Applied Thermal Engineering 98 (2016) 1044–1053

Contents lists available at ScienceDirect

Applied Thermal Engineering

j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / a p t h e r m e n g

Research Paper

Forced convection enhancement in channels with transversal ribs and

nanofluids
Assunta Andreozzi a, Oronzio Manca b,*, Sergio Nardini b, Daniele Ricci c
a Dipartimento di Ingegneria Industriale, Università degli Studi di Napoli Federico II, Piazzale Tecchio 80, 80125 Napoli, Italy
b
Dipartimento di Ingegneria Industriale e dell’Informazione, Seconda Università degli Studi di Napoli, via Roma 29, 81031 Aversa, Italy
c CIRA – Centro Italiano Ricerche Aerospaziali, Via Maiorise, 81043 Capua (CE), Italy

H I G H L I G H T S

• Turbulent forced convection in nanofluid mixture, water–Al2O3, is carried out.

• The geometry is a symmetrically heated channel with ribbed upper and lower walls.
• Analysis for ribs with different shapes: triangular, rectangular and trapezoidal.
• The numerical analysis is accomplished using the commercial code Fluent.
• Triangular ribs have the highest thermal performances but also the highest pressure losses.

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

Article history: A numerical investigation on turbulent forced convection in nanofluid mixture, water–Al2O3, in a
Received 21 September 2015 symmetrically heated channel with ribbed upper and lower walls is carried out. Different geometric rib
Accepted 29 December 2015 arrangements with triangular, rectangular and trapezoidal shapes are analyzed. The governing equa-
Available online 11 January 2016
tions are written assuming the mixture model to describe the nanofluid behavior in a two dimensional
channel heated by a constant wall heat flux. Finite volume method is used to solve the model
Keywords:
equations. Simulations for different nanoparticle volumetric concentrations from 0%, pure water, to 4%
Heat transfer enhancement
and Reynolds numbers between 20,000 and 60,000 are accomplished. As general conclusion, it is
Ribs
Nanofluids observed that the heat transfer rate increases as nanoparticle volume fraction and Reynolds number
Forced convection increase but higher pumping powers are required. The comparisons allow to determine that triangular
Turbulent flow ribs present higher thermal performances than the trapezoidal ones but also higher pressure losses.
Moreover, the rectangular-trapezoidal-shaped rib shows the best performance at a pitch–rib height
equal to 10.
© 2015 Elsevier Ltd. All rights reserved.

1. Introduction channels, circular tubes with inserts, longitudinal flows in rod

bundles and annuli, flow orthogonal to cylinders. In fact, artificial
In the last three decades, a very important point, both in the roughness is employed in several devices and systems: evapora-
research and in the industry, is represented by the heat transfer tors, condensers, steam condensers, solar air heaters, nuclear reactors
enhancement and a great attention is set on it. Forced convection heat exchangers, and gas turbine blade cooling channels, as re-
enhancement is obtained by means of several techniques in thermal ported in references from 3 to 12. Fluid dynamic and thermal
devices development and it allows to realize size reductions, effi- characteristics in different devices with turbulators have been evalu-
ciency improvement or cost savings. The realizable techniques ated experimentally and numerically [13–30]. In ribbed channels,
were listed into ‘passive’ and ‘active’ by Bergles [1]. The former the laminar sub-layer is interrupted by turbulators and a flow
techniques use coated surfaces, rough or extended surfaces, swirl separation and reattachment is determined. In this way, heat trans-
flow devices or additives in fluids, as indicated by Webb and Kim fer rate increases, due to the thermal resistance reduction, but a
[2]. A lot of information are available in single-phase forced con- pressure loss increase is detected. A growth in numerical investi-
vection flow over for flat plates or ribbed surfaces, ducts and gations with distinct arrangements of rough elements in terms of
height, shape and orientation of ribs is due to the computer power
increment and the numerical methods enhancement [14–16].
* Corresponding author. Tel.: +390815010217; fax: +390815010463. Liou et al. [14] investigated both numerically and experimentally
E-mail address: [email protected] (O. Manca). a rectangular-ribbed channel with elements mounted on one

http://dx.doi.org/10.1016/j.applthermaleng.2015.12.140
1359-4311/© 2015 Elsevier Ltd. All rights reserved.
 A. Andreozzi et al./Applied Thermal Engineering 98 (2016) 1044–1053 1045

principal wall. The importance of flow acceleration and turbu- was investigated experimentally by Wang et al. [30]. The results
lence intensity was highlighted in order to describe the main detected that the presence of ribs augments the heat transfer at
factors, influencing heat transfer coefficients. The highest enhance- the penalty of increased heat transfer non-uniformity and pres-
ment of the average heat transfer values was estimated at a pitch- sure loss.
to-rib-height ratio equal to 10. Rau et al. [15] performed an Different techniques, to improve the heat transfer enhance-
experimental investigation on a similar configuration given in ref- ment, can be used and a simple solution is to employ ribbed channels
erence 14 and they confirmed the results provided in reference with nanofluids, which allow to have fluids with higher thermal con-
14. Moreover, the importance of the geometry arrangements in ductivity values [31]. Nanoparticles with dimensions less than
the heat transfer enhancement was also remarked. A study on 100 nm are inserted in the fluid and a nanofluid mixture is ob-
turbulent flows due to turbulators in two and three-dimensional tained. Today nanofluids are becoming very popular in heat transfer
arrangements was developed numerically by Ryu et al. [17,18]. research and applications due to their possible improving of the
The fluid-dynamic and thermal behaviors were investigated on thermal performances in cooling devices and heat exchangers
ribbed channels with a dimensionless height of the ribs equal to [7,32–45].
0.1. The maximum resistance was detected at a pitch-to-height In the following a short review on investigations related to ribbed
ratio lesser than 10. An experimental investigation on the effect of channels with nanofluids is given.
rib height to the hydraulic diameter ratio on the local heat trans- A numerical study on thermal and hydraulic behaviors of tur-
fer coefficient in a double wall ribbed square channel with 90° bulent forced convection in nanofluids in rib–groove channel was
continuous attached and 60° V-broken ribs was accomplished by provided by Mohammed et al. [46]. A two dimensional model,
SriHarsha et al. [19]. The results highlighted that the heat transfer with the channel walls at assigned temperature, was assumed
augmentations in the 60° V-broken ribs are higher than those of and the finite volume method to solve the governing equations
90° continuous attached ribs. A numerical investigation on the air was used. Nine different rib–groove shapes were studied combin-
turbulent forced convection enhancement in channels with differ- ing three different rib shapes with three different groove shapes:
ently shaped ribs, mounted on two parallel walls, was accomplished rectangular, triangular and trapezoidal. Different nanofluid mix-
by Manca et al. [20]. The SST k–ω model was employed and nu- tures composed of four oxide nanoparticles, Al2O3, CuO, SiO2, and
merical model was allowed to evaluate the pitch-to-height ratio ZnO, dispersed in three base fluids (water, glycerin, engine oil)
values corresponding to maximum heat transfer coefficients and were employed in the investigation. The analysis in terms of Nusselt
friction factors. The numerical investigation was also employed to number, friction factor and thermal enhancement factor was ac-
investigate the forced convection in ribbed channels in water [21] complished in the volume fraction range from 1% to 4%, particle
and in nanofluids (Al2O3 nanoparticles in pure water) [22,23] diameter range from 25 to 80 nm, Reynolds number values from
modeled in single-phase. 5000 to 20,000 and rib–groove aspect ratio values from 0.5 to 4.
An experimental and numerical investigation on forced convec- The results showed that the highest Nusselt number among the
tion in channels with 90° ribs and V-shaped ribs was carried out rib–groove shapes was detected for the rectangular rib–triangular
by Peng et al. [24]. It was shown that both the 90° ribs and V-shaped groove and among the nanofluid mixtures it was achieved for the
ribs improve the convective heat transfer coefficient with respect SiO2 nanofluid. Moreover, the Nusselt number increased as the
to a flat wall, but the pressure drop grows. Moreover, the overall aspect ratio increased. The fluid flow and heat transfer character-
performances of the V-shaped ribs are higher than those of the istics in a vertical channel with symmetrically mounted rib heaters
90° ribs. A three-dimensional incompressible numerical study on were numerically investigated in laminar forced convection in
a ribbed channel was performed by Desrues et al. [25] to evaluate water–Al2O3 nanofluid by Gavara [47]. Results for Reynolds number
Nusselt numbers and friction factor. Pressure drop was found to values and volume fractions between 130 and 350 and 0% and
increase with Reynolds number, whereas heat transfer was aug- 10% were carried out. Asymmetric velocity and temperature fields
mented for Reynolds number values higher than a critical value. were detected for Reynolds number higher than a critical value,
The air flow in a channel with transverse and inclined vortex gen- Recr. The presence of nanoparticles determined an increase of Recr
erators was investigated experimentally by Mikielewicz et al. [26]. and it increased as nanoparticle volume fraction increased. A nu-
Results showed that the average Nusselt number values for pas- merical analysis of turbulent forced convection with nanofluid in
sages with ribs were generally two to three times higher than the a fully detached ribbed channel was accomplished by Parsazadeh
ones for the smooth channels. However, a pressure loss increase et al [48]. The investigation in a two dimensional channel, with
was detected. An experimental study to evaluate the heat transfer the bottom wall at assigned temperature and the top wall adiabat-
coefficients in a channel with both angled ribs and dimples was ic, was performed for Al2O3, CuO, SiO2, and ZnO nanoparticles in
provided by Choi et al. [27] using the transient liquid crystal tech- water to evaluate their effects in terms of nanoparticle param-
nique. Results showed that the secondary flow induced by the angled eters. The heat transfer enhancement, pressure drop and friction
ribs determined an asymmetric distribution of heat transfer coef- factor were evaluated for volumetric fraction values from 1% to
ficient and the presence of dimples between the ribs increased the 4%, Reynolds number in the range between 10,000 and 50,000
heat coefficient with an acceptable increase in pressure drop. Square and nanoparticles diameter values from 20 nm to 50 nm. The highest
cross-section ribbed channels with different arrangements of down- heat transfer enhancement was obtained for SiO2 nanofluid whereas
stream half-size ribs to determine the most optimal configurations the friction factor presented small limited variations for the differ-
for augmenting heat transfer rates with minimized pressure drop ent nanoparticles.
penalties were employed in a numerical study presented by Xie A two-dimensional analysis of forced convection in different
et al. [28]. Results showed that the application of downstream ribs nanofluids (Al2O3, CuO, ZnO and SiO2) through symmetrically
allows to decrease the pressure loss, remaining comparable the heated channels with rib–groove of triangular, square and arc
enhancement in heat transfer. Heat transfer in turbulent flow in a shapes was numerically carried out by Vatani and Mohammed
channel with two-dimensional ribs was investigated experimen- [49]. Average Nusselt number, friction coefficient and perfor-
tally and numerically by Wang et al. [29] to compare compound mance index were evaluated for channel walls heated at uniform
ribs with triangular ones. It was found that the compound ribs de- heat flux and with nanoparticle diameter between 30 and 70 nm,
termined an increase in heat transfer and a more significant decrease volumetric concentration varying from 1% to 4%, and Reynolds
in pressure drop with respect to the triangular ribs. The heat trans- numbers in the range of 30,000–50,000. It was found that the
fer and pressure drop in the turn region of a two-pass channel SiO2–water nanofluid provided the highest Nusselt number among
1046 A. Andreozzi et al./Applied Thermal Engineering 98 (2016) 1044–1053

all nanofluids considered and the highest performance evaluation 2. Physical model and governing equations
criterion (PEC) was detected for the rectangular rib–rectangular
groove. Turbulent forced convection in nanofluids for four differ- The physical problem under investigation is an incompressible
ent shapes of rib–groove channels was examined by Al-Shamani steady state turbulent forced flow in a two-dimensional (2D) ribbed
et al. [50]. A two dimensional model in a symmetrically heated channel with nanofluids. The geometrical configuration is given in
channel at uniform wall heat flux and four different nanofluid terms of dimensionless parameters as shown in Fig. 1. With refer-
mixtures, water based with nanoparticles (Al2O3, CuO, SiO2 and ence to Fig. 1a, the main parameters are the aspect ratio between
ZnO), was considered. Simulations for a volume concentration the channel length and its height, L/H, and the wall thickness-to-
between 1% and 4%, nanoparticle diameter in the range of 25– height ratio, s/H, assumed equal to 25 and 0.33, respectively. The
70 nm and Reynolds number values from 10,000 to 40,000 were channel wall is made of aluminum and its heating is symmetrical
achieved. Results allowed to detect that the trapezoidal shape and on the external surfaces and it is assumed at constant heat flux
with increasing height in the flow direction rib-trapezoidal groove equal to 10 kW/m2. The channel walls are characterized by the pres-
presented the best average Nusselt number and comparing the ence of turbulators, as reported in Fig. 1b, and different rib shapes
four nanofluids the highest value of average Nusselt number was are investigated.
presented by the SiO2 – nanofluid. Three-dimensional turbulent The governing equations in 2D form, under the aforemen-
forced convection in rectangular ribbed–grooved channel in tioned hypothesis and assuming the nanofluids thermo-physical
nanofluids at assigned wall temperature was investigated numer- properties independent of temperature, are written as
ically by Yang et al. [51]. The governing equations in single and
two-phase assumptions were considered and the standard k−ε ∂
(ρui ) = 0 (1)
turbulence model was used. It was observed that the average ∂x i
Nusselt number was found to improve more with smaller rib–
groove height ratios and some ratios of rib–groove pitch. Moreover, ∂ ∂P ∂ ⎡ ⎛ ∂u ∂u 2 ∂u ⎞ ⎤ ∂
an optimization by employing the response surface methodology ∂x j
(ρuiu j ) = − + ⎢μ ⎜ i + j − δij i ⎟ ⎥ +
∂x i ∂x j ⎣ ⎝ ∂x j ∂x i 3 ∂x j ⎠ ⎦ ∂x j
−ρui′u′j ( ) (2)
and the genetic algorithm method allowed to detect the optimal
value of performance factor at Reynolds number of 10,000 with
∂ ∂ ⎡⎛ c p μ t ⎞ ∂T ⎤
an enhancement of 18.2%. [ui (ρE + P )] = ⎜λ + ⎟ + ui ( τij )eff ⎥ (3)
In the present paper, a numerical investigation on a two- ∂x i ∂x j ⎢⎣⎝ Prt ⎠ ∂x j ⎦
dimensional turbulent forced convection in ribbed channels with
nanofluid mixtures, water/Al2O3, and symmetrically heated is carried with E the total energy, defined as E = cp T − P ρ + u2 2 , and with (τij)eff
out. The flow and the heat transfer performances are evaluated by the deviatoric stress tensor. As regards the turbulence model, the
employing the mixture model [52–54]. From the above literature following SST k–ω model [55] was adopted:
review it seems that the simple ribbed channel configuration without
∂ ∂ ⎛ ∂k ⎞ 
grooves has not been investigated for different rib shapes in tur- (ρkui ) = Γk + Gk − Yk + Sk (4)
bulent forced convection in nanofluids. This has strongly motivated ∂x i ∂x j ⎜⎝ ∂x j ⎟⎠
the present investigation. Moreover, the present study extends the
analyses, given in reference 23, and it is one of the first efforts which ∂ ∂ ⎛ ∂ω ⎞
(ρωkui ) = Γω + Gω − Yω + Dω + Sω (5)
uses the mixture model in ribbed channels. Simulations for four rib ∂x i ∂x j ⎜⎝ ∂x j ⎟⎠
shapes and different values of element pitch-to-height ratio are ac-
complished. Results for Reynolds numbers between 20,000 and ∂u j
with Gk = −ρui′u′j the production of turbulence kinetic energy, due
60,000, and particle volume concentrations in the range 0–4%, are ∂x i
carried out. to mean velocity gradients, and Gω the generation of ω, it results

a)

b)

Fig. 1. Picture of the model: (a) geometrical sketch; (b) characteristic rib parameters and considered shapes.
 A. Andreozzi et al./Applied Thermal Engineering 98 (2016) 1044–1053 1047

 = min (G , 10β*kω )
G (6) Table 1
k k Shear stress transport k–ω model constants.

σk,1 = 1.176 σk,2 = 1.0 σω,1 = 2.0 σω,2 = 1.168 α1 = 0.31 βi,1 = 0.075
α 
Gω = Gk (7) βi,2 = 0.0828 α ∞* = 1.0 α ∞ = 0.52 β ∞* = 0.09 βi = 0.072 α0 = 1/9
νt
Rβ = 8.0 Rk = 6.0 Rω = 2.95 ζ* = 1.5 σk = 2.0 σω = 2.0

In these equations νt represents the kinematic turbulent viscos-

ity and β*a model constant; α is defined as
βi = F1βi,1 + (1 − F1 )βi,2 (21)
α ∞ (α 0 + Ret Reω )
α= (8)
α* (1 + Ret Reω ) and Dω is the cross-diffusion term and it is given by

where Rω = 2.95 and α∞ is: 1 ∂k ∂ω

Dω = 2(1 − F1 )ρσ ω ,2 (22)
ω ∂x j ∂x j
α ∞ = F1α ∞ ,1 + (1 − F1 ) α ∞ ,2 (9)
Sk and Sω in Eqs. (4) and (5) are source terms.
βi,1 κ2 In Table 1, the model constants are summarized.
α ∞ ,1 = − (10)
β*∞ σ ω ,1 β*∞ The two-dimensional steady-state conduction equation in the
solid is given in the following:

βi,2 κ2 ∂ ⎛ ∂T ⎞
α ∞ ,2 = − (11) ⎜λ ⎟ =0 (23)
β*∞ σ ω ,2 β*∞ ∂x i ⎝ ∂x i ⎠

where κ = 0.41. Γk and Γω represent the effective diffusivity of k and The boundary conditions were assigned considering at the inlet
ω: section uniform velocity and temperature profiles with a turbu-
lence intensity of 1%. Moreover, on the solid walls the no-slip
μt condition was applied and a uniform heat flux was assigned on the
Γk = μ + (12)
external surface of the solid walls. In the outlet section velocity com-
σk
ponents and temperature derivatives are equal to zero.
μt Fluid-dynamic and thermal behaviors were evaluated by a nu-
Γω = μ + (13) merical study and the governing equations were solved by means
σω
of the commercial Ansys-Fluent code [56]. Pure water and a
with σk and σω the turbulent Prandtl numbers for k and ω, respec- nanofluid made of Al2O3 nanoparticles, with diameter of 38 nm, and
tively, whereas μt is the turbulent viscosity. Moreover, they are pure water were the working fluids.
evaluated by: In order to characterize the nanofluid behavior the mixture model
was assumed and the indications given in reference 56 to set the
1 numerical model were used. The mixture model describes the con-
σk = (14)
F1 σk ,1 + (1 − F1 ) σk ,2 tinuity, momentum and energy equations for the nanofluid mixture
and the volume fraction equation for the secondary phases, as well
1 as algebraic expressions for the relative velocities between the two
σω = (15)
F1 σ ω ,1 + (1 − F1 ) σ ω ,2 phases. In the mixture model, the slip between nanoparticles and
liquid and the drift velocity are considered. Thermo-physical prop-
ρk erties values for water and Al2O3 nanoparticles at a temperature of
μ t = α* (16) 300 K are given in Table 2. Constant properties were assumed because
ω
temperature differences between the channel inlet and outlet sec-
α* reduces the amplitude of the turbulent viscosity and it is esti- tions evaluated in preliminary simulations were small.
mated by: For assigned base fluid, the nanofluid thermo-physical proper-
ties depend on nanoparticle material and their concentrations. In
(α*0 + Ret Rek ) the present numerical investigation, the thermo-physical proper-
α* = α*∞ (17)
(1 + Ret Rek ) ties of the considered nanofluid, Al2O3/water, were evaluated by the
equations reported in references 58 to 61 and their values are re-
with Ret defined as ρk/μω, Rk = 6, α*0 = βi/3 and βi = 0.072. The term ported in Table 3 for volumetric concentrations equal to 0%, 1% and
F1 is evaluated by 4%,

F1 = tan (Φ14 ) (18) density: ρnf = (1 − ϕ )ρbf + ϕρp (24)

⎡ ⎛ k 500μ ⎞ 4ρk ⎤ specific heat: cp,nf = (1 − ϕ ) cp,bf + ϕcp,p (25)

Φ1 = min ⎢max ⎜ , 2 ⎟, + 2⎥
(19)
⎣ ⎝ 0.09ω y ρy ω ⎠ σ ω ,2Dω y ⎦
dynamic viscosity: μnf = μ bf (123ϕ 2 + 7.3ϕ + 1) (26)

⎡ 1 1 ∂k ∂w ⎤
Dω+ = max ⎢2ρ , 10−10 ⎥ (20)
⎣ σ ω ,2 ω ∂ x j ∂ x j ⎦
Table 2
Material properties at T = 300 K [57].
with y the distance to the next surface and Dω+ the positive portion
of the cross-diffusion term. Yk = ρβ*kω and Yω = ρβω 2 are the k Material ρ[kg/m3] cp [J/kg K] μ[Pa s] λ [W/mK]
and ω dissipations due to turbulence with β and β* constants. Al2O3 3880 773 // 36
Moreover: Water 998 4182 998 × 10−6 0.597
1048 A. Andreozzi et al./Applied Thermal Engineering 98 (2016) 1044–1053

Table 3 merically with ones calculated by using the correlations for water
Thermo-physical properties of the working fluid. and nanofluid.
φ ρ[kg/m3] cp [J/kgK] μ[Pa s] λ[W/mK] For water the following correlations were considered [57]:
0.00 996.5 4177 845 × 10−6 0.610
0.01 1025.3 4143 917 × 10−6 0.618 Dittus -Boelter : Nus = 0.024 Re0.8 Pr 0.4 (31)
0.04 1111.8 4041 1258 × 10−6 0.674
for 0.5 < Pr < 120, 6000 < Re < 107; 60 < (L/d);

Petukhov: fs = (1.84 log10 Re− 1.64 )

−2
(32)

thermal conductivity: λ nf = λ bf ( 4.97ϕ 2 + 2.72ϕ + 1) (27) for 0.5 < Pr < 200, 10,000 < Re < 5 106;

McAdams: fs = 0.184 Re−0.2 (33)

3. Roughness and dimensionless characteristic parameters
for 20,000 < Re < 300,000.
The geometrical parameters related to the internal surface ar- The chosen correlations for Al2O3/water based nanofluids were
tificial roughness of the channel, with reference to Fig. 1, are mainly the following [58–61]:
the shape of the ribs, the height (e), pitch (p) and longitudinal width
(w), in the following indicated only with width. The dimension- Pak-Cho: Nus = 0.021Re0.8 Pr 0.5 (34)
less form of characteristic geometrical parameters is expressed by
relative roughness height, e/dh, relative roughness pitch, p/e, and rel- Maiga: Nus = 0.085Re0.71 Pr 0.35 (35)
ative width, w/e. For assigned rib type with constant p/e and w/e
ratios, a family of a geometrically similar roughness is defined by The comparison in terms of average Nusselt number and fric-
changing e/dh ratio. In the present numerical study two roughness tion factor are reported in Figs. 2 and 3. Good agreements among
families were analyzed for different e/dh ratios. Moreover, the rel- the data are observed.
ative roughness pitch p/e is assigned between 4 and 20 whereas the
Reynolds number is changed from 20,000 to 60,000.
5. Results and discussion
Isosceles and rectangular trapezoidal ribs with w/e = 2.0 and
e/dh = 0.05 were considered in a channel with a hydraulic diame-
The effect of dimensionless pitches, rib shapes, Reynolds numbers
ter equal to 0.02 m. The main characteristic dimensionless
and particle volume fractions on average Nusselt number and re-
parameters are the Reynolds number, the average Nusselt number
quired pumping power are presented. Results are also reported in
and the friction factor. They are defined as
terms of temperature fields and streamline contours for signifi-
uav dh cant configurations are carried out as well.
Re = (28) Average Nusselt numbers increment to the values for smooth
ν
channel filled with water are reported in Fig. 4 as a function of the
q w dh p/e ratio for ‘iso’ ribs at Re = 40,000 and volume particle concen-
Nuav = (29) trations equal to 0, 0.01 and 0.04. For each volume concentration
(Tw − Tm ) λ f Nusselt number increment profile has a maximum value at p/e = 10
where Tm is the bulk temperature of fluid and the curves slope are higher beyond this p/e value. As expected
the presence of nanoparticles in the base fluid increases the heat
dh 1
f = 2ΔP (30)
L ρu2

4. Numerical procedure

The governing equations in the steady state regime were

solved by means of the finite volume method employing the
Ansys-Fluent code [56]. The SIMPLE scheme was chosen to
evaluate numerically the couple velocity and pressure fields. A
segregated technique, with the adoption of a second-order upwind
scheme, was used. The convergence criteria for the residuals of
velocity components and energy were set equal to 10−6 and 10−9,
respectively.
The grid independence analysis of numerical solution was ac-
complished with four different structured mesh distributions on the
channel with rectangular ribs, for p/e = 10 and w/e = 0.5. The number
of nodes in the meshes was 55,400, 108,100, 189,900 and 380,000,
respectively whereas the y+ parameter was kept under values equal
to 1 for all four grids. The comparison between the third and fourth
showed differences of 0.15 K, 0.004 m/s, 3.5 Pa in terms of wall
average temperature, average outlet velocity and pressure drops,
in the case of pure water. The third grid case was adopted because
it ensured a good compromise between the accuracy require-
ments and the computational time. The validation test was
performed in terms of average Nusselt number and friction factor
for the smooth channel by comparing the results carried out nu- Fig. 2. Validation of numerical results in terms of average Nusselt number.
 A. Andreozzi et al./Applied Thermal Engineering 98 (2016) 1044–1053 1049

Fig. 3. Validation of numerical results in terms of friction factor.

Fig. 5. Average Nusselt number enhancement profiles, referred to smooth cases in
water, as a function of p/e at Re = 40,000, for e/dh = 0.05, φ = 0.0, 0.01 and 0.04, ribs
‘rect’.
transfer coefficient. By using the mixture model, the highest values
of Nu/Nubf,s ratio are equal to 2.17, 2.24 and 2.58 for φ = 0.0, 0.01
and 0.04 (“phi” in all figures), respectively. A comparison between
average Nusselt number increment, obtained using the mixture also in this case. The highest values of Nu/Nubf,s ratio are equal to
model results, and the ones from the single-phase model, which 2.07, 2.14 and 2.39 for φ = 0.0, 0.01 and 0.04, respectively, by using
are reported in references 22 and 23, is also presented. The com- the mixture model.
parisons show that the single-phase model slightly underestimates The presence of nanoparticles in the base fluid gives an incre-
the Nusselt number values. The increment in terms of heat ment of viscosity and, therefore, higher pumping powers are needed
transfer coefficient is even higher, due to the effect of thermal con- than the ones for pure base fluid. By defining the required pumping
ductivity; in fact, its values are 2.27 and 2.85 for φ = 0.01 and 0.04, power as PP = uavAΔP, Fig. 6 shows pumping power profiles incre-
respectively. ment with respect to the smooth channel with pure water, for ‘iso’
When the ribs are ‘rect’, Nusselt number increment values are and ‘rect’ ribs, Re = 40,000, e/dh = 0.05 and φ = 0.0, 0.01 and 0.04 as
lower, as reported in Fig. 5. Maximum values are attained at p/e = 10 a function of the p/e ratio. It can be observed that mixture and

Fig. 4. Average Nusselt number enhancement profiles, referred to smooth cases in

water, as a function of p/e at Re = 40,000, for e/dh = 0.05, φ = 0.0, 0.01 and 0.04, ribs Fig. 6. Pumping power profiles, referred to smooth cases in water, as a function of
‘iso’. p/e at Re = 40,000, for e/dh = 0.05, φ = 0.0, 0.01 and 0.04, ‘iso’ and ‘rect’ ribs.
1050 A. Andreozzi et al./Applied Thermal Engineering 98 (2016) 1044–1053

Fig. 7. Average Nusselt number profiles as a function of Re, for different shapes at
significant pitches, e/dh = 0.05 and φ = 0.04. Fig. 8. Required pumping power profiles as a function of Re, for different shapes at
significant pitches, e/dh = 0.05 and φ = 0.04.

single-phase models give similar results. For each profile, the

maximum values are reached at p/e = 10 and the highest pumping (T − Tm,in )
T* = (36)
power value is for φ = 0.04. For ‘iso’ ribs PP/PPbf,s ratio is equal to about (Tm,out − Tm,in )
10 and 22 for φ = 0.01 and 0.04, respectively. Ribbed channels with
Wall temperature values behind the rib are very high whereas
trapezoidal ‘rect’ turbulators show lower values of required pumping
temperature is low on the turbulators windward surface and high
power increment, in fact this ratio is equal to about 8 and 20 at
on the leeward one. At φ = 0.04, Fig. 10b, difference of tempera-
p/e = 10 for φ = 0.01 and 0.04, respectively.
ture between the wall and the fluid decreases due to a higher fluid
Average Nusselt number and required pumping power as a func-
thermal conductivity; as a consequence the heat transfer rate
tion of Reynolds number are presented in Figs. 7 and 8 for different
increases.
rib configurations and φ = 0.04. Nusselt number profile is a nearly
linear and increasing function of Reynolds number, as shown in Fig. 7.
It can be noted that average Nusselt number for triangular ribs are 6. Conclusions
higher than the ones for trapezoidal ribs.
Trapezoidal ribs require a lower pumping power than square or Results of a numerical investigation on turbulent forced con-
rectangular ribs with w/e = 2.0, as depicted in Fig. 8. The effect of vection in a two-dimensional ribbed channel with Al2O3/water based
rib shapes on heat transfer and pressure drop shown in Figs. 7 and nanofluid are presented in this paper. The investigations are carried
8 is in agreement with the results for pure water in references 20 out for different trapezoidal rib shapes with a dimensionless height
and 21. Moreover, Fig. 8 shows that a fluid with nanoparticles de- equal to 0.05. Ribs are mounted on the heated bottom wall of the
termines larger required pumping powers in comparison with the channel and the effect of pitch-to-height ratio, which ranges from
channel filled with pure water. Pumping power profiles increase 4 to 20, is studied. The mixture model is employed for the inves-
more than linearly as Reynolds number increases. It is also clear how tigated nanofluids whose volume particle concentration ranges from
much pumping power for φ = 0.04 is greater than the ones for pure 0.01 to 0.04. The presence of nanoparticles increases average Nusselt
water. number with respect to pure base fluid but pressure drops are higher
Fig. 9 shows the streamlines for the configurations of ribbed chan- too. The maximum values of both average Nusselt number and re-
nels with ‘rect’ trapezoidal turbulators at p/e = 4 and 10 in a section quired pumping power value are reached for a relative pitch equal
near the channel exit for Re = 50,000 and φ = 0.01. In Fig. 9a the d-type to 10 for all the investigated configurations. A comparison among
roughness behavior is exhibited: a single large recirculating cell is results obtained for different rib shapes shows that trapezoidal ribs
in the region between two adjacent ribs causing a non-efficient heat have lower thermal performances than the triangular ones but this
transfer. The k-type behavior is instead clear in Fig. 9b: this is char- kind of rib shape gives also significantly lower losses. The average
acterized by the separation and reattachment of fluid flow at six- Nusselt number increases as the Reynolds number increases. For
to-eight rib height downstream from it. The core fluid is more the same Reynolds number, average Nusselt number values for tri-
influenced by the presence of ribs and the heat transfer results to angular ribs are higher than the ones for trapezoidal rib but a channel
be enhanced. with triangular ribs needs a greater pumping power. Moreover, the
Fig. 10 shows dimensionless temperature fields for p/e = 10 for average Nusselt number and pumping power values for nanofluids,
Re = 50,000, φ = 0.01, Fig. 10a, and 0.04, Fig. 10b, in the case of with particle volume concentration equal to 0.04, are much greater
‘rect’ trapezoidal ribs. The dimensionless temperature, T*, is defined than the corresponding values for pure water for each investi-
as: gated configuration.
 A. Andreozzi et al./Applied Thermal Engineering 98 (2016) 1044–1053 1051

a)

b)

Fig. 9. Stream function contours, ‘rect’ trapezoidal ribs, φ = 0.01 and Re = 50,000: (a) p/e = 4; (b) p/e = 10.

Nomenclature

A Cross section area [m2]

cp Specific heat [J/kg K]
d Duct diameter [m]
e Rib height [m]
E Total energy [J]
f Friction factor [Eq. (30)]
H Channel height [m]
k Turbulent kinetic energy [kg/(ms3)]
L Duct length [m]
Nu Nusselt number [Eq. (29)]
P Pressure [Pa]
p Rib pitch [m]
PP Required pumping power [W]
Pr = ν/α Prandtl number
q Heat flux [W/m2]
a) Re Reynolds number [Eq. (28)]
s Duct thickness [m]
T Temperature [K]
Tm Bulk temperature [K]
u Velocity component [m/s]
w Rib width [m]
x, y Spatial coordinates [m]

Greek symbols
α Thermal diffusivity [m2/s]
δ Kronecher delta function
λ Thermal conductivity [W/mK]
b)
μ Dynamic viscosity [Pa s]
Fig. 10. Dimensionless temperature fields, ‘rect’ trapezoidal ribs, p/e = 10 and ν Kinematic viscosity [m2/s]
Re = 50,000: (a) φ = 0.01; (b) φ = 0.04. ρ Density [kg/m3]
1052 A. Andreozzi et al./Applied Thermal Engineering 98 (2016) 1044–1053

σ Turbulent Prandtl number [19] V. SriHarsha, S.V. Prabhu, R.P. Vedula, Influence of rib height on the local heat
transfer distribution and pressure drop in a square channel with 90° continuous
τ Wall shear stress [kg/m]
and 60° V-broken ribs, Appl. Therm. Eng. 29 (2009) 2444–2459, doi:10.1016/
φ Nanoparticle concentration j.applthermaleng.2008.12.015.
ω Specific dissipation rate [kg/(m3s2)] [20] O. Manca, S. Nardini, D. Ricci, Numerical investigation of air forced convection
in channels with differently shaped transverse ribs, Int. J. Num. Meth. Heat Fluid
Flow 21 (5) (2011) 618–639, doi:10.1108/09615531111135855.
Subscripts [21] O. Manca, S. Nardini, D. Ricci, Numerical analysis of water forced convection
a Ambient in channels with differently shaped transverse ribs, J. Appl. Math. 2011 (2011)
av Average doi:10.1155/2011/323485 Article number 323485.
[22] O. Manca, S. Nardini, D. Ricci, A numerical investigation on nanofluids forced
bf Base fluid convection in channels with transverse ribs, Proc. 2010 ASME Int. Mech. Eng.
f Fluid Cong. & Exp. (IMECE2010), November 12-18, 2010, Vancouver, British Columbia,
h Hydraulic Canada, paper n. IMECE2010-38745, 2010. doi:10.1115/IMECE2010-38745.
[23] O. Manca, S. Nardini, D. Ricci, A numerical study of nanofluid forced convection
i, j Components in ribbed channels, Appl. Therm. Eng. 37 (2012) 280–292, doi:10.1016/
in Inlet j.applthermaleng.2011.11.030.
nf Nanofluid [24] W. Peng, P.-X. Jiang, Y.-P. Wang, B.-Y. Wei, Experimental and numerical
investigation of convection heat transfer in channels with different types of
out Outlet ribs, Appl. Therm. Eng. 31 (2011) 2702–2708, doi:10.1016/j.applthermaleng.
p Particle 2011.04.040.
s Smooth [25] T. Desrues, P. Marty, J.F. Fourmigué, Numerical prediction of heat transfer and
pressure drop in three-dimensional channels with alternated opposed ribs,
t Turbulent
Appl. Therm. Eng. 45-46 (2012) 52–63, doi:10.1016/j.applthermaleng.2012.
w Wall 03.013.
[26] D. Mikielewicz, A. Stasiek, M. Jewartowski, J. Stasiek, Measurements of heat
transfer enhanced by the use of transverse vortex generators, Appl. Therm. Eng.
49 (2012) 61–72, doi:10.1016/j.applthermaleng.2011.11.013.
References [27] E.Y. Choi, Y.D. Choi, W.S. Lee, J.T. Chung, J.S. Kwak, Heat transfer augmentation
using a rib-dimple compound cooling technique, Appl. Therm. Eng. 51 (2013)
[1] A.E. Bergles, Some perspectives on enhanced heat transfer, second-generation 435–441, doi:10.1016/j.applthermaleng.2012.09.041.
heat transfer technology, J. Heat Transf. 110 (1988) 1082–1096. [28] G. Xie, S. Zheng, W. Zhang, B. Sundén, A numerical study of flow structure and
[2] R.L. Webb, N.H. Kim, Principles of Enhanced Heat Transfer, second ed., Taylor heat transfer in a square channel with ribs combined downstream half-size
& Francis Group, New York, 2006. or same-size ribs, Appl. Therm. Eng. 61 (2013) 289–300, doi:10.1016/
[3] R.L. Webb, Advances in shell side boiling of refrigerants, Proc. Inst. Refrig. 87 j.applthermaleng.2013.07.054.
(1991) 75–86. [29] H.T. Wang, W.B. Lee, J. Chan, S. To, Numerical and experimental analysis of heat
[4] R.L. Webb, G.F. Robertson, Shell-side evaporators and condensers used in the transfer in turbulent flow channels with two-dimensional ribs, Appl. Therm.
refrigeration industry, in: R.K. Shah, E.C. Subbarao, R.A. Mashelkar (Eds.), Heat Eng. 75 (2015) 623–634, doi:10.1016/j.applthermaleng.2014.10.039.
Transfer Equipment Design, Hemisphere Pub. Corp, Washington, 1988, [30] C. Wang, L. Wang, B. Sundén, Heat transfer and pressure drop in a smooth and
pp. 559–570. ribbed turn region of a two-pass channel, Appl. Therm. Eng. 85 (2015) 225–233,
[5] B. Sunden, G. Xie, Gas turbine blade tip heat transfer and cooling: a literature doi:10.1016/j.applthermaleng.2015.03.079.
survey, Heat Transf. Eng. 31 (2010) 527–554, doi:10.1080/01457630903425320. [31] S.U.S. Choi, Enhancing thermal conductivity of fluids with nanoparticles,
[6] R. Karwa, S.C. Solanki, J.S. Saini, Thermo-hydraulic performance of solar air developments and applications of non-Newtonian flows, ASME FED 231 (1995)
heaters having integral chamfered rib roughness on absorber plates, Energy 99–105.
26 (2001) 161–176, doi:10.1016/S0360-5442(00)00062-1. [32] A.R.A. Khaled, K. Vafai, Heat transfer enhancement through control of thermal
[7] L. Syam Sundar, M.K. Singh, Convective heat transfer and friction factor dispersion effects, Int. J. Heat Mass Transf. 48 (2005) 2172–2185, doi:10.1016/
correlations of nanofluid in a tube and with inserts: a review, Renew. Sust. j.ijheatmasstransfer.2004.12.035.
Energy Rev. 20 (2013) 23–35, doi:10.1016/j.rser.2012.11.041. [33] W. Daungthongsuk, S. Wongwises, A critical review of convective heat transfer
[8] X. Liu, Z. Tao, S. Ding, G. Xu, Experimental investigation of heat transfer of nanofluids, Renew. Sust. Energy Rev. 11 (5) (2007) 797–817, doi:10.1016/
characteristics in a variable cross-sectioned two-pass channel with combined j.rser.2005.06.005.
film cooling holes and inclined ribs, Appl. Therm. Eng. 50 (2013) 1186–1193, [34] L. Cheng, E.P. Bandarra Filho, J.R. Thome, Nanofluid two-phase flow and thermal
doi:10.1016/j.applthermaleng.2012.08.001. physics: a new research frontier of nanotechnology and its challenges,
[9] A. Armellini, L. Casarsa, C. Mucignat, Experimental assessment of the aero- J. Nanoscience Nanotech. 8 (7) (2008) 3315–3332, doi:10.1166/jnn.2008.413.
thermal performance of rib roughened trailing edge cooling channels for gas [35] S.U.S. Choi, Nanofluids: from vision to reality through research, J. Heat Transf.
turbine blades, Appl. Therm. Eng. 58 (2013) 455–464, doi:10.1016/ 131 (2009) 1–9, doi:10.1115/1.3056479.
j.applthermaleng.2013.04.065. [36] S. Kakac, A. Pramuanjaroenkij, Review of convective heat transfer enhancement
[10] A. Kumar, M.-H. Kim, Convective heat transfer enhancement in solar with nanofluids, Int. J. Heat Mass Transf. 52 (13–14) (2009) 3187–3196,
air channels, Appl. Therm. Eng. 89 (2015) 239–261, doi:10.1016/ doi:10.1016/j.ijheatmasstransfer.2009.02.006.
j.applthermaleng.2015.06.015. [37] L. Godson, B. Raja, D.M. Lal, S. Wongwises, Enhancement of heat transfer using
[11] G. Xu, Y. Li, H. Deng, Effect of rib spacing on heat transfer and friction nanofluids – an overview, Renew. Sust. Energy Rev. 14 (2) (2010) 629–641,
in a rotating two-pass square channel with asymmetrical 90-deg doi:10.1016/j.rser.2009.10.004.
rib turbulators, Appl. Therm. Eng. 80 (2015) 386–395, doi:10.1016/ [38] R. Saidur, K.Y. Leong, H.A. Mohammad, A review on applications and challenges
j.applthermaleng.2015.02.011. of nanofluids, Renew. Sust. Energy Rev. 15 (3) (2011) 1646–1668, doi:10.1016/
[12] L. Chai, G.D. Xia, H.S. Wang, Numerical study of laminar flow and heat transfer j.rser.2010.11.035.
in microchannel heat sink with offset ribs on sidewalls, Appl. Therm. Eng. 92 [39] W. Yu, D.M. France, E.V. Timofeeva, D. Singh, J.L. Routbort, Comparative review
(2016) 32–41, doi:10.1016/j.applthermaleng.2015.09.071. of turbulent heat transfer of nanofluids, Int. J. Heat Mass Transf. 55 (21–22)
[13] R. Karwa, S.C. Solanki, J.S. Saini, Heat transfer coefficient and friction factor (2012) 5380–5396, doi:10.1016/j.ijheatmasstransfer.2012.06.034.
correlations for the transitional flow regime in rib-roughened rectangular [40] R. Taylor, S. Coulombe, T. Otanicar, P. Phelan, A. Gunawan, W. Lv, et al., Small
ducts, Int. J. Heat Mass Transf. 42 (1999) 1597–1615, doi:10.1016/S0017- particles, big impacts: a review of the diverse applications of nanofluids, J. Appl.
9310(98)00252-X. Phys. 113 (1) (2013) doi:10.1063/1.4754271 art. 011301.
[14] T.M. Liou, J.J. Hwang, S.H. Chen, Simulation and measurement of enhanced [41] F.S. Javadi, R. Saidur, M. Kamalisarvestani, Investigating performance
turbulent heat transfer in a channel with periodic ribs on one principal wall, improvement of solar collectors by using nanofluids, Renew. Sust. Energy Rev.
Int. J. Heat Mass Transf. 36 (1993) 507–517, doi:10.1016/0017-9310(93)80025-P. 28 (2013) 232–245, doi:10.1016/j.rser.2013.06.053.
[15] G. Rau, M. Cakan, D. Moeller, T. Arts, The effect of periodic ribs on the local [42] A. Celen, A. Çebi, M. Aktas, O. Mahian, A.S. Dalkilic, S. Wongwises, A review of
aerodynamic and heat transfer performance of a straight cooling channel, ASME nanorefrigerants: flow characteristics and applications, Int. J. Refr. 44 (2014)
J. Turbomach. 120 (1998) 368–375. 125–140, doi:10.1016/j.ijrefrig.2014.05.009.
[16] C.K. Lee, S.A. Abdel Moneim, Computational analysis of heat transfer in turbulent [43] A.M. Hussein, K.V. Sharma, R.A. Bakar, K. Kadirgama, A review of forced
flow past a horizontal surface with 2-D ribs, Int. Commun. Heat Mass Transf. convection heat transfer enhancement and hydrodynamic characteristics of a
28 (2001) 161–176, doi:10.1016/S0735-1933(01)00223-8. nanofluid, Renew. Sust. Energy Rev. 29 (2014) 734–743, doi:10.1016/
[17] D.N. Ryu, D.H. Choi, V.C. Patel, Analysis of turbulent flow in channels roughened j.rser.2013.08.014.
by two-dimensional and three-dimensional Blocks. Part I: resistance, Int. J. Heat [44] R.A. Mahdi, H.A. Mohammed, K.M. Munisamy, N.H. Saeid, Review of convection
Fluid Flow 28 (2007) 1098–1111, doi:10.1016/j.ijheatfluidflow.2006.11.006. heat transfer and fluid flow in porous media with nanofluid, Renew. Sust. Energy
[18] D.N. Ryu, D.H. Choi, V.C. Patel, Analysis of turbulent flow in channels roughened Rev. 41 (2015) 715–734, doi:10.1016/j.rser.2014.08.040.
by two-dimensional and three-dimensional blocks. Part II: heat transfer, Int. [45] S.K. Verma, A.K. Tiwari, Progress of nanofluid application in solar collectors:
J. Heat Fluid Flow 28 (2007) 1112–1124, doi:10.1016/j.ijheatfluidflow.2006. a review, Energy Conv. Manag. 100 (2015) 324–346, doi:10.1016/
11.007. j.enconman.2015.04.071.
 A. Andreozzi et al./Applied Thermal Engineering 98 (2016) 1044–1053 1053

[46] H.A. Mohammed, A.N. Al-Shamani, J.M. Sheriff, Thermal and hydraulic [53] V. Bianco, S. Nardini, O. Manca, Enhancement of heat transfer and entropy
characteristics of turbulent nanofluids flow in a rib-groove channel, Int. generation analysis of nanofluids turbulent convection flow in square
Commun. Heat Mass Transf. 39 (10) (2012) 1584–1594, doi:10.1016/ section tubes, Nanoscale Res. Lett. 6 (2011) doi:10.1186/1556-276X-6–252
j.icheatmasstransfer.2012.10.020. art. n. 252.
[47] M. Gavara, Asymmetric forced convection of nanofluids in a channel with [54] M. Mahdavi, M. Sharifpur, J.P. Meyer, CFD modelling of heat transfer and
symmetrically mounted rib heaters on opposite walls, Num. Heat Transf. Part pressure drops for nanofluids through vertical tubes in laminar flow by
A Appl. 62 (11) (2012) 884–904. Lagrangian and Eulerian approaches, Int. J. Heat Mass Transf. 88 (2015) 803–813,
[48] M. Parsazadeh, H.A. Mohammed, F. Fathinia, Influence of nanofluid on turbulent doi:10.1016/j.ijheatmasstransfer.2015.04.112.
forced convective flow in a channel with detached rib-arrays, Int. Commun. [55] F.R. Menter, Two equation eddy-viscosity turbulence models for engineering
Heat Mass Transf. 46 (2013) 97–105, doi:10.1016/j.icheatmasstransfer. applications, AIAA J. 32 (1994) 1598–1605.
2013.05.006. [56] Ansys Incorporated, Fluent 15.0 User Manual, 2014.
[49] A. Vatani, H.A. Mohammed, Turbulent nanofluid flow over periodic rib-grooved [57] W.M. Rohsenow, J.P. Hartnett, Y.I. Cho, Handbook of Heat Transfer, third ed.,
channels, Eng. Appl. Comput. Fluid Mech. 7 (3) (2013) 369–381. McGraw-Hill, New York, 1998.
[50] A.N. Al-Shamani, K. Sopian, H.A. Mohammed, S. Mat, M.H. Ruslan, A.M. Abed, [58] B.C. Pak, Y.I. Cho, Hydrodynamic and heat transfer study of dispersed fluids with
Enhancement heat transfer characteristics in the channel with trapezoidal submicron metallic oxide particles, Exp. Heat Transf. 11 (1998) 151–170.
rib-groove using nanofluids, Case Stud. Therm. Eng. 5 (2015) 48–58, [59] S.E.B. Maiga, C.T. Nguyen, N. Galanis, G. Roy, Heat transfer behaviours of
doi:10.1016/j.csite.2014.12.003. nanofluids in a uniformly heated tube, Superlattices Microstruct. 35 (2004)
[51] Y.-T. Yang, H.-W. Tang, B.-Y. Zeng, C.-H. Wu, Numerical simulation and 543–557, doi:10.1016/j.spmi.2003.09.012.
optimization of turbulent nanofluids in a three-dimensional rectangular [60] S.E.B. Maiga, N. Cong Tam, N. Galanis, G. Roy, T. Mare, M. Coqueux, Heat transfer
rib-grooved channel, Int. Commun. Heat Mass Transf. 66 (2015) 71–79, enhancement in turbulent tube flow using Al2O3 nanoparticle suspension,
doi:10.1016/j.icheatmasstransfer.2015.05.022. Int. J. Num. Meth. Heat Fluid Flow 16 (2006) 275–292, doi:10.1108/
[52] S. Mirmasoumi, A. Behzadmehr, Numerical study of laminar mixed 09615530610649717.
convection of a nanofluid in a horizontal tube using two-phase mixture model, [61] O. Manca, P. Mesolella, S. Nardini, D. Ricci, Numerical study of a confined slot
Appl. Therm. Eng. 28 (7) (2008) 717–727, doi:10.1016/j.applthermaleng. impinging jet with nanofluids, Nanoscale Res. Lett. 6 (2011) doi:10.1186/1556-
2007.06.019. 276X-6-188 art. n. 188.