Applied Thermal Engineering 206 (2022) 118117

Contents lists available at ScienceDirect

Applied Thermal Engineering

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

Research Paper

Thermal, thermodynamic and exergoeconomic investigation of a parabolic

trough collector utilizing nanofluids
Behzad Vahedi a, Ehsan Golab b, Arsalan Nasiri Sadr c, Kambiz Vafai d, *
a
Department of Mechanical Engineering, Science and Research Branch, Islamic Azad University, Tehran, Iran
b
Department of Mechanical Engineering, Sharif University of Technology (SUT), Tehran, Iran
c
Department of Mechanical and Energy Engineering, Shahid Beheshti University (SBU), Tehran, Iran
d
Mechanical Engineering Department, University of California, Riverside, CA 92521, United States

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

Keywords: The exploitation of solar energy facilitates the renewable energy paradigm. In this regard, parabolic trough
Parabolic trough solar collector collectors (PTC) are considered as a useful set-up to absorb solar energy. Simultaneous study of thermal, ther­
CFD Simulation modynamic, and exergoeconomic performance of PTC systems paves the way for designers and manufacturers to
Nanofluid
not only have a better insight into understanding the underlying concepts about the operation of PTC systems but
Thermodynamic performance
Thermal performance
also to find the most effective and cost-effective circumstances. This study aims at analyzing a practical PTC
Exergoeconomic system by considering an evacuated absorber tube with glass cover, non-uniform heat flux, and taking into
account the convective and radiative heat losses. Obtained results demonstrate that employing the glass cover,
especially in the low Reynolds numbers regime could remarkably reduce the heat losses up to 22%. The present
analysis indicated that considering oil temperature-dependent properties reduced the friction factor around 94 %
in the low Reynolds numbers regime compared to that of the cases based on constant properties. In this work a
CFD code in the OpenFOAM software was developed to simulate both laminar and turbulent regimes with Lien
cubic k − ∊ model (non-linear eddy viscosity model) by adding three types of nanoparticles (Al2 O3 , Cu,
andSWCNT) individually into the synthetic oil. Moreover, the Buongiorno’s model (BGM) which considers
Brownian, thermophoresis, and turbulence diffusion phenomena was implemented to model the nanoparticles
and base fluid interactions. It was found that adding nanoparticles and increasing the Reynolds number have no
substantial impact on thermal efficiency of the system, whereas an optimum Reynolds number was found for
exergy efficiency and net profit per unit transferred heat load (ηp ) of the system. Although injecting SWCNT into
pure oil led to augmentation of average convective heat transfer coefficient, PEC, and exergy efficiency, it was
found not to be cost-effective. It was established that the obtained results by homogenous model (single-phase
model) and BGM were roughly the same, with relative difference of less than 3.1 and 2.4% for the average
Nusselt number and friction factor, respectively.

desalination [2], heat pipes [3], vapor absorption refrigeration systems

1. Introduction [4], steam reforming for hydrogen production [5], solar air receiver [6]
or residential solar water heater [7]. In indirect systems, the heat can be
Significant adverse effects of the climate change due to global received by working fluids and stored by phase change materials for a
warming, the depletion of fossil-based energy resources, and ever- thermal energy storage tank [8] or solar ponds [9]. The absorbed heat
increasing emission of greenhouse gases and other hazardous sub­ can be utilized for the cited direct systems and converted to electricity
stances have raised global concerns and encouraged researchers and power by thermal power plants (See Fig. 1). One of the main indirect
scientists to find ways to find better solutions [1]. Clean, sustainable, systems to receive solar heat is parabolic trough collectors (PTCs). PTCs
and cost-effective energy sources, especially solar energy, are the most are one of the most prevalent, commercialized, and mature technology
promising energy sources since they are abundantly available, can be among solar thermal collectors at high temperatures [10]. Nevertheless,
easily harnessed, and can generate electricity and heat energy. Solar there are some practical limitations such as probability of deformation
energy can be directly employed as a heat source for thermal and crack formation due to high temperature and pressure gradient in

* Corresponding author.
E-mail address: [email protected] (K. Vafai).

https://doi.org/10.1016/j.applthermaleng.2022.118117
Received 23 October 2021; Received in revised form 6 January 2022; Accepted 17 January 2022
Available online 22 January 2022
1359-4311/© 2022 Elsevier Ltd. All rights reserved.
B. Vahedi et al. Applied Thermal Engineering 206 (2022) 118117

Nomenclature Greek letters

α Absorptivity
A Area, m2 δ Shape factor of SWCNT
Cp Specific heat, J kg− 1 K− 1 η Efficiency
C Concentration ratio ηp Net profit per unit transferred heat load
d Diameter, m ηc Total cost per unit transferred heat load
dnp Diameter of nanoparticle, nm ηel Electrical efficiency, 32.7%
dSWCNT Diameter of SWCNT, nm θinc Incidence angle, degrees
DB Molecular Brownian diffusion coefficient, kg m− 1 s− 1
θ Circumferential angle, degrees
DB,t Turbulent diffusion coefficient, m2 s− 1 λ Shape factor
DT Thermophoresis diffusion coefficient, kg m− 1 s− 1 μ Molecular dynamic viscosity, kg m− 1 s− 1
H Enthalpy, kg m2 s− 2 μt Turbulent dynamic viscosity, Pa s
h Convective heat transfer coefficient, W m− 2 K− 1 ν Molecular kinematic viscosity, m2 s− 1
Is Direct normal irradiation, W m− 2 ρ Density, kg m− 3
J Mass flux vector, kg m− 2 s− 1 σk Prandtl number for turbulent kinetic energy
k Thermal conductivity, W m− 1 K− 1 σ∊ Prandtl number for dissipation rate
keff Effective thermal conductivity, W m− 1 K− 1 τ Transmissivity
kB Boltzmann’s constant, (1.38064852 × 10-23) J K− 1 φ Volume fraction
L Length of the tube, m
LSWCNT Length of SWCNT, μm Subscripts
n Unit vector ave Average
P Pressure, Pa abs Absorber or absence
ΔP Pressure drop, Pa ap Aperture
PEC Performance evaluation criteria amb Ambient
Pp Pumping power, W b Bulk
Cons Constant
Q̇abs Absorbed total heat, W
ex Exergy
Q̇loss Heat loss, W f Base fluid
Q̇out Useful heat, W fr Fluid friction
Q̇s Input energy, W g Glass cover
Q̇t− g Radiative heat transfer between glass and tube, W ht Heat transfer
q’’ Heat flux, W m− 2 in Inlet
s Entropy generation rate per unit length, W m− 1
K− 1 Ini Initial
stot Total entropy generation, W m− 3 K− 1 K Kinetic dissipation
T Temperature, K lam laminar
Ta Ambient temperature, K loc Local
Tsun Sun temperature, K nf Nanofluid
Tg Average glass cover temperature, K np Nanoparticle
out Outlet
Tt Average tube temperature, K
op Optical
V Velocity vector, m s− 1
ov Overall
Va Ambient air velocity, m s− 1
p Pump
χ̇ d Exergy destruction, W
Prs Presence
χ̇ u Useful exergy, W
PTC Parabolic trough collector
r, θ, z Cylindrical coordinate, m
s sun
Dimensionless number t tube
Be Bejan number th Thermal
Le Lewis number tot Total
Nu Nusselt number tur Turbulence
NBT Ratio of Brownian to thermophoretic diffusivity u Useful
Pe Peclet number w Wall
Pr Prandtl number var Variable
Prt Turbulent Prandtl number
Superscripts
Re Reynolds number
− Time-averaged variable
Ra Rayleigh number
’ Fluctuating variable
Sc Schmidt number
Sct Turbulent Schmidt number

the absorber tube, and thermal degradation of the working fluid [11]. several working fluids, such as air, water, molten salt [13], and syn­
Moreover, economical limitations include the need for pumping power thesized oils [14], can be exploited in the PTC systems. The combination
to provide high flow rate, high cost of evacuated absorber tube, need for of nanotechnology and solar collectors culminates in a substantial heat
a durable absorber tube to resist high pressure and temperature, and transfer rate and thermal efficiency augmentation.
providing solar tracking system [12]. Depending on the application, There are several techniques to improve the thermal efficiency of the

2
B. Vahedi et al. Applied Thermal Engineering 206 (2022) 118117

solar collector’s absorber tube and especially of PTCs, for instance the CuO-H2O nanofluid on a direct absorption PTCs’ efficiency. Their
injecting solid fine particles to the base fluid, disturbing the fluid flow by results showed that the nanoparticles concentration and the nanofluid
baffles, increasing the surface of heat transfer by utilizing fins, embed­ flow rate profoundly impact the thermal efficiency. However, some
ding a screw shaft or twisted tape as a turbulator [15], and inserting studies [26] yielded no significant results about the thermal efficiency
porous structures [16]. Since nanofluids possess advantages and the augmentation of PTCs. Mwesigye et al. [27] presented a numerical
potential to harvest solar energy more efficiently than traditional fluids, investigation of SWCNT/Therminol VP-1 nanofluid in a PTC. Based on
they have gained popularity and applications in PTC systems [17]. their numerical findings, the convective heat transfer coefficient was
Nanofluids are dilute suspensions of solid nanoscale-sized particles, enhanced by 234%, and the entropy generation decline was roughly
including metallic such as silver (Ag), gold (Au), and Copper (Cu) [18], 70%. However, there was no substantial thermal efficiency
nonmetallic such as Al2 O3 , Fe3 O4 , andSWCNT [19], and hybrid nano­ improvement.
particles [20], dispersed in a base fluid. Adding nanoparticles into the To simulate the flow and heat transfer in nanofluids, there are
base fluid can alter the thermophysical properties of the working fluid generally two procedures: homogenous and non-homogenous such as
because of their exceptional properties which depends on concentration, the Lagrangian-Eulerian model (or discrete-phase model) [28], the
size, shape, and movement of particles [21]. According to Khanafer and Eulerian-Mixture model (EMM) [28], and Buongiorno’s model (BGM)
Vafai [22], adding 4% volumetric fraction of Al2 O3 solid particles to [29]. The main assumption of the homogenous or single-phase method
H2 O varied density, specific heat capacity, thermal expansion coeffi­ (SPM) is that the slip velocity is negligible between the particles and the
cient, dynamic viscosity, and thermal conductivity by + 10%, − 10.7%, base fluid. The discrete-phase model (DPM) proposes a continuous and
− 16.2%, +80%, and + 16% at room temperature, respectively. Even discrete phase, which Eulerian and Lagrangian approaches are deployed
though nanofluids boost the heat transfer rate, they have some disad­ to solve the mentioned phases, respectively. Albojamal and Vafai [30]
vantages, such as high cost of production, sedimentation, clogging of implemented the DPM procedure in the forced convection study by
particles, agglomeration, erosion, corrosion to the heat transfer system’s considering Van Der Waals force between particles and wall surface,
components [23]. Olia et al. [24] presented a comprehensive literature drag, gravitational, lift, thermophoresis, and Brownian forces. They
review regarding the effects of various factors, such as nanoparticles and concluded that the most dominant mechanisms were Brownian, ther­
base liquid type, nanoparticle size, and particle concentration on PTCs’ mophoretic, and lift forces, respectively. The EMM computes mo­
performance. They declared that nanofluids augmented the heat transfer mentum and energy equations for mixture phase coupled with
rate and exergy efficiency of the system and diminished the entropy concentration equation for dispersed particles and mass conservation for
generation of the system. Menbari et al. [25] carried out combined mixture fluid [31]. Despite the fact that the governing equations and the
modeling and experimental examination to scrutinize the influences of solution process for the BGM approach is similar to the EMM approach,

Fig. 1. Direct and indirect use of absorbed heat by parabolic trough collector system.

3
B. Vahedi et al. Applied Thermal Engineering 206 (2022) 118117

Brownian movement and thermophoresis are the most predominant reflector and solar dish concentrators are utilized [38]. A parabolic
mechanisms in comparison to the forces as mentioned earlier with trough collector (PTC), made of a parabolic mirror and a receiver tube
respect to the nanoscale-sized particles [30,32]. Since nanofluids are mounted at the focal axis of parabola, was employed (see Fig. 2). The
inherently two-phase, the two-phase approach gives more precise results parabolic mirror is commonly coated with polished aluminum that has a
than the single-phase approach. Nonetheless, the single-phase simula­ high reflectivity. Designing a suitable receiver plays a pivotal role to
tion requires less computational cost. A comparison between the nu­ reach the optimum performance of a solar collector system. The receiver
merical results of homogenous and non-homogenous models and tube comprises two parts, a Pyrex glass cover, and a stainless-steel tube.
experimental data was carried out for forced convection of Al2 O3− H2 O The reason for using a glass cover is to capture the solar irradiance be­
by Albojamal and Vafai [33]. Their results elucidated that the obtained tween the absorber tube and the glass cover. Another reason to install a
average convective heat transfer coefficient (hAve ) of DPM and SPM were glass cover is to prevent the absorber tube from direct contact with the
almost the same, with a difference of less than 1% with the experimental surroundings to reduce the thermal losses. To minimize the thermal
measurements. However, the obtained hAve for the EMM differed losses due to the convective heat transfer between the absorber tube and
23.5–60.1% from the experimental outcomes. Accordingly, the EMM the glass cover, the annulus pressure was set to be very low of the order
approach is associated with some fundamental defects to model nano­ of 0.0103 Pa [39]. The absorber tube surface was coated with cermet to
fluid flow and its distribution. The DPM and SPM approaches yielded elevate its absorptivity of the penetrated solar radiation from the glass
more precise results in comparison to the EMM. Buongiorno [29] pro­ cover, as well as to reduce its emissivity at high temperatures [40]. The
posed a non-homogenous equilibrium model by taking seven slip optical analysis of a PTC deals with some geometric parameters to reach
mechanisms into account. Based on Buongiorno’s observation, the dif­ higher optical performance. To obtain optimum interception of the
ferences between SPM and BGM results are insignificant for volume incoming solar beam, it is better to calculate the parabola focal distance.
fraction of less than 5%. The BGM has been widely applied to many CFD In focusing collectors, another important factor to fabricate a PTC sys­
analyses to study the natural, mixed, and forced convection of nano­ tem is the concentration ratio (C), which is the ratio of the area of
fluids for different applications. Sayyar and Saghafian [34] implemented aperture to the area of the absorber tube. The diameter of the absorber
the BGM model to analyze natural and forced convection inside a ver­ tube was specified by formula number 5 in Table 1 based on the rim
tical pipe filled with nanofluids. They concluded that the results of NuAve radius (rr ) and half acceptance angle (θm ) [41]. Finally, it should be
and Pressure drop corresponding to BGM and SPM were almost identical noted that the thickness of the pipe and glass were not considered. In
with a maximum 5% difference. Table 1, the row numbers 1–5 were used to describe the design formula.
Exergoeconomic is defined as a branch of engineering that considers Also, Table 2 provides the values of the design parameters.
exergy and economic principles to analyze the cost-effectiveness of a
system. Exergy analysis is based on the first and second laws of ther­ 3. Mathematical formulation
modynamics studied for each system that deals with energy. One of the
main systems that deals with huge amount of energy is thermal systems 3.1. Dimensional form of governing equations
where high exergy destructions occur in their components [35]. In this
regards, thermal collectors exploit free energy from solar irradiance, The heat transfer fluid (HTF) flowing through a horizontal circular
while they need an initial investment to be established. Wu et al. [36] absorber tube was considered steady, incompressible, Newtonian, and
studied the exergoeconomic of a heat exchanger system by considering three-dimensional. The absorbed solar thermal energy was transferred
exergy destructions due to heat transfer and fluid friction irrever­ to HTF through conductive and convective heat transfer mechanisms.
sibilties. Their Findings indicated that the inlet velocity and working Since the dominant mechanism of heat transfer was forced convection,
temperature have important impacts on exergoeconomic analysis. and the Richardson number was much less than unity, the effect of
Makkeh et al. [37] thermodynamically investigated the exergoeconomic buoyancy was considered to be negligible [42]. In light of the above
of an organic Rankine cycle (ORC) which the PTC was used as a heat assumptions, the governing equations including continuity, momentum,
source. Their results demonstrated that the exergy destruction of their energy, and nanoparticles distribution for both laminar and turbulent
PTC system was of the order of 39.5%. Thus, both numerical and ther­ regimes were employed, which can be represented as [43]:
modynamic studies help designers to accurately calculate the exergy ( )
∇. ρnf V = 0 (1)
destruction of PTC systems.
In the last few years, although several researchers have simulated a ( )
variety of HTF flow inside absorber tube of PTCs, they have not ∇. ρnf VV = ∇.(σ + R) (2)
comprehensively analyzed various aspects of a practical PTC system. So, The Navier Stokes equations are averaged by time resulting in the
there is a scientific gap in this field of study and the present study bridges Reynolds-averaged Navier–Stokes (RANS) which appears through the
the gap by studying various aspects of a practical PTC system including Reynolds stress term (R). To compute turbulence viscosity, the non-
its design, HTF flow simulations, first and second laws of thermody­ linear eddy-viscosity model based on Lien cubic k − ∊ model was
namics, and exergoeconomic parameters. Due to the non-uniformity of implemented [44]. In the k − ∊ model the Reynolds stress has a linear
solar heat flux on the surface of the absorber tube and heat losses, it is relationship with the deformation rate, whereas in the Lien cubic k − ∊
better to consider them to model practical PTC systems. Three different model the Reynolds stress has a non-linear and linear relation with
practical nanoparticles such as Al2 O3 , Cu, andSWCNT are added to the deformation and rotation, respectively. By considering the aforemen­
synthetic oil which the oil properties depends on temperature. Unlike tioned hypotheses, two PDE s were solved for Lien cubic k − ∊ model that
the previous studies, laminar and turbulence regimes based on Lien are called turbulent kinetic energy (TKE) and turbulent dissipation rate
cubic k − ∊ model are implemented in the present simulation to find the (TDR) as shown below [44]:
optimum Reynolds number. In addition, a well-known two-phase ( ) ( )
mathematical model called BGM is applied to compare the BGM and ∇. ρnf Vk = ∇. ρnf Dk ∇k + Gk − ρnf ∊ (3)
SPM. Finally, the exergy destructions are obtained by CFD procedure to
perform the exergoeconomic assessment. ( ∊ ∊2
∇. ρnf V∊) = ∇.(ρnf D∊ ∇∊) + C∊1 Gk − C∊2 ρnf (4)
k k
2. Designing a physical model Calculating the temperature distribution of HTF needs to solve the
energy equation (summation of enthalpy and kinetic), which the left
In order to harness the solar energy more efficiently, solar thermal side in the Eq. (5) denotes the convective and kinetic terms due to
collectors such as parabolic trough, compound parabolic, linear Fresnel

4
B. Vahedi et al. Applied Thermal Engineering 206 (2022) 118117

Fig. 2. Illustration of the PTC in present study.

respectively. The effective Brownian diffusion (DB,eff ) contains molecular

Table 1
Brownian (DB ) and turbulent diffusion (DB,t ). The nanoparticle distri­
Design parameters values for the present study [41].
bution can be described as [46]:
Num. Definition Formula
1 ( )
1 Equation for a parabola (y) y2 = 4fx V.∇(φ) = − ∇. Jnp (6)
ρnp
2 Parabola focal distance (f) Wa
4tan(θr )
3 Concentration ratio (C) Wa ∇T
πdt Jnp = − ρnp DB,eff ∇φ − ρnp DT (7)
4 Rim radius (rr ) 2f T
1 + cos(θr )
5 Diameter of the absorber tube (dt ) 2rr sin(θm ) In the above equations, V is the time-averaged velocity vector, T is
6 Area of the aperture (Aap ) L × Wa the time-averaged temperature,H is the time-averaged enthalpy, J and φ
denote the time-averaged mass flux and nanoparticles volume fraction,
respectively. Also, the subscripts np and nf denote nanoparticle and
nanofluid, respectively. Moreover, ρ, CP , and ∊ represent density, spe­
Table 2
The values of designing parameters for present study. cific heat capacity, and dissipation rate of turbulent kinetic energy. The
essential relations to complete the defined PDEs (1–7) are listed in
Parameters Value
Table 3. Furthermore, Table 4 represents the used constant values to
Diameter of the absorber tube (dt ) 0.09 m execute turbulence model.
Diameter of the glass cover (dg ) 0.135 m
Length of the receiver (L) 1m
Width of the aperture (Wa ) 5.64 m 3.2. Thermophysical properties
Area of the aperture (Aap ) 5.64 m2
Parabola focal distance (f) 0.81 m Within the solar thermal collector, the absorber tube experiences an
Concentration ratio (C) 20.03 intense solar heat flux, which leads to a large increase in the surface
Rim angle (θr ) 114
◦
temperature of the absorber tube. As such, the main reason for
Rim radius (rr ) 3.24 m employing thermal oils as the base fluid in solar thermal collectors is the
Half acceptance angle (θm ) 0.79
◦
higher boiling temperature in comparison with other conventional
Incidence angle (θinc ) 0
◦ fluids. In the present investigation, a synthetic oil was used and its
properties in Table 5 were formulated for a temperature range of
233–677 K [47].
movement of fluid, respectively. So, the energy equation can be written In the present work three types of nanoparticles including aluminum
as [45]: oxide, copper, and single-wall carbon nanotube (SWCNT) were indi­
( ) ( ) vidually added to pure oil. The main purpose of using these nano­
∇. ρnf VH + ∇.(VK) = ∇. keff ∇T + ΦD − CP,np Jnp .∇T (5) particles were their wide application and remarkable different
On the right side the conductive term is due to molecular and tur­ properties, which are suitable to observe their impact on heat transfer.
bulence diffusions, the heat generation term is due to mechanical The Einstein-stokes and the Hamilton-crosser model [48] were used for
dissipation, and the mass flux term is due to the movement of the dynamic viscosity and thermal conductivity of the nanofluid, respec­
nanoparticles. In the nanoparticles distribution equation (NDE), it is tively. It is worth noting that λ is equal to 3 for Al2 O3 and Cu spherical
assumed that the volume fraction of the nanofluid is less than 5%, and nanoparticles. Furthermore, precise nanofluid’s thermal conductivity
that the nanoparticles and the base fluid are locally in thermal equilib­ and dynamic viscosity (δ = 80) models were derived for spherical
rium. The NDE consists of a convective term on the left side, and the SWCNT with 10 nm diameter (dSWCNT ) and 5 μm length (LSWCNT ). Ac­
Brownian motion term due to volume fraction gradient and thermo­ cording to formula number 4 in Table 6, k11 c and k33 c represent the
phoresis phenomena due to temperature gradient in the right side, thermal conductivities along transverse and longitudinal axes of a
composite unit cell [49].

5
B. Vahedi et al. Applied Thermal Engineering 206 (2022) 118117

Table 3 3.3. Boundary conditions

The essential relations to complete the cited PDEs [29,44].
Num. Definition Relation The mentioned PDEs were solved by means of the following bound­
ary conditions. Generally, an entrance length of hydrodynamically fully
1 Field variable (X) X(x) + X (x, t) ’
developed flow (Le ) is defined in internal flows, which shows the
2 Mechanical stress tensor (σ) τ − PI convergence length of velocity boundary layers. This length can be
3 Linear deformation rate (γ̇) (∇V) + (∇V)T calculated for both laminar and turbulent regimes as follows:
{
4 Vorticity tensor (Ω) (∇V) − (∇V)T Le 0.05Reinlaminar
= (8)
dt 1.36Re1/4 inturbulance
5 Shear stress (τ) 2
μnf γ̇ − μ ∇.Vδij
3 nf
In the present designed PTC system, the Le for the laminar (Re =
6 Reynolds stress tensor (R) − ρnf V’i V’j
2000) and turbulent (Re = 5 × 104 ) regimes equal to 9 and 1.83 m,
7 Turbulence kinetic energy (k) 1( 2 ) respectively. Due to a distance between the pump and the tube’s inlet
u’ +v’2 +w’2
2
and the obtained values for Le , it can be said that the flow was consid­
8 Effective diffusivity for TKE (Dk ) νt
νnf +
σk ered fully developed in the entrance of the absorber tube. In order to
9 Production rate of TKE (Gk ) R : ∇V make the CFD simulation more accurate with the practical system, it is
10 Effective diffusivity for TDR (D∊ ) νnf +
νt logical to choose a suitable velocity boundary condition at the tube’s
σ∊ inlet and outlet. So, the parabolic and 15 power law profiles of the velocity
11 Turbulent dynamic viscosity (μt ) k2
ρnf Cμ should be implemented for the laminar and turbulent regimes at the
∊
12 Effective thermal conductivity (keff ) knf + kt absorber tube’s inlet, respectively [51]. According to previous argu­
ments, the HTF flow in tube’s outlet was completely fully developed. In
13 Turbulent thermal conductivity (kt ) CP,nf
μt
Prt conclusion, employing zero velocity gradient is correct for the velocity
14 Enthalpy (H) CP,nf T boundary condition at the outlet.
15 Kinetic energy (K) 1 The inlet temperature (T in ) and initial volumetric fraction of nano­
ρ |V|2
2 nf particles (φIni ) at the absorber tube inlet was constant and uniform;
16 Heat generation due to mechanical dissipation (ΦD ) ∇.((τ + R).V )
pressure was set to be zero gradient. The required Neumann boundary
17 Effective Brownian diffusivity (DB,eff ) DB + DB,t
condition for the energy equation on the absorber tube wall is discussed
18 Turbulent diffusivity (DB,t ) 1 μt
ρnf Sct
in the upcoming section. Furthermore, the no-slip condition and nano­
19 Brownian diffusivity (DB ) kB T
particles zero mass flux [52] were taken into consideration at the
3πμf dnp absorber tube wall. Lastly, the zero gradient boundary conditions
20 Thermophoresis diffusivity (DT ) kf μf φ assumption due to fully developed flow were applied for the tempera­
0.26
2kf + knp ρf ture and volume fraction at the absorber tube outlet. Also, the pressure
was set to be zero. Complete boundary conditions used for these simu­
lations are summarized in Table 8.
In Table 8, V is the velocity vector, T the temperature, n a unit vector
In order to calculate nanofluid properties in Table 6, the constant
thermophysical values such as density (ρ), specific heat capacity (Cp ),
dynamic viscosity (μ), thermal conductivity (k), and diameter of each Table 6
nanoparticle (dnp ) were introduced, which in Table 7 shows them at T = Formulas to determine nanofluid properties [48,49].
296.15K [50]. The subscripts nf, np, andf denote nanofluid, nano­ Num. Nanofluid Type Formula
particle, and fluid, respectively. properties

1 ρnf All ρnf = (1 − φ)ρf + φρnp

Table 4 2 CP,nf All (1 − φ)(ρCp )f +φ(ρCP )np
CP,nf =
Empirical constants to compute the turbulence model. ρnf
3 μnf Al2 O3 , μnf = μf /(1 − φ)2.5
Constant parameter Value Constant parameter Value Cu

Prt 0.85 Cβ1 3 SWCNT μnf = μf (1 + δφ)

Sct 0.85 Cβ2 15 4 knf Al2 O3 , knf knp + (λ − 1)kf + (λ − 1)φ(knp − kf )
=
Cu kf knp + (λ − 1)kf − φ(knp − kf )
σk 1.0 Cβ3 − 19 ( )
SWCNT 3
( )
σ∊ 1.3 Cγ1 16 knf + φ βx + βy 2 k11 c − kf
= , βx = c , βy =
C∊1 1.44 Cγ2 16 kf 3 − φβx k11 + kf
C∊2 1.92 Cγ4 − 80 k33 c k np
− 1, k11 c = , k33 c =
kf 2knp ak
Cμ1 1.25 Cμ 0.09 1+
dSWCNT kf
Cμ2 0.9 Aν 0.0198 knp
,ak = Rk kf ,Rk = 8 × 10− 8 (m2 KW− 1 )
Cβ 1000 A∊ 0.00375 2knp ak
1+
LSWCNT kf

Table 5
The properties formula of synthetic oil in 233 ≤ T ≤ 677K [47].
Num. Thermophysical properties Equations

1 μf 9.88562 × 10 − 7.30924 × 10− 1 T +2.21917 × 10− 3 T2 − 3.42377 × 10− 6 T3 +2.66836 × 10− 9 T4 − 8.37194 × 10− 13 5
T (mPas)
( )
2 ρf 1.2691 × 103 − 1.52115T + 1.79133 × 10− 3 T2 − 1.67145 × 10− 6 T3 kg/m3
3 Cf 1.10787 + 1.70736 × 10− 3 T(kJ/kgK)
4 kf 1.90134 × 10− 1 − 1.88053 × 10− 4 T(W/mK)

6
B. Vahedi et al. Applied Thermal Engineering 206 (2022) 118117

Table 7
The constant thermophysical properties of nanoparticles and synthetic oil at T = 296.15K [50].
Properties Name
Al2 O3 Cu SWCNT Synthetic oil

( )
$ 150 400 30,000 -
Price
( ) Kg
kg 3970 8933 2600 930
ρ 3
(m )
W 40 401 6600 0.13
k
mK )
(
J 765 385 425 1613
Cp
kgK
dnp (nm) 20 50 10 −
( )
kg − − − 6.7 × 10− 3
μ
ms

Table 8
Governing boundary conditions for a realistic PTC system [51,52].
Inlet Wall Outlet
[ ( ( r )2 ) ]
Velocity Laminar V =0 ∇V.n = 0
V in = 2Vmean 1 − ̂z
R
Turbulent [ 1 ]
( r )5
V in = 1.32Vmean 1 − ̂z
R
Pressure ∇P.n = 0 ∇P.n = 0 P=0
⎛ ⎞
Temperature Tin = 296.15K ∇T.n = 0
1 ⎜ Q̇t− g ⎟
∇T.n = ⎝q’’abs − ⎠
keff At
Concentration φIni = 0.04 − DT ∇T ∇φ.n = 0
∇φ.n = .n
DB,eff T

normal to tube walls, and P and φ denote the pressure and the volume
fraction, respectively. The subscripts in and Ini denote the inlet and
initial.
The boundary condition for the absorber tube wall consists of non-
uniform heat flux (q’’abs ) and heat transfer between absorber tube and
glass cover (Q̇t− g ). First, the mirror concentrator was exposed to the solar
irradiance which reflects it by the concentrator to the glass cover. Part of
the incident solar radiation at the glass cover is absorbed (αg ), partly
transmitted (τg ), and partly reflected. The penetrated solar radiation
from the glass cover is partly absorbed by the absorber tube (αt ), and
partly reflected. In addition, the reflected solar radiation onto the
absorber tube was distributed non-uniformly as a function circumfer­
ential angle (θ) [41]. Jeter [53] studied the local concentration ratio
(LCR), which represents the ratio of the concentrated solar flux to the
incident solar radiation for different incidence angles. In the current Fig. 3. The comparison of the LCR on the absorber tube for θinc = 0 between
◦

simulation, a curve fitting method was applied to estimate the LCR. The the present study and Jeter [53] calculated by curve fitting method.
comparison between the present study and Jeter’s study for variation of
LCR around the absorber’s tube outer surface as a function of circum­ tween glass cover and the absorber tube (Q̇t− g ) (annulus) an accurate
ferential angle (θ) was illustrated in Fig. 3. It is worth mentioning that model was used to simulate the complicated mathematical radiative
the average LCR in the present study is equivalent to the concentration model. Previous studies [54] have demonstrated that the radiation
ratio (C). The required values and LCR function were listed in Table 9. shape factor (RSF) is considered as an appropriate choice to calculate the
In the present study, the thermal losses from the glass cover (Q̇loss,g ) radiative heat transfer between two surfaces. Thus, in the present study,
encompasses convective heat transfer between the glass cover and the the RSF procedure was implemented, but the average glass cover tem­
ambient air, and radiative heat transfer between the glass cover and the perature (T g ) was unknown. The energy conservation around the glass
sky. To make the present simulation more accurate with realistic con­ cover accounts for the thermal equilibrium between Q̇loss,g , Q̇t− g , and
ditions, the ambient air velocity (Va ) and temperature (Ta ) were
Q̇abs,g . Trial and error method can determine the Tg . Consequently, in the
considered to account for the convective heat transfer from the surface
current simulation, a function (fg ) with required parameters and for­
of the glass cover [13]. The average glass cover temperature (Tg ) was
also required to calculate the thermal losses. mulas (Table 10) were defined. T g was determined using the Bisection
In order to realistically account for the radiative heat transfer be­ method (see Fig. 4). Eventually, Q̇t− g for the boundary condition can be

7
B. Vahedi et al. Applied Thermal Engineering 206 (2022) 118117

Table 9
The required values and LCR function to obtain q’’ abs [41].
Definition Formula and value

Absorbed solar heat flux τg αt ηop Is LCR(θ)

by absorber tube wall
(q’’ abs )
Optical efficiency (ηop ) ωκk(θinc )
Incidence angle modifier 1
(cosθinc + 0.000884θinc − 0.00005369θinc 2 )
k(θinc ) cosθinc
Absorber tube absorptivity 0.93
(αt )
Glass cover transmissivity 0.97
(τg )
Direct normal irradiation 800W/m2
(Is )
Intercept factor (κ) 1
Reflectance of the 0.827
concentrator (ω)
LCR(θ) 0 ≤ θ ≤ 68 − 2.588×10− 7 θ5 + 4.175×10− 5 θ4 − 2.258×10− 3 θ3 +
◦

0.04643θ2 − 0.01496θ + 26.65

68 < θ ≤ 114 4.071×10− 4 θ3 − 0.1121θ2 + 9.099θ − 181.1
◦

114 < θ ≤ 180 − 2.157×10− 6 θ3 + 1.135×10− 3 θ2 − 0.1821θ + 9.668

◦

Table 10
The required parameters and formulas to calculate Q̇t− g [45,54].
Num. Definition Formula Fig. 4. The process of calculating T g based on Bisection method.
1 Glass cover absorbed total heat (Q̇abs,g ) Q̇abs,g = ηop Is Cαg Ag
2 Area of glass cover (Ag ) Ag = Lπdg
( ) Table 11
3 Heat loss of glass cover (Q̇loss,g ) Q̇loss,g = ha Ag Tg − Ta +
( ) The essential constant values for boundary condition section.
4
εg σsb Ag Tg − Tsky
4
Parameters Value
4 Convective Heat transfer coefficient of ha = Va 0.58 dg − 0.42

outer wall (ha ) Glass cover absorptivity (αg ) 0.02

5 Sky temperature (Tsky ) Tsky = 0.0522Ta 1.5 Air velocity (Va ) 7m/s

6 Heat transfer between absorber tube and

(
4 4
) Ambient temperature (Ta ) 286.15 K
σsb At Tt − Tg
glass cover due to radiation (Q̇t− g ) Q̇t− g = Sun temperature (Tsun ) 5770 K
1 1 − εg dt
+ The emittance of the absorber tube with cermet (εt ) 0.062 − 10− 7 (Tt − 273.15)2
εt εg dg
7 Area of the absorber tube (At ) At = Lπdt Glass cover emissivity (εg ) 0.01
∫ L ∫ 2π
8 Average temperature of tube surface (Tt ) Tt rt dθdz Stefan-Boltzmann constant (σsb )
(
W
)
Tt = ∫0 L ∫0 2π 5.67 × 10− 8
m2 K4
0 0 [ rt dθdz ]
9 Function of calculating Tg (fg )
fg = Q̇t− g − Q̇abs,g − Q̇loss,g

Table 12
Entropy generation and exergy destruction formulas due to fluid friction and
computed when T g is determined. Note that the required constant values
heat transfer [55-57].
for all of the abovementioned formulas given earlier in the boundary
Num. Definition Formula
condition’s section are classified in Table 11.
1 Entropy generation of molecular heat knf
(∇T : ∇T)
transfer (sht,lam ) T
2
3.4. Entropy generation and exergy
2 Entropy generation of turbulent heat keff 1
(∇T : ∇T)
transfer (sht,tur ) knf T2
Generally, entropy in thermodynamic is basically defined as the 3 Total entropy generation of heat transfer
∫ L ∫ rt ∫ 2π ( )
0 0 0 sht,lam +sht,tur rdθdrdz
measure of randomness and disorder in a system which is a function of (sht,tot )
state. On the other hand, the entropy generation is a function of the 4 Entropy generation of molecular fluid μnf
(∇V : ∇V)
process which is produced due to the existence of resistance in energy friction (sfr,lam ) T
transfer for instance thermal resistance (conductivity) in heat energy 5 Entropy generation of turbulence ρnf
ε
transfer and viscous force in kinetic energy transfer. The two main dissipation (sfr,tur ) T
∫ L ∫ rt ∫ 2π ( )
dominant mechanisms in entropy generation in heat and mass transfer 6 Total Entropy generation of fluid friction
0 0 0 sfr,lam +sfr,tur rdθdrdz
(sfr,tot )
process are heat transfer and fluid friction entropy generation [55]. The
7 Total entropy generation (stot ) sht,tot + sfr,tot
CFD accurately computes the entropy generations in comparison with
8 Bejan number (Be) sht,tot
the thermodynamic approach because of consideration of the entropy stot
generation at each point of the system. The molecular and turbulent 9 Exergy destruction due to heat transfer Ta sht,tot
entropy generation due to heat transfer and fluid friction in the CFD code (χ̇ d,ht )

was listed in Table 12 [56]. Exergy is the amount of useful work of a 10 Exergy destruction due to fluid friction Ta sfr,tot
(χ̇ d,fr )
system that can be achieved by equilibrium with its surrounding. Be­
11 Total exergy destruction (χ̇ d,tot )
sides, exergy destruction is available due to entropy generation in all
χ̇d,ht + χ̇d,fr

energy transfer process which Table 12 displays the exergy destruction

8
B. Vahedi et al. Applied Thermal Engineering 206 (2022) 118117

of fluid flow and heat transfer [57]. Table 14

The required knf 0 for each material to
calculate the average Nusselt number.
3.5. Thermal performance parameters of PTC systems
Type knf0

Many practical PTC systems deal with flow and heat transfer mech­ Oil 0.131
anisms which requires some vital parameters such as friction factor, Al2 O3 0.151
pump power, outlet bulk temperature, convective heat transfer coeffi­ Cu 0.152
cient, and Nusselt number. The pump power (Pp ) for an incompressible SWCNT 0.542

flow must not only overcome viscous force resistance (Ẇfr ), but it must
also provide the kinetic energy dissipation (ẆK ) [12]. If the correct same value. Eventually, the exergy efficiency expresses the ratio of
velocity boundary condition is implemented at the tube’s inlet, the ef­ output useful exergy (χ̇ u,PTC ) to input useful exergy (χ̇ u,tot ) [27]. To have a
fects of the kinetic energy dissipation on the pumping power could be
better insight into the main concepts of first and second laws of ther­
negligible. Moreover, the bulk temperate (Tb ) related to the ratio of
modynamic of a PTC system, a schematic energy and exergy diagrams
integral of enthalpy rate to the integral of the heat capacity rate. The
are depicted in Figs. 5 and 6. The performance and thermodynamic
performance evaluation criterion (PEC) [58] is defined to compare the
parameters were comprehensively arranged in Table 15.
heat transfer rate to friction relative to base condition (pure oil in Re =
1000) owing to inclusion of nanoparticles and Reynolds number
increment. Table 13 and 14 presents the mentioned parameters and 3.7. Exergoeconomic analysis of PTC systems
their formulas in integral form.
After design and analysis, looking into the economic aspects of a
system have a crucial role in optimization of designing and fabrication.
3.6. Thermodynamic analysis of PTC systems The exergoeconomic analysis is comprised of exergy analysis and eco­
nomic principles assessment to evaluate the tradeoff between the per­
Our CFD analysis provides numerous impacts such as the design, formance and economic parameters, as well as to produce a cost-
flow, material, and ambient parameters. On the other hand, our ther­ effective production. In this regard, in the current study, the exergy
modynamic analysis introduces some useful and practical definitions destruction of fluid friction and heat transfer were considered in exergy
such as thermal efficiency (ηth ), overall efficiency (ηov ), and exergy ef­ assessment. In other words, the initial investments for the pump,
ficiency (ηex ). Thermal efficiency refers to the ratio of benefit to cost nanoparticles, and PTC system including the parabolic mirror, evacu­
which is proposed here as the ratio of output’s received heat rate (Q̇out ) ated absorber tube, cermet coating, and connections were taken into
to summation of received heat rate for a PTC system (Q̇s ) and pump account in economic evaluation [60].
power (Pp ). The, overall efficiency (ηov ) indicates the ratio of useful One of the important parameters in exergoeconomic assessment is
energy (Eusf ) to input energy (Ein ) for the system, which is an important the net profit per unit of transferred heat load (ηP ), which explains the
parameter for powerplants. Another concept for sustainability of ther­ ratio of annual income (NPV) due to useful heat load by separating a
modynamic systems is exergy efficiency (ηex ), which expresses the en­ portion of the initial investment to the total received heat [36]. It is
ergy losses due to thermal (Q̇loss,g ), optical (Q̇loss,opt ) and dissipations worth mentioning that the designed PTC system is cost-effective, when
the ηP is greater than zero. Another pivotal parameter in exer­
because of heat transfer rate and fluid friction [59]. Lack of energy losses
goeconomic is total cost per unit transferred heat load (ηC ) which means
and dissipation causes the thermal and exergy efficiency to have the
the ratio of annual cost (TCV) due to exergy destructions and initial
investment to the total received heat. It can be mentioned that the
Table 13
designed PTC system would be less costly when the ηC tends to zero. All
Integral form of performance parameters formulas [58].
the required formulas and constant values were arranged in Tables 16
Num. Definition Formula and 17 to calculate the defined parameters in our exergoeconomic
1 Pumping power (Pp ) Ẇfr + ẆK assessment.
2 Pumping power due to friction (Ẇfr ) out
∫ [ ∫ rt ∫ 2π ]
0 0 PV.nrdθdr
in 3.8. Justification of using glass cover and temperature-dependent
3 Pumping power due to kinetic dissipation out
∫ [∫ rt ∫ 2π ]
(ẆK ) 0 0 KV.nrdθdr properties
in

4 Pressure drop (ΔP) o∫ut[ ∫

rt ∫ 2π ]
ΔP = 0 0 Prdθdr As it was mentioned earlier, although adding a glass cover to the
in
5 Dimensionless friction factor (fr) ΔP/L absorber tube has a relatively high cost for the PTC system, it could
1
ρin dVin 2 notably mitigate the thermal losses from the absorber tube [61]. The aim
∫2 rt ∫ 2π
6 Bulk temperature (Tb )
0 0 ρnf CP,nf TV.nrdθdr of this section is to show the impact of the presence and absence of a
glass cover on the thermal loss values for both laminar and turbulence
∫ rt ∫ 2π
0 0 ρnf CP,nf V.nrdθdr
7 Output’s received heat rate (Q̇out = out
∫ [ ∫ rt ]
0
∫ 2π
0 ρnf CP,nf TV.nrdθdr regimes. To provide an explanation to compare the variation of the heat
ṁCP ΔT) in
loss (VARloss ) for presence and absence of a glass cover, a relation (Eq.
8 Received heat flux from the absorber tube keff ∇T.n
(9)) is proposed to calculate it for each Reynolds number, which is
wall (q˝w )
∫ L ∫ 2π depicted in Fig. 7.
9 Received heat rate from the absorber tube q˝w rt dθdz
0 0
wall (Q̇w ) Q̇out,prs − Q̇out,abs
10 Local convective heat transfer coefficient q˝w VARloss = × 100% (9)
Q̇out,abs
(hloc ) Tt − Tb
∫ L ∫ 2π
11 Average convective heat transfer
0 0 hloc rt dθdz As it inferred from the presented results, utilizing a glass cover for
coefficient (have ) ∫ L ∫ 2π
0 0 rt dθdz laminar regime (Re = 1000, 2000) and lower Reynolds numbers up to
12 Average Nusselt number (Nuave ) have d
(Re = 5000) could be beneficial. When the Re is low and the tempera­
knf0
13 Performance evaluation criteria (PEC) Nuave /Nuave,0 ture of the absorber tube is high, the heat transfer between absorber tube
( )
1 wall and surrounding is high as such, the glass cover significantly re­
f/f0 3 duces the thermal losses from absorber tube. On the other hand, VARloss

9
B. Vahedi et al. Applied Thermal Engineering 206 (2022) 118117

Fig. 5. Schematic energy diagram for a PTC system.

Fig. 6. Schematic Exergy diagram of PTC system to show different types of exergy losses and destructions.

tends to zero when the Reynolds number is around 20,000 and higher. According to Eq. (2), the friction factor is a function of the wall shear
Therefore employing a glass cover is not useful or effective for Reynolds stress. Moreover, at the same Reynolds number, the wall shear stress
number greater than 20000. This is due to the fact that for high Reynolds only depends on dynamic viscosity. Dynamic viscosity also depends on
numbers the temperature of the absorber tube wall was less than the temperature which is enhanced by temperature increment. As it can be
glass cover temperature (the glass cover was able to absorb the solar seen from Fig. 8(a), at low Reynolds numbers (laminar regime), the
irradiation proportional to αg ). difference between frvar and frcons was high since the wall temperature in
Previous studies [10] have shown that the oil properties depend on the low Re region was high and the dynamic viscosity had low values. As
the temperature for the examined PTC system which substantially af­ a result, frvar values were less than frcons . Furthermore, by increasing the
fects the results. So, the aim of this section is to investigate the tem­ Reynolds number, VARfr tends to zero. This is due to the fact that
perature dependency and independency of oil properties in the present increasing the Re declined the wall temperature and the variation of the
designed PTC system for a wide range of Reynolds numbers. Further­ dynamic viscosity relative to the incoming oil. Finally, it can be seen that
more, to provide a definition to compare these effects, the variation of the oil properties depending on temperature can be ignored at high
friction factor was presented as follow: Reynolds numbers.
frvar − frcons
VARfr = × 100% (10)
frcons
Moreover, VARfr was depicted in Fig. 8 for (1 × 103 ≤ Re ≤ 5 × 104 ).

10
B. Vahedi et al. Applied Thermal Engineering 206 (2022) 118117

Table 15 Table 17
Performance and thermodynamic parameters of the PTC system [27]. The required constant values for exergoeconomic formulas.
Num. Definition Formula Parameters Value

1 Thermal efficiency (ηth ) Benefit Q̇out Price per unit length of the tube (λl ) 170 $/m
= × 100%
Cost Q̇s + Pp Annual operation time (τ) 10512 × 103 s
2 Received heat rate to PTC system (Q̇s ) Is Aap
Unit cost of exergy associated with heat transfer (εq ) 3.4 × 10− 8
$/J
3 Output’s received or useful heat rate Q̇s − Q̇loss,tot + Pp
Weighing factor (ξ) 4
(Q̇out )
( ) The average lifespan of tube (n) 18 Years
4 Reflected heat loss of mirror (Q̇loss,opt ) 1 − ηopt Q̇s
Salvage value of tube (SV) 0
5 Total heat loss (Q̇loss,tot ) Q̇loss,opt + Q̇loss,g Interest rate (i) 0.15
6 Received heat rate to HTF from Q̇out − Pp orQ̇s − Q̇loss,tot Pump price Ap = 0.3$Bp = 10$
absorber tube wall (Q̇w )
7 Overall efficiency (ηov ) Usefulenergy Eu
= × 100%
Inputenergy Ein
8 Useful energy (Eu ) Q̇out − Pp /ηel
9 Input energy (Ein ) Q̇s
10 Exergy efficiency (ηex ) outputusefulexergy χ̇u,PTC
= ×
inputusefulexergy χ̇u,tot
100%
11 Output useful exergy of PTC system Q̇out − χ̇ d,tot
(χ̇ u,PTC )
[ ) ]
12 Useful received exergy from sun (χ̇ u,s ) (
4 Ta
) (
1 Ta 4
Q̇s 1 − −
3 Tsun 3 Tsun
13 The useful exergy of pump (χ̇ u,p ) Ta
Pp
T
14 The input useful exergy of the PTC χ̇u,s + χ̇u,p
system (χ̇ u,tot )

Table 16
The required formulas to calculate the exergoeconomic parameters [36].
Num. Definition Formula

1 Net profit value (NPV)

( ) Fig. 7. The variation of heat loss (VARloss ) for the presence and absence of a
τ εq χ̇ u,ht − εp χ̇d,fr − [Itot − SV(P/F, i, n) glass cover.
](A/P, i, n)
2 Useful exergy of heat transfer Q̇out − χ̇ d,ht
(χ̇ u,ht )
n
3 Worth factor (P/F, i, n) (1 + i)−
4 Capital recovery factor (A/P, i, i/[1 − (1 + i)− n ]
n)
5 Net profit per unit transferred NPV
heat load (ηp ) τQ̇s
( )
6 Total cost value (TCV) τεq ΔET +ξΔEp + [Itot − SV(P/F, i, n)
](A/P, i, n)
7 Total cost per unit transferred TCV
heat load (ηC ) τQ̇s
8 Initial investment of the PTC IPTC + Ip + Inp
system (Itot )
9 Initial investment of a PTC λl L
(IPTC )
10 Initial investment of pump (Ip ) Ap Pp
+ Bp
ηel
11 Initial investment of mnp Anp
nanoparticle (Inp )
12 Nanoparticle mass (mnp ) ρnp φIni Vnf
13 Nanofluid volume (Vnf ) Lπrt 2
14 Unit cost of exergy associated εp = εq ξ
with flow pressure drop (εp )

Fig. 8. The variation of friction factor to investigate the temperature de­

4. Numerical modeling procedure, grid independence, and code pendency and independency of oil properties.
validation
PLE algorithm) to handle the pressure–velocity coupling. Next, the
4.1. Numerical modeling approach SimpleFoam solver was developed by adding energy and nanoparticles
distribution equations, variable properties, and Bisection algorithm.
The SimpleFoam solver was chosen from Open Field Operation and Also, a non-linear eddy viscosity model, Lien cubic k − ∊, was employed
Manipulation (OpenFOAM) based on finite volume method (FVM) which to simulate turbulence regime which is more accurate than the other
implements Semi-Implicit Method for Pressure Linked Equations (SIM­ k − ∊ models because of consideration of the deformation rate and

11
B. Vahedi et al. Applied Thermal Engineering 206 (2022) 118117

rotation in the Reynolds stress term. The second-order accurate Gauss y+ = 1 and it can be obtained by formulas in Table 18 for fully devel­
linear corrected and the second-order accurate bounded Gauss upwind oped turbulence flow [62].
schemes were used to discretize the non-linear PDEs into linear algebraic In order to reduce the computational costs and increase the accuracy
equations for Laplacian and Divergence terms, respectively. In order to of the results in turbulence simulation, empirical correlations (wall
have convergence in an iterative method, residuals must be defined and functions) can be employed to estimate the various fields in the inner
determined. In fact, residual demonstrates the ratio of difference be­ region of the turbulence regime which is a bridge between the wall and
tween current and previous iterations to current iteration. Moreover, the the outer region [63]. It is worth mentioning that the hc can be estimated
under-relaxation factors are defined in CFD simulations to have stability for y+ = 10 by implementing the wall functions. In the current study, the
in order to reduce the effects of the prior iteration on the current iter­
ation. Eventually, a flowchart was graphically depicted in Fig. 9 to
Table 18
summarize the implemented algorithm including convergence criteria, Required formula to calculate the height of the nearest cell to wall in fully
residuals, and under-relaxation factor. developed turbulence flow [62].
Num. Definitions Formula
4.2. Grid independence assessment
1 Height of the nearest cell to y+ μf
wall (hc ) ρf uτ
There are two main regions in the turbulence regime, called outer 2 Shear velocity (uτ )
√̅̅̅̅̅
τw
region and inner region including viscous sublayer, buffer layer, and log ρf
3 Wall shear stress (τw ) frm
layer. There is a non-dimensional number in characteristic of turbulence ρ Vin 2
8 f
modeling, well-known Y plus (y+ ), which indicates the aforementioned 4 Friction factor of fully [0.79lnRe − 1.64]− 2 for3×103 ≤ Re ≤ 5 ×
regions. In the CFD simulation of turbulence flow, the mesh must be developed turbulence flow 106
finer in the proximity of the wall region to cover the viscous sublayer (frm )

which the height of nearest cell to the wall must be equal and less than

Fig. 9. Flowchart of the present developed code to correct the V, P, T, andφ iteratively.

12
B. Vahedi et al. Applied Thermal Engineering 206 (2022) 118117

hc must be smaller than 3.5 × 10− 4 .

For each CFD simulation and FVM, the computational domain must
be divided into small cells, in which structured mesh were generated for
the absorber tube in the present study. Moreover, the meshes were
denser in the radius direction, whereas a uniform mesh was utilized in
longitudinal direction. Also, the maximum aspect ratio of the cells are
less than 25. Fig. 10(a) illustrates the generated mesh for the current
modeling.
In order to provide precise numerical results and mesh indepen­
dence, five grid setups were tested to check the mesh independence for
results for the pure oil and Re = 5 × 104 . Fig. 10(b) illustrates Nuave
relative to the grid numbers to investigate mesh sensitivity. It can be
observed that the relative difference percentage for the Nuave between
the reference (6125 × 70) and 4500 × 60 cells is fewer than other de­
viations. Thus, the grid with 4500 × 60 cells could provide reasonable
run time and adequate accuracy.
Fig. 11. Comparison of Nuave versus different Peclet number for φ = 0.03 be­
tween the numerical results and the experimental results of Zeinali et al. [64].
4.3. Validation

This section was devoted to assessing the validity of the current (Eq. (11)). In their investigation, the Nuave values were examined for
developed CFD code. The obtained data were comprehensively γ − Al2 O3 and water nanofluid flow in a circular pipe with 0.104mm
compared with pertinent experimental and numerical cases in the diameter and 4.8m length. Their correlation is valid under the condition
literature. of Re ranging from 1 × 104 to 1 × 105 , Pr from 6.54 to 12.33, and the
volume fraction from 0 to 0.03. For the present validation, similar
4.3.1. Laminar flow and two-phase model evaluation conditions were duplicated with an average uniform heat flux boundary
In the first case, the laminar flow (2000 ≤ Pe ≤ 7000) of Al2 O3 -H2 O condition of 99.5 Kw
m2
. Alumina nanoparticles with volume fraction of
nanofluid was investigated through a copper tube with 6mm diameter 2.78%, Pr = 12.33, knf = 0.745, CP,nf = 4084, ρnf = 985, μnf =
and 1m length with constant wall temperature [64]. According to Zeinali 4.6(mPa.s), and dnp = 13nm were selected. Also, all the nanofluid
et al. [64], the migration of nanoparticles due to temperature gradient properties were calculated at a constant temperature (T = 300K).
(Thermophoresis) and Brownian motion, especially near the tube wall
should be considered to have a better understanding about interpreta­ Nuave = 0.021Re0.8 Pr0.5 (11)
tion of heat transfer enhancement. Accordingly, in the present valida­ In Fig. 12, a comparison was performed between their measured
tion, the two-phase Buongiorno model was used to make the validation Nuave and those obtained in this study. As it can be inferred from Fig. 12,
process precise enough by considering the behavior and interactions there is only subtle differences (maximum 7.8%) between the present
between the nanoparticles and the base fluid. The two-phase Nuave results and the experimentally measured results.
against the nanofluid Peclet number for φ = 0.03 was compared with
their experimental data. As represented graphically in Fig. 11, an 4.3.3. Thermal efficiency evaluation
acceptable average relative error of 10% from the experimental results To validate the thermal performance of a practical PTC with our CFD
was observed which establishes the reliability of this method. code, a well-known experimental study by Dudley et al. [66] for a
SEGSLS − 2 collector with cermet coating and vacuum in the annulus
4.3.2. Turbulent flow evaluation was chosen. They proposed a correlation (Eq. (12)) based on the
In order to assess the Nuave under the condition of turbulence regime, regression of obtained data during various tests to predict the thermal
the present results were compared with Pak and Cho [65] correlation

Fig. 10. (a) The illustration of generated mesh (number of mesh in cross section × length direction) and (b) Mesh independence versus Nuave .

13
B. Vahedi et al. Applied Thermal Engineering 206 (2022) 118117

Rayleigh number (Ra) is demonstrated in Fig. 14, and the obtained

numerical data were compared with the results of Sheikhzadeh et al.
[67] and Ho et al. [68]. Hence, as can be inferred from the results, the
maximum difference percentages of Nuave between the present study,
and experimentally measured data and numerical results were 12% and
9%, respectively. Even though the BGM includes some major slip
mechanisms, considering all hydro-thermal behaviors of nanofluids is
implausible, especially at higher nanoparticle concentrations [69]
leading to some minor errors in the final results.
For the last case, the natural convection of alumina-water nanofluid
inside a square cavity was presented for Ra = 106 , φave = 0.02, θ = 0 ,
◦

5 6 4
2.62 × 10 ≤ Le ≤ 1.05 × 10 , Pr = 4.623, Sc = 3.55 × 10 , and NBT =
1.1. The obtained isotherms and volume fraction contours of the present
code were evaluated with the results of Yekani et al. [69]. As depicted in
Fig. 15, the present results were in good agreement.
Fig. 12. Comparison of Nuave versus different Reynolds number for φ = 0.0278
between the present numerical results and Pak and Cho’s correlation [65].
4.3.5. Error analysis
The differences between the numerical and experimental results and
efficiency, which is valid at the average direct normal irradiance (Is,ave = sources of errors in this study are categorized as follows:
940 mw2 ). The thermal efficiency values were scrutinized for Syltherm- Experimental errors: Measurement inaccuracies in the laboratory
800 oil flow in a steel absorber tube with 7cm diameter, 11.5cm Pyrex investigation arising from systematic errors due to not calibrating the
glass diameter,4m length, Concentration ratio (C = 71), and Aap = measuring equipment, random errors due to variations of experimental
39.2m2 . The SEGSLS − 2 collector were examined for different operating conditions and the surrounding, and the observer’s errors. Moreover,
conditions including inlet temperatures, ranging from approximately repeatability and uncertainty of a measured value can create additional
100to360c , volumetric flow rate ranging from 0.795 to 0.938 Ls, and Is experimental deviations [66].
◦

from 807to982 mw2 . Lastly, For the current evaluation, the similar condi­ Numerical errors: Sources of numerical errors arise before and
tion based on the aforementioned operating conditions were employed. during computations from modeling approximations and assumptions
due to simplifications, discretization errors, mesh-related errors due to
ηth = 73.1 + 0.0012ΔT − 0.00085ΔT 2 (12) the quality and quantity of a generated mesh, round-off errors, and
And ΔT is defined as follows: truncation errors [70].
( ) Providing more accurate mathematical, physical models, and
ΔT =
Tin + Tout
− Tamb (13) empirical correlations helps researchers to reduce these errors. In fact,
2 utilizing well-known correlations to estimate nanofluid properties,
Based on the plotted data in Fig. 13, the average relative error be­ choosing an appropriate turbulence model, and considering dominant
tween the experimental and present numerical results is 4.4%, whereas interactions and behaviors of nanoparticles in a base fluid pave the way
the average estimated error due to systematic uncertainties of the for obtaining more realistic results rather than idealized results.
experimental test is 2%, which shows that the present CFD code is
consistent with real operating conditions of a PTC system. 5. Results and discussion

4.3.4. Buongiorno two-phase model, isotherms and nanoparticles The effects of the Reynolds number (Re), and different nanoparticles
distribution evaluation such as Al2 O3 , Cu, andSWCNT at the same volume fraction (φIni = 0.04)
The laminar free convection of aluminum oxide–water nanofluid was and synthetic oil are discussed for Tin = 296.15K on the performance
simulated inside a square enclosure for 2 ≤ ΔT ≤ 10, dP = 33nm, Pr = parameters including dimensionless friction factor (fr), average
4.623, the average volume fraction of 4%, 3.37 × 105 ≤ Ra ≤ 1.68 ×
106 , and the inclination angle of θ = 0 . The variation of the Nuave with
◦

Fig. 14. Comparison of the current study results for Nuave with those of
Sheikhzadeh et al. [67], and Ho et al. [68] for various Rayleigh numbers when
Fig. 13. Thermal efficiency comparison between present validation and dp = 33nm, 2 ≤ ΔT ≤ 10, 3.37 × 105 ≤ Ra ≤ 1.68 × 106 , Pr = 4.62, andφave =
experimentally measured data of Dudley et al. [66] for a SEGS LS-2 collector. 0.03.

14
B. Vahedi et al. Applied Thermal Engineering 206 (2022) 118117

Fig. 15. The comparison of the present validation case results for isotherms and nanoparticles distribution with the numerical results of Yekani et al. [69] (a)
Contours of nanoparticles volume fraction, and (b) Isotherms (K).

convective heat transfer coefficient (have ), PEC, and thermodynamic Nusselt number outweighs the reduction in the friction factor. However,
parameters such as thermal efficiency (ηth ), exergy efficiency (ηex ), adding SWCNT promotes the PEC, since the Nuave enhancement affects
overall efficiency (ηov ), total exergy destruction (χ̇ d,tot ), Bejan number the PEC more than the fr enhancement.
(Be), and exergoeconomic parameters including net profit per unit A summary of the results can be expressed as:
transferred heat load (ηp ), and total cost per unit transferred heat load
(ηC ). The results are acquired for a wide range of Reynolds numbers from • In the laminar regime, increasing the Reynolds number from 1000 to
103 to 5 × 104 . It is important to mention that adding 4% volume 2000, the fr is reduced by 56%, 52.8%, 54.4%, and 48.9% for oil,
fraction of Al2 O3 , Cu, andSWCNT to the oil altered the dynamic viscosity Al2 O3 , Cu, andSWCNT, respectively.
by + 10.7, +10.7, and + 320%, thermal conductivity by + 15.3%, • In the turbulence regime, increasing the Reynolds number from 5000
+16%, and + 313.7%, density by + 13.1%, +34.4%, and + 7.2%, and to 50,000 causes a reduction in the friction factor by 40.1%, 40.3%,
heat capacity (ρCp ) by + 4.1%, +5.2%, and − 1.1%, respectively. 38.5%, and 42.5% for oil, Al2 O3 , Cu, andSWCNT, respectively.
• In the turbulence regime and Re = 20000, adding 4% volume frac­
tion of Al2 O3 , Cu, andSWCNT lead to a variation in fr by − 7.6%,
5.1. Thermal performance analysis − 19.1%, and + 269.5%, respectively.
• Increasing the Reynolds number in the laminar regime from 1000 to
The dimensionless friction factor (fr) has a linear and inverse relation 2000 enhances the have by 26.6%, 19.9%, 19.9%, and 11% for oil,
with pressure drop (ΔP) and square of inlet velocity (Vin 2 ), respectively Al2 O3 , Cu, andSWCNT, respectively.
as can be seen in Table 13. The variation of the friction factor with • Increasing the Reynolds number in the turbulence regime from 5000
respect to the Reynolds number for both laminar and turbulence regimes to 50,000 increases the have by 5.06, 4.97, 4.95, and 4.34 times for
are plotted in Fig. 16(a) with the crosshatched zone representing the oil, Al2 O3 , Cu, andSWCNT, respectively.
transition region. As it can be inferred from the results in the laminar • In the laminar regime and Re = 2000, adding 4% volume fraction of
regime, fr is reduced by increasing the Reynolds number since the in­ Al2 O3 , Cu, andSWCNT causes an increase in have by 9.6%, 10.2%, and
ertial force is dominant in comparison to the pressure force. In addition, 184.7%, respectively.
injecting 4% volume fraction of the SWCNT increases the fr because of • In turbulence regime and Re = 20000, adding 4% volume fraction of
the dynamic viscosity increment, whereas the Al2 O3 andCu reduce it. In Al2 O3 , Cu, andSWCNT gives rise to variations in have by + 9.3%,
the turbulence regime, for all cases, fr is reduced by increasing the +8.7%, and + 480.3%, respectively.
Reynolds number from 5000 to 50,000 since the inertial forces are • Increasing the Reynolds number in the laminar regime from 1000 to
dominant. 2000 enhances the PEC by 26.2%, 19.8%, 20%, and 15.2% for pure
Fig. 16(b) demonstrates the variations of the average convective heat oil, Al2 O3 , Cu, andSWCNT, respectively.
transfer coefficient (have ). For all Reynolds numbers, inclusion of the • Increasing the Reynolds number in the turbulence regime from 5000
nanoparticles to the base fluid enhances the have due to the conduction to 50,000 augments the PEC by 4.97, 4.97, 4.95, and 4.39 times for
enhancement. oil, Al2 O3 , Cu, andSWCNT, respectively.
As expected, increasing the Reynolds number causes an increase in • In the laminar regime and Re = 2000, adding 4% volume fraction of
PEC as can be seen in Fig. 16(c). Also, employing Al2 O3 andCu nano­ Al2 O3 , Cu, andSWCNT leads to a drop in PEC by 5.1%, 4.4%, and
particles results in a reduction in PEC, since the reduction of average 10.5%, respectively.

15
B. Vahedi et al. Applied Thermal Engineering 206 (2022) 118117

Fig. 16. The variation of (a) dimensionless friction factor (fr), (b) average convective heat transfer coefficient (have ), and (c) performance evaluation criteria (PEC)
against Reynolds number for laminar and turbulence regimes and three various nanoparticles (Al2 O3 , Cu, andSWCNT) and synthetic oil.

• In the turbulence regime and Re = 20000, adding 4% volume frac­ result, the pump power plays an important role in thermal efficiency of
tion of Al2 O3 , Cu, andSWCNT contributes to variations in PEC by the system (See Table. 15 formula number 1). The thermal efficiency
− 5.2%, − 11.8%, and + 40.1%, respectively. results for both laminar and turbulence regimes are in the same range
with respect to pure oil at Re = 1000 due to low pumping power.
5.2. Thermodynamic performance analysis Fig. 17(b) represents the overall efficiency of the system. The value
of the thermal and overall efficiency for all cases are of the same order
The optical loss in this study is taken to be constant and equal to for 1000 ≤ Re ≤ 20000 while by increasing the Reynolds number to
17.3% of the total received heat rate to the PTC system (Q̇s ). Thus the 50,000 the overall efficiency decreases notably for all cases, especially
maximum value of the thermal efficiency is 82.7%. The other losses are for SWCNT.
related to the reflectivity and absorptivity of the glass and the absorber As was mentioned earlier, in addition to optical and thermal losses,
tube. The radiative thermal resistance (Rrad ) between the glass cover and the irreversibilities due to heat transfer and fluid friction affect the
the absorber tube can be estimated as: exergy efficiency. In the present study, the variation of Reynolds number
affects the temperature and velocity gradients. Furthermore, adding
1 σ sb At ( 2 )
2 ( )
nanoparticles influences the thermal conductivity, dynamic viscosity,
= Tt + Tg Tt + Tg (14)
Rrad ε1 + 1−ε εg ddt and the temperature gradient. The exergy destruction and Bejan number
t g g

data are presented in Fig. 17(c) and (d). Increasing the Reynolds number
The Rrad for the critical conditions in this study (Tt = 500KandTg = from 1000 to 20,000 can reduce the total exergy destruction. This is due
Ta ) is 1.56 × 104 W
K
which translates to a loss of 1.37 × 10− 2 (W). As such to the fact that the heat transfer irreversibility decreases due to the
the thermal loss is insignificant compared to the received solar heat. As a

16
B. Vahedi et al. Applied Thermal Engineering 206 (2022) 118117

Fig. 17. The variation of (a) thermal efficiency (ηth ), (b) overall efficiency (ηov ), (c) total exergy destruction (χ̇ d,tot ), (d) Bejan number, and (e) exergy efficiency (ηex )
against Reynolds number for both laminar and turbulence regimes, and three different nanoparticles (Al2 O3 , Cu, andSWCNT) and synthetic oil.

17
B. Vahedi et al. Applied Thermal Engineering 206 (2022) 118117

temperature gradient reduction, but increasing the Reynolds number • Adding SWCNT to pure oil leads to an increment in exergy efficiency
does not increase the fluid friction irreversibility since the Bejan number of the system from 74.5 to 79.9 % at an optimum Reynolds number of
is of the order of 1 (domination of heat transfer over the fluid friction (Re = 20000).
irreversibility). On the other hand, although increasing the Reynolds
number from 20,000 to 50,000 causes a reduction in the heat transfer
5.3. Exergoeconomic performance analysis
irreversibility (temperature difference reduction), it can elevate the total
exergy destruction. This is attributed to the fact that the fluid friction
The variation of net profit per unit transferred heat load (ηp ) with
irreversibility enhances due to the velocity gradient. As it is evident from
respect to Reynolds number for three different nanoparticles and pure
Bejan number results in Fig. 17(d), increasing the Reynolds number
oil is plotted in Fig. 18(a). Based on the obtained results, increasing the
from 20,000 to 50,000 reduces the Be as a result of fluid friction irre­
Reynolds number from 1000 to 20,000 increases the ηp due to the total
versibility increment and its dominance over the heat transfer irre­
versibility. It is worth mentioning that adding Al2 O3 andCu have no exergy destruction decrement for all nanoparticles and oil. However,
substantial impacts on irreversibilities, whereas SWCNT diminishes the increasing the Reynolds number to 50,000 reduces the ηp owing to the
total exergy destruction as can be seen in Fig. 17(c). This is due to a more total exergy destruction augmentation. Moreover, utilizing Al2 O3
substantial temperature gradient reduction relative to the dynamic nanoparticles have no impact on ηp , but the Cu nanoparticles decrease
viscosity and thermal conductivity enhancement. As seen in Fig. 17(c), the ηp due to the higher initial investment (cost of adding Cu nano­
because of heat transfer dominance over the fluid friction irreversibility particle). On the other hand, injecting SWCNT significantly lessens the
for Al2 O3 andCu in the range of Reynolds numbers of 1000 ≤ Re ≤ ηp because the effect of its staggering amount of initial cost outweighs its
20000 and SWCNT range of 1000 ≤ Re ≤ 10000, the Bejan number ability to reduce the total exergy destruction. Also, it can be mentioned
equals unit. However, increasing the Reynolds more than that reduces that the values of ηp for oil, Al2 O3 and Cu are greater than zero, but it is
the Bejan number since the fluid friction irreversibility outweighs the smaller than zero for SWCNT. In fact, the ηp for greater and smaller than
heat transfer irreversibility.
zero stands for the cost-effectiveness and cost-ineffectiveness of a
As it was mentioned before, the exergy efficiency would be equal to
system.
the thermal efficiency in the absence of exergy destruction due to heat
In order to represent total cost per unit transferred heat load (ηC ), the
transfer and fluid flow irreversibilities. In this regards, exergy efficiency
obtained data are graphically depicted in Fig. 18(b) which shows an
highly depends on the total exergy destruction of heat transfer and fluid
exact opposite trend to that of ηp . Employing Cu nanoparticle contrib­
flow. In comparison between Fig. 17(c) and (e), it can be seen that the
utes to a fall in ηC values since the effect of its initial investment on ηC is
trend of exergy efficiency is quite the opposite of exergy destruction. The
negligible in comparison to its ability to diminish the total exergy
following observations can be attained based on our results:
destruction, whereas SWCNT elevates the ηC owing to its huge cost of
utilization. The obtained results are represented in percentages as
• Inclusion of nanoparticles have no impact on the overall efficiency of
follows:
the system up to Reynolds number of 50,000 which the overall ef­
ficiency declines by 3% for oil, Al2 O3 , andCu and 11.6% for SWCNT.
• There is an optimum Reynolds number for ηp andηC which equals to
• The optimum Reynolds number to have less exergy destruction is
20000.
20,000 for all nanoparticles and oil for which exergy destruction falls
• Increasing Reynolds number from 1000 to 20,000 leads to an
about 64% for oil, Al2 O3 , andCu and 44.8% for SWCNT.
enhancement for ηp in order of 80% for synthetic oil.
• Adding Al2 O3 , Cu, andSWCNT nanopartciles into pure oil have sub­
• Adding Cu and SWCNT nanoparticles into synthetic oil diminishes
stantial impacts on the exergy destruction reduction of the order of
the ηp on the order of 13.3 and 285% forRe = 20000, respectively.
− 14.7%, − 27.8%, and − 54.9%, respectively.
• Injecting Cu and SWCNT nanoparticles into pure oil increases the ηc
by 2 and 23 times forRe = 20000, respectively.

Fig. 18. The variation of (a) net profit per unit transferred heat load (ηp ) and (b) total cost per unit transferred heat load (ηc ) against Reynolds number for both
laminar and turbulence regimes, and three different nanoparticles (Al2 O3 , Cu, andSWCNT) and synthetic oil.

18
B. Vahedi et al. Applied Thermal Engineering 206 (2022) 118117

5.4. Comparison between BGM and SPM in a PTC system • The optimum Reynolds number for fr equals to 20,000 for all
nanoparticles and synthetic oil.
In this study, the ratio of molecular Brownian diffusion to thermo­ • have and PEC have a direct non-linear relationship with the Reynolds
phoresis (NBT ) is 0.03, 0.53, and 2.2 for Al2 O3 , Cu, andSWCNT, respec­ number for all nanoparticles and pure oil.
tively. According to references [34,71] as the NBT increases, the effects • Utilizing SWCNT had the highest impact on have , fr, and PEC
of thermophoretic diffusion becomes weaker, leading to a more uniform enhancement relative to other nanoparticles. However, the thermal
nanoparticles distribution and as a result the Buongiorno’s model (BGM) efficiency of SWCNT − oil was reduced by 2.8% compared to that of
becomes closer to the single-phase model (SPM). Xu et al.[71] reported pure oil case due to need for higher pumping power.
that increasing the NBT from 0.1 to 10 led to a 7.5% rise in Nuave for • Adding nanoparticles and increasing the Reynolds number have no
volume fraction of 4% for forced convection. In fact, the closer the remarkable effects on the thermal efficiency of the system for Rey­
calculated NBT to one or higher, the more nanofluid distribution be­ nolds number up to 50000.
comes uniform, and therefore the more two-phase results tend towards • The trend of the exergy efficiency was opposite of the exergy
the single-phase results. In the present study, some cases were consid­ destruction. Moreover, the obtained maximum exergy efficiency was
ered for Al2 O3 with volume fraction of 4% as the BGM, and the ratio of of the order of 74.4% for oil, Al2 O3 , andCu and 79.9% for SWCNT
the BGM to SPM for Nuave and fr are compared as shown in Fig. 19. In the which occurred at Re = 20000.
laminar regime, the values of fr for BGM are less than SPM. This is due to • The ηp values for oil, Al2 O3 , andCu is higher than zero, whereas it is
the fact that the the nanoparticles concentration near tube wall reduces less than zero for SWCNT for all Reynolds number.
due to the domination of thermophoresis relative to the Brownian • Optimum Reynolds number for ηp was 20000, where ηp was 1.75 and
diffusion and advection. As a result, the dynamic viscosity reduces. In 1.24 times higher than the ηp at Re = 1000 and 50000.
the turbulence regime, the value of fr and Nuave for BGM and SPM are the • It was established that for forced convection problems by considering
same due to the advection dominance. Higher values of Nuave for BGM the value of NBT , it is efficient to use the SPM approach to model
relative to SPM is related to the decrease of the thermal conductivity nanofluid flow by considering the importance of analyzing the heat
near the tube wall. The final results show a maximum of 3.1 and 2.4% transfer. However, to show the nanoparticles distribution the non-
difference between the BGM and SPM for Nuave and fr, respectively. homogenous model should be used.

6. Conclusions Perspectives

In the present study, the oil based working fluid flow inside the • Implementing other two-phase methods to predict and consider all
absorber tube of the designed parabolic trough collector was investi­ possible behaviors of nanoparticles in the base fluid to gain more
gated. Non-uniform heat flux and thermal losses were incorporated into realistic results about the solid nanoparticles distribution.
the OpenFoam code. In the energy equation the dissipation term, and • Controlling high-temperature gradient effect especially in lower half
nanoparticle distribution with Brownian movement and thermophoresis of the absorber tube by utilizing partial porous media, installing fins,
phenomena were considered. In addition to the laminar regime, the using hybrid nanofluids, and turbulator to uniformly diffuse the heat
turbulent regime based on the Lien cubic k − ∊ model was implemented. through the fluid to reduce the absorber tube wall temperature.
The impact of the different Reynolds numbers and adding nanoparticles • Reducing the heat losses from the absorber tube by adding two or
including Al2 O3 , Cu, andSWCNT based synthetic oil were investigated on more glass envelopes, designing various shapes for glass cover, and
the dimensionless friction factor (fr), average convective heat transfer using cutting-edge insulating materials such as aerogel in the
coefficient (have ), performance evaluation criteria (PEC) and thermo­ annulus.
dynamic parameters such as thermal efficiency (ηth ), exergy efficiency • The surface of the absorber tube is vulnerable to thermal stresses
(ηex ), overall efficiency (ηov ), total exergy destruction (χ̇ d,tot ), Bejan especially when it is coated with ceramic because of high-
number (Be) and exergoeconomic parameters such as net profit per unit temperature gradient. In this regard, it is better to study the mate­
transferred heat load (ηp ), and total cost per unit transferred heat load rials that have a high thermal expansion coefficient to resist rapid
(ηC ). The following significant observations were obtained:

Fig. 19. Comparison of (a) friction factor (fr) and (b) average Nusselt number (Nuave ) between the Buongiorno’s model (BGM) and single-phase model (SPM).

19
B. Vahedi et al. Applied Thermal Engineering 206 (2022) 118117

changes in the temperature of the absorber tube and reduce the Sustain. Energy Technol. Assess. 50 (2022) 101888, https://doi.org/10.1016/j.
seta.2021.101888.
probability of deformation and crack formation.
[21] I. Zahmatkesh, M. Sheremet, L. Yang, S.Z. Heris, M. Sharifpur, J.P. Meyer,
• When the temperature of the heat transfer fluid exceeds the M. Ghalambaz, S. Wongwises, D. Jing, O. Mahian, Effect of nanoparticle shape on
maximum level, thermal degradation of working fluid occurs. In the performance of thermal systems utilizing nanofluids: A critical review, J. Mol.
order to promote the PTC’s working temperature, it is highly rec­ Liq. 321 (2021) 114430, https://doi.org/10.1016/j.molliq.2020.114430.
[22] K. Khanafer, K. Vafai, A critical synthesis of thermophysical characteristics of
ommended to study different types of oils to boost their boiling nanofluids, Int. J. Heat Mass Transf. 54 (19-20) (2011) 4410–4428.
temperature and prevent them from degradation. [23] M. Awais, N. Ullah, J. Ahmad, F. Sikandar, M.M. Ehsan, S. Salehin, A.A. Bhuiyan,
Heat transfer and pressure drop performance of Nanofluid: A state-of-the-art
review, Int. J. Thermofluids 9 (2021) 100065, https://doi.org/10.1016/j.
Declaration of Competing Interest ijft.2021.100065.
[24] H. Olia, M. Torabi, M. Bahiraei, M.H. Ahmadi, M. Goodarzi, M.R. Safaei,
Application of nanofluids in thermal performance enhancement of parabolic
The authors declare that they have no known competing financial trough solar collector: state-of-the-art, Appl. Sci. 9 (3) (2019) 463.
interests or personal relationships that could have appeared to influence [25] A. Menbari, A.A. Alemrajabi, A. Rezaei, Heat transfer analysis and the effect of
CuO/Water nanofluid on direct absorption concentrating solar collector, Appl.
the work reported in this paper.
Therm. Eng. 104 (2016) 176–183.
[26] A. Mwesigye, Z. Huan, J.P. Meyer, Thermodynamic optimisation of the
References performance of a parabolic trough receiver using synthetic oil–Al2O3 nanofluid,
Appl. Energy 156 (2015) 398–412.
[27] A. Mwesigye, İ.H. Yılmaz, J.P. Meyer, Numerical analysis of the thermal and
[1] N. Abas, N. Khan, Carbon conundrum, climate change, CO2 capture and
thermodynamic performance of a parabolic trough solar collector using SWCNTs-
consumptions, J. CO2 Util. 8 (2014) 39–48.
Therminol® VP-1 nanofluid, Renew. Energy 119 (2018) 844–862.
[2] N. Zhang, S.-Y. Yin, K. Vafai, Transient Performance of a Solar Humidification-
[28] A. Albojamal, H. Hamzah, A. Haghighi, K. Vafai, Analysis of nanofluid transport
Dehumidification Desalination System based on Hollow Fiber Membrane,
through a wavy channel, Numerical Heat Transfer, Part A: Applications 72 (12)
J. Comput. Des. Eng. 8 (2021) 923–934, https://doi.org/10.1093/jcde/qwab023.
(2017) 869–890.
[3] R. Senthil, R. Madurai Elavarasan, R. Pugazhendhi, M. Premkumar, E. Vengadesan,
[29] J. Buongiorno, Convective transport in nanofluids, 2006.
S. Navakrishnan, M.R. Islam, S.K. Natarajan, A holistic review on the integration of
[30] A. Albojamal, K. Vafai, Analysis of particle deposition of nanofluid flow through
heat pipes in solar thermal and photovoltaic systems, Sol. Energy 227 (2021)
porous media, Int. J. Heat Mass Transf. 161 (2020) 120227, https://doi.org/
577–605.
10.1016/j.ijheatmasstransfer.2020.120227.
[4] S.S. Christopher, R. Santosh, M. Ponrajan Vikram, R. Prabakaran, A.K. Thakur,
[31] M. Goodarzi, M.R. Safaei, K. Vafai, G. Ahmadi, M. Dahari, S.N. Kazi, N. Jomhari,
H. Xu, Optimization of a solar water heating system for vapor absorption
Investigation of nanofluid mixed convection in a shallow cavity using a two-phase
refrigeration system, Environ. Prog. Sustain. Energy 40 (1) (2021), https://doi.org/
mixture model, Int. J. Therm. Sci. 75 (2014) 204–220.
10.1002/ep.v40.110.1002/ep.13489.
[32] N. Shehzad, A. Zeeshan, R. Ellahi, K. Vafai, Convective heat transfer of nanofluid in
[5] F. Wang, J. Tan, Y. Shuai, L. Gong, H. Tan, Numerical analysis of hydrogen
a wavy channel: Buongiorno’s mathematical model, J. Mol. Liq. 222 (2016)
production via methane steam reforming in porous media solar thermochemical
446–455.
reactor using concentrated solar irradiation as heat source, Energy Convers.
[33] A. Albojamal, K. Vafai, Analysis of single phase, discrete and mixture models, in
Manage. 87 (2014) 956–964.
predicting nanofluid transport, Int. J. Heat Mass Transf. 114 (2017) 225–237.
[6] P. Wang, K. Vafai, D.Y. Liu, “Analysis of Radiative Effect under Local Thermal Non-
[34] R.O. Sayyar, M. Saghafian, Numerical simulation of convective heat transfer of
Equilibrium Conditions in Porous Media-Application to a Solar Air Receiver,
nonhomogeneous nanofluid using Buongiorno model, Heat Mass Transf. 53 (8)
Numerical Heat Transfer J. 65 (2014) 931–948.
(2017) 2627–2636.
[7] T. Salameh, Y. Zurigat, A. Badran, C. Ghenai, M. El Haj Assad, K. Khanafer,
[35] Y. Chen, J. Xu, D. Zhao, J. Wang, P.D. Lund, Exergo-economic assessment and
K. Vafai, Numerical Investigation of Two-Phase Flow Over Unglazed Plate Collector
sensitivity analysis of a solar-driven combined cooling, heating and power system
Covered With Porous Material of Wire Screen for Solar Water Heater Application,
with organic Rankine cycle and absorption heat pump, Energy 230 (2021) 120717,
J. Sol. Energy Eng. 141 (3) (2019), 031009.
https://doi.org/10.1016/j.energy.2021.120717.
[8] E.K. Summers, M.A. Antar, J.H. Lienhard, Design and optimization of an air heating
[36] S.-Y. Wu, L. Xiao, S.-M. Chen, Y.-R. Li, J.-L. Xu, C. Liu, Exergo-economic criteria for
solar collector with integrated phase change material energy storage for use in
performance evaluation of enhanced heat transfer duct with constant wall
humidification–dehumidification desalination, Sol. Energy 86 (11) (2012)
temperature, Appl. Therm. Eng. 36 (2012) 393–402.
3417–3429.
[37] S.A. Makkeh, A. Ahmadi, F. Esmaeilion, M.A. Ehyaei, Energy, exergy and
[9] C. Angeli, E. Leonardi, L. Maciocco, A computational study of salt diffusion and
exergoeconomic optimization of a cogeneration system integrated with parabolic
heat extraction in solar pond plants, Sol. Energy 80 (11) (2006) 1498–1508.
trough collector-wind turbine with desalination, J. Cleaner Prod. 273 (2020)
[10] E. Bellos, C. Tzivanidis, Alternative designs of parabolic trough solar collectors,
123122, https://doi.org/10.1016/j.jclepro.2020.123122.
Prog. Energy Combust. Sci. 71 (2019) 81–117.
[38] Z. Su, S. Gu, K. Vafai, Modeling and Simulation of Ray Tracing for Compound
[11] W. Fuqiang, C. Ziming, T. Jianyu, Y. Yuan, S. Yong, L. Linhua, Progress in
Parabolic Thermal Solar Collector, Int. Commun. Heat Mass Transfer 87 (2017)
concentrated solar power technology with parabolic trough collector system: A
169–174, https://doi.org/10.1016/j.icheatmasstransfer.2017.06.021.
comprehensive review, Renew. Sustain. Energy Rev. 79 (2017) 1314–1328.
[39] S.C. Mullick, S.K. Nanda, An improved technique for computing the heat loss factor
[12] M. Mehos, C. Turchi, J. Jorgenson, P. Denholm, C. Ho, K. Armijo, On the path to
of a tubular absorber, Sol. Energy 42 (1) (1989) 1–7.
SunShot-advancing concentrating solar power technology, performance, and
[40] R. Forristall. Heat transfer analysis and modeling of a parabolic trough solar
dispatchability (No. NREL/TP-5500-65688 SAND-2016-2237 R). EERE Publication
receiver implemented in engineering equation solver. National Renewable Energy
and Product Library, 2016.
Lab., Golden, CO.(US); 2003. NREL/TP-550-34169.
[13] D. Ercole, O. Manca, K. Vafai, An investigation of thermal characteristics of
[41] S.K. Solar, Energy Engineering: Processes and Systems, 2nd ed., 1125 Academic
eutectic molten salt-based nanofluids, Int. Commun. Heat Mass Transfer 87 (2017)
Press, United Kingdom, 2013.
98–104.
[42] M.V. Bozorg, M. Hossein Doranehgard, K. Hong, Q. Xiong, CFD study of heat
[14] H.M. Sandeep, U.C. Arunachala, Solar parabolic trough collectors: A review on
transfer and fluid flow in a parabolic trough solar receiver with internal annular
heat transfer augmentation techniques, Renew. Sustain. Energy Rev. 69 (2017)
porous structure and synthetic oil–Al2O3 nanofluid, Renew. Energy 145 (2020)
1218–1231.
2598–2614.
[15] E. Bellos, C. Tzivanidis, Enhancing the performance of a parabolic trough collector
[43] N.Z.A.E.R.V.K. Shehzad, A. Zeeshan, R. Ellahi, K. Vafai, Convective heat transfer of
with combined thermal and optical techniques, Appl. Therm. Eng. 164 (2020)
nanofluid in a wavy channel: Buongiorno’s mathematical model, J. Mol. Liq. 222
114496, https://doi.org/10.1016/j.applthermaleng.2019.114496.
(2016) 446–455.
[16] K. Milani Shirvan, M. Mamourian, S. Mirzakhanlari, R. Ellahi, K. Vafai, Numerical
[44] F.S. Lien, W.L. Chen, M.A. Leschziner, Low-Reynolds-Number Eddy-Viscosity
investigation and sensitivity analysis of effective parameters on combined heat
Modelling Based on Non-Linear Stress-Strain/Vorticity Relations, Eng. Turbulence
transfer performance in a porous solar cavity receiver by response surface
Model. Experiments 3 (1996) 91–100.
methodology, Int. J. Heat Mass Transf. 105 (2017) 811–825.
[45] H. Khakrah, A. Shamloo, S. Kazemzadeh Hannani, Exergy analysis of parabolic
[17] K. Khanafer K. Vafai, Applications of Nanomaterials in Solar Energy and
trough solar collectors using Al2O3/synthetic oil nanofluid, Sol. Energy 173 (2018)
Desalination Sectors” Advances in Heat Transfer, Vol. 45, Chapter 5, pp. 303-329.
1236–1247.
https://doi.org/10.1016/B978-0-12-407819-2.00005-0, (2013).
[46] S.Y. Motlagh, E. Golab, A.N. Sadr, Two-phase modeling of the free convection of
[18] B. Bakthavatchalam, K. Habib, R. Saidur, B.B. Saha, K. Irshad, Comprehensive
nanofluid inside the inclined porous semi-annulus enclosure, Int. J. Mech. Sci. 164
study on nanofluid and ionanofluid for heat transfer enhancement: A review on
(2019), 105183.
current and future perspective, J. Mol. Liq. 305 (2020) 112787, https://doi.org/
[47] İ.H. Yılmaz, A. Mwesigye, T.T. Göksu, Enhancing the overall thermal performance
10.1016/j.molliq.2020.112787.
of a large aperture parabolic trough solar collector using wire coil inserts,
[19] A. Borode, N. Ahmed, P. Olubambi, A review of solar collectors using carbon-based
Sustainable Energy Technol. Assess. 39 (2020) 100696, https://doi.org/10.1016/j.
nanofluids, J. Cleaner Prod. 241 (2019) 118311, https://doi.org/10.1016/j.
seta.2020.100696.
jclepro.2019.118311.
[48] R.L. Hamilton, O.K. Crosser, Thermal conductivity of heterogeneous two-
[20] S. Samiezadeh, R. Khodaverdian, M.H. Doranehgard, H. Chehrmonavari, Q. Xiong,
component systems, Ind. Eng. Chem. Fundam 1 (3) (1962) 187–191.
CFD simulation of thermal performance of hybrid oil-Cu-Al2O3 nanofluid flowing
through the porous receiver tube inside a finned parabolic trough solar collector,

20
B. Vahedi et al. Applied Thermal Engineering 206 (2022) 118117

[49] C.-W. Nan, G. Liu, Y. Lin, M. Li, Interface effect on thermal conductivity of carbon [62] J.M. McDonough, Introductory lectures on turbulence: physics, mathematics and
nanotube composites, Appl. Phys. Lett. 85 (16) (2004) 3549–3551. modelling, 2007.
[50] A. Mwesigye, J.P. Meyer, Optimal thermal and thermodynamic performance of a [63] T.J. Craft, A.V. Gerasimov, H. Iacovides, B.E. Launder, Progress in the
solar parabolic trough receiver with different nanofluids and at different generalization of wall-function treatments, Int. J. Heat Fluid Flow 23 (2) (2002)
concentration ratios, Appl. Energy 193 (2017) 393–413. 148–160.
[51] N. Afzal, A. Seena, A. Bushra, Power law velocity profile in fully developed [64] S.Z. Heris, S.G. Etemad, M.N. Esfahany, Experimental investigation of oxide
turbulent pipe and channel flows, J. Hydraul. Eng. 133 (9) (2007) 1080–1086. nanofluids laminar flow convective heat transfer, Int. Commun. Heat Mass Transfer
[52] S. Goudarzi, M. Shekaramiz, A. Omidvar, E. Golab, A. Karimipour, A. Karimipour, 33 (4) (2006) 529–535.
Nanoparticles migration due to thermophoresis and Brownian motion and its [65] B.C. Pak, Y.I. Cho, Hydrodynamic and heat transfer study of dispersed fluids with
impact on Ag-MgO/Water hybrid nanofluid natural convection, Powder Technol. submicron metallic oxide particles, Experimental Heat Transfer Int. J. 11 (2)
375 (2020) 493–503. (1998) 151–170.
[53] S.M. Jeter, Calculation of the concentrated flux density distribution in parabolic [66] V.E. Dudley, G.J. Kolb, A.R. Mahoney, T.R. Mancini, C.W. Matthews, Michael
trough collectors by a semifinite formulation, Sol. Energy 37 (5) (1986) 335–345. Sloan, D. Kearney, Test results: SEGS LS-2 solar collector (No. SAND94-1884).
[54] Y.A. Cengel, A.J. Ghajar, Heat and mass transfer. A practical approach, 2007. Sandia National Lab.(SNL-NM), Albuquerque, NM (United States), 1994.
[55] A. Bejan, A Study of Entropy Generation in Fundamental Convective Heat Transfer, [67] G.A. Sheikhzadeh, M. Dastmalchi, H. Khorasanizadeh, Effects of nanoparticles
J. Heat Transf. 101 (1979) 718–725, https://doi.org/10.1115/1.3451063. transport mechanisms on Al2O3–water nanofluid natural convection in a square
[56] H. Herwig, F. Kock, Direct and indirect methods of calculating entropy generation enclosure, Int. J. Therm. Sci. 66 (2013) 51–62.
rates in turbulent convective heat transfer problems, Heat Mass Transf. 43 (3) [68] C.J. Ho, W.K. Liu, Y.S. Chang, C.C. Lin, Natural convection heat transfer of
(2007) 207–215. alumina-water nanofluid in vertical square enclosures: An experimental study, Int.
[57] R. Petela, Engineering thermodynamics of thermal radiation: for solar power J. Therm. Sci. 49 (8) (2010) 1345–1353.
utilization, McGraw-Hill Education, 2010. [69] S.Y. Motlagh, H. Soltanipour, Natural convection of Al2O3-water nanofluid in an
[58] E. Bellos, C. Tzivanidis, D. Tsimpoukis, Optimum number of internal fins in inclined cavity using Buongiorno’s two-phase model, Int. J. Therm. Sci. 111 (2017)
parabolic trough collectors, Appl. Therm. Eng. 137 (2018) 669e677. 310–320.
[59] E. Bellos, C. Tzivanidis, K.A. Antonopoulos, G. Gkinis, Thermal enhancement of [70] S. Goudarzi, S. Yekani Motlagh, H. Soltanipour, Two-phase simulation of MHD and
solar parabolic trough collectors by using nanofluids and converging-diverging FHD impacts on the natural convection of a magnetic nanofluid within a porous
absorber tube, Renew. Energy 94 (2016) 213–222. cavity, AUT J. Mech. Eng. 6 (1) (2022) 9.
[60] H. Price. A parabolic trough solar power plant simulation model. In: International [71] H. Xu, Z. Xing, K. Vafai, Analytical considerations of flow/thermal coupling of
solar energy conference (Vol. 36762, pp. 665-673), 2003, January. nanofluids in foam metals with local thermal non-equilibrium (LTNE) phenomena
[61] W. Fuqiang, T. Jianyu, M. Lanxin, W. Chengchao, Effects of glass cover on heat flux and inhomogeneous nanoparticle distribution, Int. J. Heat Fluid Flow 77 (2019)
distribution for tube receiver with parabolic trough collector system, Energy 242–255.
Convers. Manage. 90 (2015) 47–52.

21