolds number for all
MF. Also, Fe3O4 ferrofluid shows better performance than CuO in terms
working fluids, and also nanoparticles help to possess higher HTC
of velocity distribution at the outlet, comparing Fig. 6a and c with
compared to the base fluid. In addition, the HTC shows an increasing
Fig. 6b and d.
behavior for higher values of φ, as clearly can be seen in Fig. 4b and c,

9
M. Malekan, et al. Applied Thermal Engineering 163 (2019) 114435

Fig. 9. Variation of the friction factor (f) in the presence of MFs: (a) comparison between two particle sizes, (b) particle size of 10 nm, and (c) particle size of 20 nm.

Fig. 7 shows the temperature distributions at the outlet for both Nusselt number.
ferrofluids with volume fraction of 4%, Re of 31,000, Tin equal to 503 K, Friction factor variation for both CuO and Fe3O4 ferrofluids are
particle size of 10 nm, and MF of 0 and 250G. Though surface tem- shown in Fig. 9, with and without attendance of MF, different volume
perature of the receiver tube for B = 0 reaches a higher value than fractions and particle sizes. Higher density and viscosity for ferrofluids
B = 250, MF forces the majority of fluid to flow with the maximum result in higher friction factor than the base fluid. In addition, the
temperature. In addition, maximum temperature of CuO ferrofluid is highest friction factor from all cases is for Fe3O4 ferrofluid with 20 nm
bigger than Fe3O4, for both MFs of 0 and 250G. However, larger portion particle size and 250G MF; see Fig. 9a. Fig. 9b and c show the friction
of the Fe3O4 case reach a bigger temperature value, see minimum factor for particle size of 10 nm and 20 nm, respectively. More im-
temperature of all cases from Fig. 8. portantly, friction factor increases in the attendance of MF, according to
Fig. 8 presents the variation of Nusselt number in the presence of Fig. 9.
MF and for two different nanoparticle sizes. Similar to the HTC varia- According to the results presented in Fig. 9, friction factor drops
tion from Fig. 4, smaller particle size improves better the variation of with the higher values of Reynolds number. As it was presented in Eq.
the Nusselt number than bigger particle sizes, by comparison between (18), there is an inverse relationship between friction factor and velo-
Fig. 8a and b. Nusselt number grows with Reynolds number for all city. Velocity streamlines at the receiver tube outlet are presented in
working fluids, while ferrofluid experiences higher values compared to Fig. 10, for B = 0 and 250G, Reff = 31, 000 , φ = 4%, and particle size of
the base fluid. As it was stated before, incorporating nanoparticles 10 nm. It could be noticed that the MF forces the boundary layers to
within ferrofluids led to have higher thermal conductivity, mainly due move to the tube central parts, similar to the distribution of outlet
to the high conductivity of nanoparticles. Therefore, HTC of the ferro- temperature. As can be observed from Fig. 10c and d, force generated
fluid are increased more than base fluid. Higher φ values lead to have by the MF in the transverse plane produces secondary flows and two
bigger effective thermal conductivity for ferrofluid, and hence, higher vortices. This force can be augmented by enhancing the intensities of

10
M. Malekan, et al. Applied Thermal Engineering 163 (2019) 114435

Fig 10. Outlet velocity streamline (in m/s) for Re = 31,000, Tin = 503 °K, φ = 4% and particle size of 10 nm, for: (a) CuO with B = 0, (b) Fe3O4 with B = 0, (c) CuO
with B = 250G, and (d) Fe3O4 with B = 250G.

MF. On the other hand, the maximum outlet velocity for B = 0 is The receiver thermal loss can be calculated using the following re-
concentrated at the tube central part, see Fig. 10a and b. Due to the lation [27]:
presence of MFs, ferrofluid particles are more concentrated near the Qloss = Qabs − Qu (26)
tube bottom side, which clarifies why there are vortices in Fig. 10c and
d, and no vortices in Fig. 10a and b. where Qabs is defined as the absorbed solar energy and can be estimated
Performance evaluation criteria (PEC) has been introduced as a as follows [27]:
proven approach to appraise the thermo-hydraulic manner of flowing Qabs = Qs ·ηopt,max ·K (θ) (27)
working fluid in the receiver tube [46], which is calculated by:
Nu/ Nu 0 in which ηopt,max is the maximum optical efficiency and for current work
PEC = is equal to 75.5%, and K (θ) is the incident angle modifier. This last
(f / f0 )1/3 (25)
parameter depends on the incident angle on the collector aperture (θ )
in which Nu0, f0, f and Nu are the Nusselt number and friction factor of and for the present study is considered equal to 1 since the PTS collector
the receiver tube working with base fluid and ferrofluid, respectively. is surveyed for zero incident angle. Fig. 12 present the receiver thermal
Accordingly, for an effective approach the PEC value is greater than loss variation versus Reynolds number for all the cases with smooth
one, which means a better performance from thermo-hydraulic point of receiver tube, utilizing the Eq. (26). As it was expected, the thermal loss
view. PEC variation for both ferrofluids with volume fractions of 2 and for the PTSC with the base fluid is bigger than all the other cases, and
4%, both particle sizes, and different MF intensities, is shown in Fig. 11. lower heat loss was obtained with Fe3O4/Therminol 66 ferrofluid of 4%
As it can be seen from these two figures, Fe3O4/Therminol 66 with volumetric fraction and with MF intensities of 250 G.
φ = 4% and B = 250 shows the best thermo-hydraulic performance. As The thermal efficiency of the solar collector is defined using the
it was explained before, MF leads to a bigger HTC and friction factor. In following equation [27]:
addition, the enhancements in HTCs are bigger than friction factors,
Qu
and as result, the PEC for ferrofluids with B = 250 show an increasing ηth =
Qs (28)
behavior.

11
M. Malekan, et al. Applied Thermal Engineering 163 (2019) 114435

Fig. 11. Variation of PEC in the presence of MF, for: (a) particle size of 10 nm, and (b) particle size of 20 nm.

12
M. Malekan, et al. Applied Thermal Engineering 163 (2019) 114435

Fig. 12. Variation of heat loss in the presence of MF, for: (a) particle size of 10 nm, and (b) particle size of 20 nm.

The main parameter to evaluate the solar collector is its thermal higher thermal conductivity for the ferrofluids, and according to Fig. 13
efficiency. Variation of thermal efficiency for all working fluids, with this leads to an increase in the efficiency. Therefore, more SE can be
different φ and B values, and with particle size of 10 nm and 20 nm is absorbed by the receiver tube with ferrofluids, and hence resulting in
presented in Fig. 13. MF helps to have higher thermal efficiency for the more thermal energy conversion. Korres et al. [14] have reported that
collector. As it was stated before, the bigger values of φ lead to have the thermal efficiency of a PTSC with Syltherm 800/CuO nanofluid

13
M. Malekan, et al. Applied Thermal Engineering 163 (2019) 114435

Fig. 13. Thermal efficiency for all working fluids: (a) particle size of 10 nm, and (b) particle size of 20 nm.

14
M. Malekan, et al. Applied Thermal Engineering 163 (2019) 114435

Fig. 14. MF and fins effects on the: (a) HTC, (b) Nusselt number, (c) friction factor, and (d) PEC, for nanoparticle size of 10 nm.

showed an average of 1.24% enhancement compared to Syltherm 800 from this figure, finned tube shows a better performance in terms of all
case. According to Fig. 14, the average of enhancement in the thermal these parameters, where the best enhancement behavior is for the
efficiency for CuO/Therminol 66 is around 1.35% for B = 0 and around finned tube with Fe3O4/Therminol 66 working fluid and in the presence
4.0% for B = 250 G . of MF.
Fig. 15 presents distribution of the outlet temperature and outlet
velocity streamline for finned tube with base fluid and Fe3O4/Ther-
3.2. Results for finned tube minol 66 ferrofluid (4% volume fraction). Particle size is 10 nm, Rey-
nolds number is 30,000 and MF is 250G. The effects of fins to increase
According to Bellos et al. [47] and Olia et al. [48], incorporating the outlet temperature are clarified by comparing Fig. 15a with 5a and
fins within collector tube can enhance the performances of the PTC 15b with 7d, with average outlet temperatures of 513.85, 514.25,
system. Four fins with the length of 10 mm were utilized for the finned 514.41, and 514.56, respectively. In addition, velocity streamline
receiver tube, as shown in Fig. 1c. In addition, as it was discussed in magnitudes and configurations are affected by the presence of fins,
Section 3.1, nanoparticle size of 10 nm show better performance than comparing Fig. 15d with Fig. 10d. Fig. 15c is used to compare the base
the 20 nm particle size. Also, Fe3O4/Therminol 66 ferrofluid delivered fluid velocity distribution with ferrofluid distribution in the presence of
better performance than the CuO/Therminol 66, and of course with fins and MF.
larger volume fraction, i.e., 4%. Therefore, Fe3O4/Therminol 66 ferro- Similar to PTC with smooth tube, the heat loss of the PTC with
fluid with φ = 4% and particle size of 10 nm are considered for the finned tube along with corresponding results for the smooth tube are
finned tube, to analyze the effect of MF along with the fins on the shown in Fig. 16. As can be seen from this figure, the PTC with finned
performance of the PTC system. tube delivered the lower heat loss compared with the smooth receiver
Fig. 14 presents variation of heat transfer coefficient, Nusselt tube. In addition and similar to the previous findings, the presence of
number, friction factor, and performance evaluation criteria for finned MFs helped to have even lower hear loss in comparison to PTC without
tube with and without presence of MFs and for base and Fe3O4 nano- the presence of MFs.
fluid with 4% volume fraction of 10 nm particle size. As can be seen

15
M. Malekan, et al. Applied Thermal Engineering 163 (2019) 114435

Fig. 15. (a) and (b) outlet temperature (K) for base fluid and Fe3O4 with B = 250G, (c) and (d) outlet velocity streamline (in m/s) for base fluid and Fe3O4 with
B = 250G. Other parameters are: Re = 31,000, Tin = 503 °K, φ = 4% and particle size of 10 nm, for finned tube.

As a result of enhancing outlet temperature due to the presence of calculated for all the operating PTC lifetime. Considering the typical
fins, the solar collector efficiency is also increases as presented in lifetime of mechanical equipments, parameter N for the current study is
Fig. 17. This figure is also for the base fluid and Fe3O4 ferrofluid with considered to be around 24,000 h (total life of 20 years, with 1200 h per
particle size of 10 nm. MF intensities are 0 and 250G and volume year). The reader should note that the LCOE estimates presented here
fraction of nanoparticles are 4%. As can be seen from this figure, finned are only cost indications of the PTC collector part with the working
receiver tube with ferrofluid in the presence of MF delivers the highest fluids studied here, not for the whole solar power plant.
efficiency compared to the other cases. Fig. 18 presents the LCOE variation with Reynolds number for the
Another important parameter to be examined is the financial as- current PTC with smooth and finned tubes, and for different working
sessment of the collector with smooth and finned tubes. To do this, the fluids. According to Bellos et al. [47], the LCOE for smooth and finned
levelized cost of energy (LCOE) was proven to be the best approach tubes are almost the same for inlet temperature of 493–503 K, as can be
[28,49]. According to Conrado et al. [49], the PTC cost with smooth seen in Fig. 18a. While LCOE would be larger for finned tube that
receiver tube is about 200 Euros/m2, while cost of the finned tube is smooth one for lower inlet temperatures. Therefore, PTC with smooth
about 3% higher [28]. Current PTC collector has 85 m2 aperture, and tube must be selected for low inlet temperatures, while for higher inlet
hence, the cost of PTC with smooth and finned tubes is around 17,000 temperatures than (higher than 503 K) the finned tube must be selected.
and 17,510 Euros, respectively. Following equation can be used to
calculate the LCOE [47]: 4. Conclusions
CO CO
LCOE = = In the current research, a numerical simulation was exerted for the
Qu N ηth Qs N (29)
comparison between CuO/Therminol 66 and Fe3O4/Therminol 66 at
in which N is the system total operation hours. This parameter is the presence of MF, which were proposed as working fluid for a

16
M. Malekan, et al. Applied Thermal Engineering 163 (2019) 114435

Fig. 16. Heat loss of the solar collector with smooth and finned tubes, for base and Fe3O4/Therminol 66 fluids. Particle size is 10 nm, and MFs are B = 0 and 250.

Fig. 17. Thermal efficiency of the solar collector with smooth and finned tubes, for base and Fe3O4/Therminol 66 fluids. Particle size is 10 nm, and MFs are B = 0
and 250.

17
M. Malekan, et al. Applied Thermal Engineering 163 (2019) 114435

Fig. 18. LCOE for collector with smooth and finned tubes, for different working fluids and MFs are B = 0 and 250, and for: (a) particle size of 10 nm, and (b) particle
size of 20 nm.

18
M. Malekan, et al. Applied Thermal Engineering 163 (2019) 114435

parabolic trough solar collector (PTSC). This comparison was done collectors under condition of multiple optical factors, Appl. Therm. Eng. 160 (Sep.
using CFD approach. Besides, the central receiver was designed with 2019) 114070.
[8] F.A. Al-Sulaiman, M.I. Zubair, M. Atif, P. Gandhidasan, S.A. Al-Dini, M.A. Antar,
internal fins and the distribution of the thermal and fluid flow char- Humidification dehumidification desalination system using parabolic trough solar
acteristics of the ferrofluids impressed by MF for the finned tube and air collector, Appl. Therm. Eng. 75 (Jan. 2015) 809–816.
smooth tube were compared and evaluated. The influence of MF and [9] B. Zou, H. Yang, Y. Yao, Y. Jiang, A detailed study on the effects of sunshape and
incident angle on the optical performance of parabolic trough solar collectors, Appl.
particle size of nanoparticles over the collector efficiency were analyzed Therm. Eng. 126 (Nov. 2017) 81–91.
and discussed. The following results were drawn from the current [10] A. Khosravi, M. Malekan, M.E.H. Assad, Numerical analysis of magnetic field effects
study: on the heat transfer enhancement in ferrofluids for a parabolic trough solar col-
lector, Renew. Energy 134 (Apr. 2019) 54–63.
[11] S. Pavlovic, A.M. Daabo, E. Bellos, V. Stefanovic, S. Mahmoud, R.K. Al-Dadah,
▪ Enhancing Reynolds number rises the convective heat transfer Experimental and numerical investigation on the optical and thermal performance
coefficient (CHTC) for the base fluid and ferrofluids. of solar parabolic dish and corrugated spiral cavity receiver, J. Clean. Prod. 150
(May 2017) 75–92.
▪ When the volume fraction and particle size of nanoparticles en-
[12] P.D. Tagle-Salazar, K.D.P. Nigam, C.I. Rivera-Solorio, Heat transfer model for
hance, the CHTC augments as well. The maximum CHTC belongs to thermal performance analysis of parabolic trough solar collectors using nanofluids,
4% Fe3O4/Therminol 66 with the particle size of 10 nm and in the Renew. Energy 125 (Sep. 2018) 334–343.
presence of MF. Rising the particle size decreases the CHTC. The [13] H. Khakrah, A. Shamloo, S. Kazemzadeh, Hannani, Exergy analysis of parabolic
trough solar collectors using Al2O3/synthetic oil nanofluid, Sol. Energy 173 (2018)
same outcomes were obtained for the Nusselt number. 1236–1247.
▪ Investigation of the friction factor for the ferrofluids and base fluid [14] D. Korres, E. Bellos, C. Tzivanidis, Investigation of a nanofluid-based compound
demonstrated that the friction factor of the ferrofluids is higher than parabolic trough solar collector under laminar flow conditions, Appl. Therm. Eng.
149 (Feb. 2019) 366–376.
the base fluid. The 4% Fe3O4-Therminol 66 with the particle size of [15] M.S. Dehaj, M.Z. Mohiabadi, Experimental investigation of heat pipe solar collector
20 nm and under MF has the maximum friction factor. using MgO nanofluids, Sol. Energy Mater. Sol. Cells 191 (Mar. 2019) 91–99.
▪ Performance evaluation criteria (PEC) was considered as an eva- [16] M.A. Sharafeldin, G. Gróf, Efficiency of evacuated tube solar collector using WO3/
Water nanofluid, Renew. Energy 134 (Apr. 2019) 453–460.
luation parameter for the solar collector. The analyses have shown [17] M.M. Tafarroj, R. Daneshazarian, A. Kasaeian, CFD modeling and predicting the
that increasing the MF and decreasing the particle size can augment performance of direct absorption of nanofluids in trough collector, Appl. Therm.
the PEC. The same consequences were obtained for the efficiency of Eng. 148 (Feb. 2019) 256–269.
[18] Y. Xiong, et al., Performance enhancement of bromide salt by nano-particle dis-
the collector. The solar collector working with 4% Fe3O4-Therminol persion for high-temperature heat pipes in concentrated solar power plants, Appl.
66 under MF with the particle size of 10 nm indicates better per- Energy 237 (Mar. 2019) 171–179.
formance compared to the base fluid as well as CuO/Therminol 66. [19] X. Wei, Y. Yin, B. Qin, J. Ding, J. Lu, Thermal conductivity improvement of liquid
Nitrate and Carbonate salts doped with MgO particles, Energy Proc. 142 (Dec.
▪ Although CuO nanoparticles have a higher thermal conductivity
2017) 407–412.
compared to Fe3O4 particles, this study has shown that under [20] D. Shin, D. Banerjee, Enhanced specific heat of silica nanofluid, J. Heat Transfer 133
magnetic field Fe3O4 performs better than CuO for improving the (2) (Feb. 2011) 024501.
collector efficiency. Moreover, PTC with finned tube illustrates [21] M. Malekan, A. Khosravi, Investigation of convective heat transfer of ferrofluid
using CFD simulation and adaptive neuro-fuzzy inference system optimized with
better proficiency in terms of enhancing heat transfer characteristics particle swarm optimization algorithm, Powder Technol. 333 (2018) 364–376.
from tube to the working fluids. Finally, PTCs with smooth and [22] M. Malekan, A. Khosravi, X. Zhao, The influence of magnetic field on heat transfer
finned tubes and Fe3O4/Therminol 66 with 4% volume fraction as of magnetic nanofluid in a double pipe heat exchanger proposed in a small-scale
CAES system, Appl. Therm. Eng. 146 (Jan. 2019) 146–159.
working fluid and in the presence of MF have almost the same le- [23] A. Jafari, T. Tynjälä, S.M. Mousavi, P. Sarkomaa, Simulation of heat transfer in a
velized cost of energy, while the other cases show higher values for ferrofluid using computational fluid dynamics technique, Int. J. Heat Fluid Flow 29
this parameter. The average of enhancement in the thermal effi- (4) (2008) 1197–1202.
[24] A. Khosravi, M. Malekan, Effect of magnetic field on heat transfer coefficient of
ciency for CuO/Therminol 66 is around 1.35% for B = 0 , and Fe3O4-water ferrofluid using artificial intelligence and CFD simulation, Eur. Phys.
around 4.0% for B = 250G , while an enhancement of 1.24% was J. Plus 134 (3) (2019) 88, https://doi.org/10.1140/epjp/i2019-12477-5.
reported by Korres et al. [14] for Syltherm 800/CuO for a PTSC with [25] H. Aminfar, M. Mohammadpourfard, F. Mohseni, Two-phase mixture model simu-
lation of the hydro-thermal behavior of an electrical conductive ferrofluid in the
B = 0 , i.e. no magnetic field.
presence of magnetic fields, J. Magn. Magn. Mater. 324 (5) (2012) 830–1542.
[26] H. Aminfar, M. Mohammadpourfard, Y.N. Kahnamouei, Numerical study of mag-
Acknowledgments netic field effects on the mixed convection of a magnetic nanofluid in a curved tube,
Int. J. Mech. Sci. 78 (2014) 81–90.
[27] E. Bellos, I. Daniil, C. Tzivanidis, Multiple cylindrical inserts for parabolic trough
The first author would like to acknowledge the financial support solar collector, Appl. Therm. Eng. 143 (July) (2018) 80–89.
from the Aarhus University, Denmark. The second and third authors [28] J. Muñoz, A. Abánades, Analysis of internal helically finned tubes for parabolic
acknowledge Aalto University, Finland, for the financial support of this trough design by CFD tools, Appl. Energy 88 (11) (Nov. 2011) 4139–4149.
[29] Ali Shakiba, Khodadad Vahedi, Numerical analysis of magnetic field effects on
project. hydro-thermal behavior of a magnetic nanofluid in a double pipe heat exchanger, J.
Magn. Magn. Mater. 402 (2016) 131–142.
References [30] E. Tzirtzilakis, N. Kafoussias, Three-dimensional magnetic fluid boundary layer flow
over a linearly stretching sheet, J. Heat Transf. 132 (1) (2010) 011702.
[31] S.V. Mousavi, M. Sheikholeslami, M. Gorji Bandpy, M. Barzegar Gerdroodbary, The
[1] Global Warming of 1.5 °C. [Online]. Available: < https://www.ipcc.ch/sr15/ Influence of magnetic field on heat transfer of magnetic nanofluid in a sinusoidal
> (accessed: 21-Jan-2019). double pipe heat exchanger, Chem. Eng. Res. Des. 113 (2016) 112–124.
[2] A. Khosravi, R.N.N. Koury, L. Machado, J.J.G. Pabon, Prediction of wind speed and [32] B.E. Launder, D.B. Spalding, The numerical computation of turbulent flows,
wind direction using artificial neural network, support vector regression and Comput. Methods Appl. Mech. Eng. 3 (2) (Mar. 1974) 269–289.
adaptive neuro-fuzzy inference system, Sustain. Energy Technol. Assessments 25 [33] S.E. Ghasemi, A.A. Ranjbar, Thermal performance analysis of solar parabolic trough
(Feb. 2018) 146–160. collector using nanofluid as working fluid: A CFD modelling study, J. Mol. Liq. 222
[3] A. Khosravi, S. Syri, M. El Haj Assad, M. Malekan, Thermodynamic and economic (2016) 159–166.
analysis of a hybrid ocean thermal energy conversion/photovoltaic system with [34] J.L. IV and J. L. V, A Heat Transfer Textbook, fourth ed., Philogiston Press, 2012.
hydrogen-based energy storage system, Energy 172 (2019) 304–319. [35] I. Leinhard, J. Leinhard, A Heat Tranfer Textbook, fourth ed., Philogiston Press, Saint
[4] A. Khosravi, R.O. Nunes, M.E.H. Assad, L. Machado, Comparison of artificial in- Petersburg, 2012.
telligence methods in estimation of daily global solar radiation, J. Clean. Prod. 194 [36] P. Bs, Heat transfer in turbulent pipe flow with variable physical properties, Harnett
(Sep. 2018) 342–358. JP, Ed. Adv. Heat Transf. 6 (1970) 504–564.
[5] F. Ullah, M. Kang, Impact of air flow rate on drying of apples and performance [37] A. Kasaeian, S. Daviran, R.D. Azarian, A. Rashidi, Performance evaluation and
assessment of parabolic trough solar collector, Appl. Therm. Eng. 127 (Dec. 2017) nanofluid using capability study of a solar parabolic trough collector, Energy
275–280. Convers. Manag. 89 (2015) 368–375.
[6] P.D. Malali, S.K. Chaturvedi, R. Agarwala, Effects of circumsolar radiation on the [38] M. Sheikholeslami, F. Bani Sheykholeslami, S. Khoshhal, H. Mola-Abasia,
optimal performance of a Stirling heat engine coupled with a parabolic dish solar D.D. Ganji, H.B. Rokni, Effect of magnetic field on Cu-water nanofluid heat transfer
collector, Appl. Therm. Eng. 159 (Aug. 2019) 113961. using GMDH-type neural network, Neural Comput. Appl. 25 (1) (2014) 171–178.
[7] B. Zou, Y. Jiang, Y. Yao, H. Yang, Optical performance of parabolic trough solar [39] R. Hamilton, O. Crosser, Thermal conductivity of heterogeneous two-component

19
M. Malekan, et al. Applied Thermal Engineering 163 (2019) 114435

systems, Ind. Eng. Chem. Fundam. 1 (3) (1962) 187–191. two-pass internal cooling channel, J. Heat Transfer 133 (5) (2011) 051701.
[40] E. Abu-Nada, Effects of variable viscosity and thermal conductivity of CuO-water [45] E. Bellos, Z. Said, C. Tzivanidis, The use of nanofluids in solar concentrating tech-
nanofluid on heat transfer enhancement in natural convection: mathematical model nologies: a comprehensive review, J. Clean. Prod. 196 (Sep. 2018) 84–99.
and simulation, J. Heat Transfer 132 (5) (2010) 052401. [46] W. Rl, Performance evaluation criteria for use of enhanced heat transfer surfaces in
[41] K. Ravi Kumar, K.S. Reddy, Thermal analysis of solar parabolic trough with porous heat exchanger design, Int. J. Heat Mass Transf. 24 (1981) 715–726.
disc receiver, Appl. Energy 86 (9) (2009) 1804–1812. [47] E. Bellos, C. Tzivanidis, I. Daniil, Thermal and exergetic evaluation of parabolic
[42] W. Siddique, I.V. Shevchuk, L. El-Gabry, N.B. Hushmandi, T.H. Fransson, On flow trough collectors with finned absorbers operating with air, Proc. Inst. Mech. Eng.
structure, heat transfer and pressure drop in varying aspect ratio two-pass rectan- Part A J. Power Energy 231 (7) (2017) 631–644.
gular channel with ribs at 45°, Heat Mass Transf. 49 (5) (May 2013) 679–694. [48] H. Olia, M. Torabi, M. Bahiraei, M.H. Ahmadi, M. Goodarzi, M.R. Safaei,
[43] W. Siddique, L. El-Gabry, I.V. Shevchuk, N.B. Hushmandi, T.H. Fransson, Flow Application of nanofluids in thermal performance enhancement of parabolic trough
structure, heat transfer and pressure drop in varying aspect ratio two-pass rectan- solar collector: state-of-the-art, Appl. Sci. 9 (3) (Jan. 2019) 463.
gular smooth channels, Heat Mass Transf. 48 (5) (May 2012) 735–748. [49] L. Salgado Conrado, A. Rodriguez-Pulido, G. Calderón, Thermal performance of
[44] I.V. Shevchuk, S.C. Jenkins, B. Weigand, J. von Wolfersdorf, S.O. Neumann, parabolic trough solar collectors, Renew. Sustain. Energy Rev. 67 (2017)
M. Schnieder, Validation and analysis of numerical results for a varying aspect ratio 1345–1359.

20