International Communications in Heat and Mass Transfer 91 (2018) 239–247

Contents lists available at ScienceDirect

International Communications in Heat and Mass Transfer

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

Two-phase numerical simulation of hybrid nanofluid heat transfer in T

minichannel heat sink and experimental validation
⁎
Vivek Kumar, Jahar Sarkar
Department of Mechanical Engineering, Indian Institute of Technology (B.H.U.), Varanasi, UP 221005, India

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

Keywords: Nanofluid cooled mini-micro channel heat sink has become a pleasant alternative for electronics and thermal
Mini/microchannel heat sink applications recently due to its compactness and enhanced heat transfer characteristics. In the present study, a
Hybrid nanofluid numerical simulation on laminar forced convection flow of DI water based Al2O3 nanofluid and Al2O3-MWCNT
Mixture model hybrid nanofluid in minichannel heat sink has been performed using two-phase mixture model to investigate the
Multi-walled carbon nanotube
heat transfer and pressure drop characteristics. The experimental study on hybrid nanofluid flow in minichannel
Experimentation
Heat transfer coefficient
heat sink has also been conducted and validated the numerical model. Effect of some important parameters, such
as, hydraulic diameter, channel aspect ratio, composition of Al2O3 and MWCNT in hybrid nanofluid and Reynold
number has been investigated as well. Two-phase (heterogeneous) model has good agreement with the ex-
periment result as compared to single phase (homogenous) approach. Maximum heat transfer coefficient has
been found for 0.01 vol% (Al2O3 + MWCNT) (7:3) hybrid nanofluid for minichannel depth of 0.5 mm. Pressure
drop has been found maximum for minichannel of 0.5 mm channel depth. The developing length can increase by
using nanofluid. Maximum heat transfer coefficient improvement of 15.6% has been observed with no appre-
ciable increment in pressure drop by using hybrid nanofluids.

Nomenclatures vof volume fraction

wch channel width (mm)
A effective heat transfer area (m2)
Ar channel aspect ratio Greek symbols
cp specific heat (J·kg− 1 K− 1)
dh hydraulic diameter (mm) μ dynamic viscosity (Pa·s)
f friction factor ρ density (kg·m− 3)
h heat transfer coefficient (W·m− 2 K− 1) φ volume concentration
hch channel height (mm)
k thermal conductivity (W·m− 1 K− 1) Subscripts
ṁ mass flow rate (kg·s− 1)
Nu Nusselt number bf base fluid
p pressure (Pa) ch channel
Pr Prandtl number eff effective
Q heat transfer rate (W) f fluid
Re Reynolds number in inlet
t fin width (mm) l channel length
t1 thickness of asbestos plate (mm) m mixture
t2 distance from channel base to minichannel base (mm) nf nanofluid
T temperature (°C) np,p nanoparticle
u,v,w,V velocity (m·s− 1) out outlet

Abbreviation: MWCNT, multi walled carbon nanotube; MCHS, mini/micro channel heat sink
⁎
Corresponding author.
E-mail address: jsarkar.mec@itbhu.ac.in (J. Sarkar).

https://doi.org/10.1016/j.icheatmasstransfer.2017.12.019

0735-1933/ © 2017 Elsevier Ltd. All rights reserved.

V. Kumar, J. Sarkar
s surface
1. Introduction
Use of liquid-cooled Mini/Micro Channel Heat Sink (MCHS) is the
effective way to tackle high heat dissipation in space constraint elec-
tronic devices and keep them in their desired designation for good
functionality. Due to increasing demand of energy density, the further
improvement in term of higher heat dissipation capability can be ob-
tained by using liquid of better thermophysical properties and nano-
fluids [1] as well as hybrid nanofluids [2] have emerged as a viable
candidate for their better heat transfer characteristics. In much prac-
tical application, it is required to adjust between different properties
because a single material does not hold all the favorable characteristics
required for a particular purpose, so hybrid terms come here [2]. Alu-
mina nanoparticle is widely used in nanofluids due to low cost, avail-
ability, chemical stability and higher heat transfer improvement to
pumping power increase ratio [3]. On the other hand, multi-walled
carbon nanotube (MWCNT) has attracted many researchers due to their
higher thermal conductivity and very high aspect ratio for application
in nanofluids [4]. Hence, alumina and MWCNT can be the best com-
bination in hybrid nanofluid for MCHS applications. Although, the ex-
perimental studies on heat transfer and fluid flow characteristics of
nanofluid in MCHS have been conducted by many investigators [5–12],
whereas, investigations on hybrid nanofluids flow in MCHS are very
limited in the open literature [2,13].
Numerical simulation of nanofluid flow in MCHS is still challenging
issue and many models can be applied to capture the complex me-
chanisms for heat transfer enhancement. Single-phase (homogeneous)
model has been adopted in many investigations on MCHS with various
nanofluids and hybrid nanofluids [14–18]. However, as the homo-
geneous model is unable to capture the heat transfer mechanism of
nanofluids accurately, two-phase mixture model has been adopted re-
cently for numerical simulation of nanofluids in MCHS. Kalteh et al.
[19] simulated steady state, laminar flow and constant wall tempera-
ture conditions for Cu-water nanofluids in microchannels using Eu-
lerian-Eulerian two phase flow approach. Their result reveals that
mixture model is in good agreement with experimental data and mix-
ture model theory is possibly applied to nanofluids flow in micro-
channels for laminar flow. Kalteh et al. [20] also studied the laminar
convective heat transfer characteristics of an Al2O3-water nanofluid
studied inside a wide rectangular microchannel heat sink both nu-
merically by adopting two-phase Eulerian-Eulerian method using the
finite volume approach and experimentally. Their study reveals that the
two-phase Eulerian- Eulerian method results are in better agreement
with experimental results than the homogeneous (single-phase) model.
Moraveji and Ardehali [21] did a CFD modeling of laminar forced
convection on Al2O3 nanofluid in mini-channel heat sink using four
models (single phase, VOF, mixture, Eulerian). It found that two-phase
models were more precise than single phase model when compared
with experimental reference data. Single and two phase model has been
used by Naphon and Nakharintr [22] to study laminar convective heat
transfer of TiO2 nanofluids in a minichannel heat sink. Two phase
model results are closer to experimental result when comparing to
single phase model. Esmaeilnejad et al. [23] numerically investigated
the convection heat transfer and laminar flow of nanofluids in micro-
channel using two-phase mixture model. The result showed that the
thermal resistance reduces about 27.2% and pressure drop increases
approximately 50.7% with Peclet number (Pe) of 700 and 4 vol%
concentration of nanoparticles with using shear thinning non-New-
tonian base fluid. Recently, Ambreen and Kim [24] compared various
numerical models (Homogeneous, discrete phase and Eulerian-Eulerian
(Mixture, Volume of Fluid, Eulerian)) to investigate laminar forced
convection of Al2O3 + water and TiO2 + water nanofluids in the mini/
micro channels and found Eulerian-Eulerian models is better as com-
pared to homogenous model. Hence, the two phase model study carried
240
International Communications in Heat and Mass Transfer 91 (2018) 239–247
the numerical calculations at least one step forward in terms of simu-
lating the physics of nanofluids. The mixture model has been success-
fully used in numerical simulation of nanofluids in other heat transfer
application also [25–29]. Hence, two phase mixture model has been
employed in this study to simulate the hybrid nanofluids in MCHS.
However, the studies on numerical simulation of hybrid nanofluids
flow in MCHS using mixture model is very limited. Nimmagadda &
Venkatasubbaiah [30] employed two-phase mixture model to analyzed
the heat transfer characteristics of (Al2O3 + Ag) hybrid nanofluid at
different Reynold number and particle volume concentrations in a wide
rectangular micro-channel. Two-phase mixture model has been used by
Nimmagadda & Venkatasubbaiah [31] also for the numerical in-
vestigation to analyze the performance of microchannel under forced
convection laminar flow using hybrid nanofluid (Cu + Al, water
+ methanol). However, a single geometry and single nanoparticles
composition were considered in their study. Hence, with best of the
authors' knowledge, no numerical simulation of hybrid nanofluids in
MCHS using mixture model to consider geometric and nanoparticle
composition (ratio of different nanoparticles) effects is not available in
open literature.
The objective of present study is to numerically analyze the per-
formance of minichannel heat sink of different geometries using Al2O3-
MWCNT/DI water hybrid nanofluids using the mixture model. The
experimental study has been conducted on minichannel heat sink to
validate the numerical model. Uniform and constant heat fluxes have
been applied from the bottom and all other surfaces are adiabatic for no
heat losses. Effects of aspect ratio, hydraulic diameter, flow rate, tem-
perature and nanoparticle composition have been discussed. Velocity
and temperature profile along the channel length has also been studied
as well using mixture model.
2. Formulation of the problem
Water, Al2O3/water nanofluids and Al2O3-MWCNT/water hybrid
nanofluids have been used as a coolant in a minichannel heat sink
(MCHS). The volume fraction of alumina used in Al2O3/water nanofluid
is 0.01%. In hybrid nanofluid, Al2O3 and MWCNT have been mixed in
three different ratios (9:1, 8:2 and 7:3) with the total volume fraction of
0.01%. The numerical analysis has been performed at the constant heat
flux of 8.3 W/cm2 and at the laminar flow regime 50 < Re < 500.
The minichannel heat sink is made of aluminum, covered by an
adiabatic acrylic plate on top. A schematic of the structure of a rec-
tangular minichannel heat sink is shown in Fig. 1. The bottom surface of
the heat sink is uniformly heated using heater supplied with AC current.
Three different geometries are considered for numerical study based on
the channel height taking 3 mm, 1 mm and 0.5 mm (aspect ratio 3, 1
and 2, respectively). Minichannel heat sink consists of 9 parallel rec-
tangular shaped minichannels having length 30 mm, having channel
width and fin width of 1 mm each. It may be noted that the mini-
channels are supposed to be identical in terms of both heat transfer and
hydrodynamics.
3. CFD methodology
3.1. Governing equations for single phase model
Assuming incompressible, Newtonian and laminar flow thorough
MCHS, the dimensional governing equations for steady state condition
(continuity, momentum, and energy) using the single phase model are
as follows:
Continuity equation
∂u ∂v ∂w
+ + =0
∂x ∂y ∂z (1)
x momentum equation:
V. Kumar, J. Sarkar International Communications in Heat and Mass Transfer 91 (2018) 239–247

Fig. 1. Mini/microchannel heat sink

(w = wch + t).

∂u ∂u ∂u ∂p ∂ 2u ∂ 2u ∂ 2u The viscosity of the nanofluid or hybrid nanofluid is calculated

ρf ⎛⎜u +v + w ⎞⎟ = − + μ f ⎛⎜ 2 + + 2 ⎞⎟
∂x ∂y ∂z ∂x ∂ x ∂ y 2 ∂z ⎠ (2) using Batchelor relation [34] which was proposed for nanofluids with
⎝ ⎠ ⎝
the high volumetric concentration in which hydrodynamic interactions
y momentum equation: and nanoparticle aggregation are important.
∂v ∂v ∂v ∂p ∂ 2v ∂ 2v ∂ 2v μ eff = (1 + 2.5φ + 6.2φ2)μ bf (9)
ρf ⎛⎜u +v + w ⎟⎞ = − + μ f ⎛⎜ 2 + 2 + 2 ⎞⎟
⎝ ∂x ∂y ∂z ⎠ ∂y ⎝ ∂ x ∂ y ∂z ⎠ (3)
Hamilton and Crosser [35] developed one of the basic models for
z momentum equation: the prediction of thermal conductivity of nanofluids: (consider sphe-
rical nanoparticles shape factor is equal to 3)
∂w ∂w ∂w ⎞ ∂p ∂ 2w ∂ 2w ∂ 2w ⎞
ρf ⎛⎜u +v +w ⎟ = − + μ f ⎛⎜ 2 + 2
+ ⎟
∂x ∂y ∂z ∂z ∂ x ∂ y ∂z2 ⎠
⎝ ⎠ ⎝ (4)
k eff
=
2+ ( ) + 2φ ⎡⎣ ( ) − 1⎤⎦
knp
k bf
knp
k bf
Energy equation for fluid: kbf
∂T ∂T ∂T ∂ 2T ∂2Tf ∂2Tf ⎞
2+ ( ) − φ ⎡⎣ ( ) − 1⎤⎦
knp
k bf
knp
k bf (10)
ρf c pf ⎛⎜u f + v f + w f ⎞⎟ = kf ⎛⎜ 2f + 2
+ ⎟

⎝ ∂x ∂y ∂z ⎠ ⎝ ∂x ∂y ∂z2 ⎠ (5) Similarly, thermal conductivity of hybrid nanofluid has been cal-
culated by [36],
Energy equation for the solid wall
(1/φ) ∑ φnp knp + 2kbf + 2 ∑ φnp knp − 2φkbf
∂ 2T ∂2Ts ∂2Ts ⎞ k eff
0 = kS ⎜⎛ 2s + 2
+ ⎟ =
np np

⎝ ∂x ∂y ∂z2 ⎠ (6)
In this work, if a nanofluid is used as the coolant, the thermo-
physical properties in the governing equations are replaced by those of
the nanofluid. That is, the effective density ρf, heat capacity cpf, dy-
namic viscosity μf, and thermal conductivity kf in the governing equa-
tions are replaced by ρeff, cp,eff, μeff, and keff, respectively which were
given by Eqs. (7)–(12).
Nanofluid properties are normally expressed in nanoparticle volume
fraction φ. The density of nanofluid or hybrid nanofluid has been cal-
culated from [32]:
ρeff = (1 − φ)ρ bf + ∑ φnp ρnp
np (7)
where, np and bf refer to nanoparticles and base fluid, respectively. For
nanofluid, np = 1 and for hybrid nanofluid, np = 2, φ is the overall
volume concentration of two different types of nanoparticles dispersed
in hybrid nanofluid and is calculated as: φ = φnp1 + φnp2.
Similarly, the specific heat of nanofluid has been defined as [33],
kbf (1/φ) ∑ φnp knp + 2kbf − ∑ φnp knp + φkbf
np np (11)
3.2. Governing equation for Multiphase Mixture Model
The mixture model can simulate multi-phases (fluid or particulate)
by solving the momentum, continuity and energy equations for the
mixture, the volume fraction equations for the secondary phases and
algebraic expressions for the relative velocities. The mixture model
allows selecting granular phases and calculating all properties of the
granular phases. This can be applicable for liquid-solid flows. The
mixture model is based on the following assumptions as given below:
i) Same pressure is shared by all phases.
ii) Nanoparticles are spherical and uniform in size and shape.
iii) The interactions between different dispersed phases in nanofluid
are neglected.
iv) Fluid flow is incompressible, Newtonian and laminar.

ρeff c p,eff = (1 − φ)ρ bf c p,bf + ∑ φnp ρnp c p,np Based on the above assumptions, steady-state governing equations
np (8)
of fluid flow and heat transfer for the two-phase mixture (n = 2) model

241
V. Kumar, J. Sarkar International Communications in Heat and Mass Transfer 91 (2018) 239–247

Table 1 condition as velocity inlet and outflow, respectively. No slip boundary

Effect of grid size on heat transfer coefficient and pressure drop. condition is applied on the channels walls. The constant heat flux
(83 kW/m2) is applied to the bottom surface and all others surfaces of
Grid type Number of elements Heat transfer coefficient Pressure drop
(W/m2 K) (Pa) minichannel heat sink is assumed to be adiabatic.
At the microchannel inlet (x = 0),
Coarse 1,840 1002 20.0
Medium 48,942 1949 47.4
Vx = 0, Vy = V, Vz = 0, T = Tin
Fine 90,768 1956 54.9
Very Fine 9,26,250 1961 58.7
At the exit (x = l), the pressure is assumed to be atmospheric
1331658 1965 61.4 pressure.

3.4. Numerical method and Grid Independence Test

are as follows [37–39]:
Continuity equation: The computational fluid dynamics (CFD) calculations are done by
→ using FLUENT (ANSYS 18). A two phase Mixture Model Theory is used
∇⋅(ρm Vm) = 0 (12) for simulations. The First order upwind numerical scheme and SIMPLE
Momentum equation (Navier-Stokes equations) algorithm are used to discretize the governing equations. The conver-
ging criterions are taken as 10− 6 for all the parameters. The appro-
n
→→ → ⎛ → → ⎞ priate quality and quantity of grids are very important in order to en-
∇⋅(ρm Vm Vm) = −∇pm + ∇⋅(μ m∇ Vm) + ∇⋅⎜ ∑ φk ρk Vdr,k Vdr,k⎟
⎝ k = 1 ⎠ (13) sure the accuracy and time of CFD numerical computation. In this
study, the distribution of quadrilateral cells in the computational do-
Energy equation: main has been determined from tests with the different number of cells
n
→ for the grid independence test. The grid used have number of elements
⎡
∇⋅⎢ ∑ φk Vk (ρk Hk + pm] = ∇⋅(Km ∇T) of 1,840 (coarse), 48,942 (intermediate), 90,768 (fine), 9,26,250 (very
⎣ k=1 (14) fine) and 1331658. The results of the average heat transfer coefficient
Volume fraction equation: and pressure drop for different number of elements for water in a
→ → minichannel of cross-section area of dimensions 1 mm × 1 mm are
∇⋅(φnp ρnp Vm) = −∇⋅(φnp ρnp Vdr,p) (15) presented in Table 1. It has been observed that, there is very less dif-
ference (0.20%) in the calculated results after further increasing from
where mixture velocity, density and viscosity are, respectively,
grid 9,26,250 to 1331658 grid. As a result, by comparing time usage
n → and accuracy, the very fine grid (926250) has been selected for the
→ ∑ φk ρk Vk
Vm = k = 1 simulation.
ρm (16)
n 4. Experimental methodology
ρm = ∑ φk ρk
k=1 (17)
4.1. Experimental setup
n
μm = ∑ φ k μk The schematic diagram of the experimental system is illustrated in
(18)
k=1
Fig. 2. Four main units exist in the system; flow loop, cooling, heating
The drift velocity of kth phase is and measuring unit. The flow loop consists of pipes, a test section, a
→ → → flow meter and a pump. An in-built positive displacement pump at-
Vdr,k = Vk − Vm (19) tached to the constant temperature bath with an inverter, which is
The slip velocity (relative velocity) is defined as the velocity of a suitable for certain flow range, has been used. The flow rate can be
secondary phase (p) relative to the velocity of the primary phase (f): adjusted in the range of 0–40 LPH. The flow rates have been measured
using a variable area flowmeter (CVG Technocrafts India) with a
→ ⎯→
⎯ ⎯→
⎯
Vpf = Vp − Vf (20) measuring range of 16–160 LPH. A water bath circulator has been used
→ just before the test section so that the working fluid can enter the test
The drift velocity ( Vdr,k ) is related to the relative velocity as: section at the inlet conditions. The test section includes heater car-
n ⎯⎯⎯→ tridges, top and bottom plates, thermocouples, and acrylic plate ma-
→ → ∑ φk ρk Vfk
Vdr,p = Vpf − k = 1 terials. The acrylic plate was used as an insulator to cover the heat sink
ρm (21) from top. Five holes are drilled into the top of the acrylic plate for
temperature measurements along the length of the heat sink. Likewise,
→ ρp d 2p (ρp − ρm ) →
Vpf = a two holes are also drilled onto the bottom plexiglass to measure the
18μ mf drag ρp (22) pressure drop between channel inlet and outlet by using a differential
The drag function is determined by Schiller and Naumann [40], pressure transducer.
In order to get constant heat flux boundary condition, a cartridge
0.687
⎧ 1 + 0.15Re p Re p≤1000 heater, has a power of 50 W, with a width of 10 mm and a length of
f drag =
⎨ 0.0183Re p Re p > 1000 (23) 30 mm are placed into the bottom of the heat sink. The cartridge hea-
⎩
ters are connected to a variac transformer, and a constant heat flux of
→ → →
where Rep = (Vmdnp)/νm and a = g − (Vm ∙∇) Vm. 83 kW/m2 is applied to the bottom surface of the channel. The ex-
Where, g is gravitational acceleration. periments were first conducted by using distilled water and then using
Al2O3–MWCNT hybrid nanofluids. The fluid taken from the supply re-
3.3. Boundary condition servoir into the system at a constant inlet temperature of 300 K was
driven by the pump, and the heat energy was supplied to the system
The nanofluid enters the channel with uniform velocity and tem- after the maximum Reynolds number was reached. The average of the
perature with the assumption that both the base fluid and particles have readings taken from the five surface thermocouples was used as the
the same velocity. Channel inlet and the channel outlet have boundary surface temperature in the calculations. After reaching steady state

242
V. Kumar, J. Sarkar International Communications in Heat and Mass Transfer 91 (2018) 239–247

T1
T2 Data
T3 Acquisition
T4
W P T5 System
Variac
T6

Ts1
Ts2
Test module
(mini/micro
channel)
Valve
Copper plate,
Constant Heater
Temperature
Bath

Flowmeter
Valve

W-Wattmeter, P-Digital Pressure Gauge, T-Thermocouple

Pump
Fig. 2. Schematic diagram of experimental work.

conditions, i.e. thermal equilibrium, in 45–50 min, all measured tem- The uncertainties in percentage during the measurements of the
peratures were recorded. experimental parameters such as inlet temperature, outlet temperature,
thermal conductivity, density, specific heat, viscosity, and flow rate
4.2. Data analysis and experimental uncertainties are ± 0.33, ± 0.33, ± 2.5, ± 2.0, ± 2.0, ± 3.0 and ± 0.5 respec-
tively. Based on the measured parameters, the total uncertainties found
By measuring flow rate, fluid inlet and outlet temperatures, the heat for estimated convection heat transfer coefficient, Nusselt number and
transfer has been calculated by, Reynolds number are ± 3.2%, ± 4.1% and ± 3.6% respectively.
Q = ṁ c p (Tout − Tin) (24)
5. Results and discussion
Thus, the average convection heat transfer coefficient can be given
by, 5.1. Experimental validation
Q
h=
A(Ts − Tm ) (25) The variation of surface heat transfer coefficient with inlet velocity
for different working fluids has been compared with the experimental
where, Ts and Tm are average wall temperature and mean fluid tem- results. It was seen that the heat transfer coefficient increases with inlet
perature, respectively. Then the Nusselt number is calculated as below: velocity. The increase in heat transfer coefficient with the addition of
hdh nanoparticles which was nearly 9% for 0.01%vof Al2O3-MWCNT (9:1).
Nu =
k (26) It was seen that the Nusselt number increases with increase in both
Reynolds number and addition of nanoparticles. This increase in
The Reynolds number is defined by using,
Nusselt number is due to increase in heat transfer coefficient.
ρu m dh The Tables 2 and 3 compare the experimental result for water and
Re =
μ (27) 0.01%vof Al2O3-MWCNT Hybrid Nanofluid (9:1) with the numerical
results. The average deviation of numerical results from experimental
By measuring the flow rate and pressure drop, the friction factor is
results for water for heat transfer coefficient was 5.9%. From the table it
given by,
can be seen that average deviation of single phase model from the ex-
2 perimental results for 0.01%vof Al2O3-MWCNT Hybrid Nanofluid was
l ρu m
Δp = f 5.6% and for multi-phase mixture model it was about 4.5%, so it can be
dh 2 (28)
concluded that multi-phase mixture model gives more accurate results
During experiments with MCHS, the temperatures, flow rates and than the single phase model as similar to previous studies. Thus further
pressure loss have been measured with appropriate instruments. The investigation has only been carried out on multi-phase mixture model
following Kline and McClintock [41] equation has been used for the at the inlet temperature of 30 °C.
uncertainty analysis of the convection heat transfer coefficient, Nusselt
number, friction factor and Reynolds number.
5.2. Numerical results based on multi mixture model
2 2 2 1/2
W=⎡ ⎛ ∂R ⎞ ⎛ ∂R ⎞ ⎛ ∂R ⎞ ⎤
⎢ ∂x1 w1 + ∂x2 w2 + …………….+ ∂xn wn ⎥ The variations of surface heat transfer coefficient with inlet fluid
⎜ ⎟ ⎜ ⎟ ⎜ ⎟

⎣ ⎝ ⎠ ⎝ ⎠ ⎝ ⎠⎦ (29)
velocity for different working fluids are shown in Fig. 3 for the aspect

243
V. Kumar, J. Sarkar International Communications in Heat and Mass Transfer 91 (2018) 239–247

Table 2
Comparison of experimental and numerical results for heat transfer coefficient.

Working fluid Reynolds number Experimental result Model

(W/m2 K)
Single phase Deviation Multi-phase mixture Deviation
W/m2 K W/m2 K

Water 307 3033 3140 3.6% _ _

368 3081 3416 10.9% _ _
430 3260 3556 9.1% _ _
Al2O3-MWCNT Nanofluid (9:1) 307 3433 3205 6.7% 3534 2.9%
368 3577 3473 3.0% 3820 5.8%
430 3816 3558 6.8% 3941 3.3%

ratio of 1, 3, and 2. It can be observed that the heat transfer coefficient

increases with increase in inlet velocity due to thickening of the thermal
boundary layer. The maximum value of heat transfer coefficient is 4000
5205.5 W/m2 K for Al2O3-MWCNT (7:3)/water hybrid nanofluid for

h (W/m2K)
0.5 mm (Ar = 2) channel height. It is well-known fact that the heat
transfer coefficient increases with the decrease in hydraulic diameter 3000 Al2O3 nf 0.01%vof
and hence it is maximum for 0.5 mm channel height. There is found an Al2O3-MWCNT vof 0.01% (9:1)
Al2O3-MWCNT 0.01%vof (8:2)
increment in heat transfer coefficient with the addition of nano- Al2O3-MWCNT 0.01%vof (7:3)
particles, which may be due to several reasons including an increase in 2000 water
thermal conductivity, different slip mechanisms and the nano-porous 0.09 0.14 0.19 0.24 0.29
Inlet velocity (m/sec)
and nano-fin effects [3]. It is interesting to note that the heat transfer
coefficient increases with increase in the ratio of MWCNT for same total (a)
volume fraction (0.01%) and the reason for increment in heat transfer
4800
coefficient is mainly due to high high thermal conductivity of MWCNT.
Within the studied range, the maximum improvement in heat transfer
coefficient has been observed as nearly 4% for 0.01%vof Al2O3 nano- 3800
h (W/m2K)

fluid, 9% for 0.01%vof Al2O3-MWCNT (9:1), 12% for 0.01%vof Al2O3-

MWCNT (8:2) and 15.6% for 0.01%vof Al2O3-MWCNT (7:3) hybrid water
nanofluids. 2800 Al2O3 0.01%vof
Al2O3-MWCNT 0.01% vof (9:1)
The variation of Nusselt number with Reynolds number is shown in Al2O3-MWCNT 0.01%vof (8:2)
Fig. 4 for different aspect ratios. It is well-known fact that the Nusselt Al2O3-MWCNT 0.01%vof (7:3)
1800
number increases with increase in Reynolds number (as predicted in 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
figures). An increase in Nusselt number can be observed in addition of Inlet velocity (m/sec)
nanoparticles and this increase in Nusselt number is due to increase in
(b)
heat transfer coefficient as discussed above. From the above graphs, it
can be seen that there is a decrease in the Nusselt number value with 5600
the decrease in hydraulic diameter. This can be explained with help of 5100
definition of the Nusselt number as in Eq. (26). Thus if hydraulic dia- 4600
h (W/m2K)

meter decreases there will be a decrease in the value of the Nusselt 4100
number. Again, the Nusselt number increases with increase in the ratio 3600 water
Al2O3 0.01%vof
of MWCNT for same total nanoparticle volume fraction. As shown in 3100 Al2O3-MWCNT 0.01%vof (9:1)
figures, the hybrid nanofluid shows very similar behavior with base 2600
Al2O3-MWCNT 0.01% vof (8:2)
Al2O3-MWCNT 0.01% vof (7:3)
fluid, although dissimilar to the nanofluids the Nusselt number for 2100
hybrid nanofluids is strongly dependent on the volume ratio of nano- 0.1 0.2 0.3 0.4 0.5 0.6
particles also. Inlet velocity (m/sec)

The variations of pressure drop with Reynolds number for different

working fluids showing the effect of aspect ratio are shown in Fig. 5. It
can be observed that there is a significant increase in pressure drop with
the decrease in hydraulic diameter, and the difference of increase in
pressure drop also increases significantly with increase in Reynolds
(c)
Fig. 3. Variation of heat transfer coefficient with inlet velocity for channel height of (a)
3 mm (Ar = 3), (b) 1 mm, (Ar = 1), (c) 0.5 mm (Ar = 2).

Table 3
Comparison of experimental and numerical results for Nusselt number.

Working fluid Reynolds number Parameter Experimental result Model

Single phase Deviation Multiphase mixture Deviation

Water 307 Nusselt number 7.58 7.83 3.3% _ _

368 7.70 8.55 11.1% _ _
430 8.15 8.89 9.1% _ _
Al2O3-MWCNT Nanofluid (9:1) 307 Nusselt number 8.52 7.92 7.0% 8.74 2.5%
368 8.85 8.61 2.7% 9.49 6.2%
430 9.44 9.00 4.7% 9.76 3.3%

244
V. Kumar, J. Sarkar International Communications in Heat and Mass Transfer 91 (2018) 239–247

11 1200
10 water 3mm channel height
1000

Pressure drop (Pa)

9 water 1mm channel height
800 water 0.5mm channel height
8
Nu

600
7 water
0.01% vof Al2O3 400
6
0.01% vof Al2O3-MWCNT (9:1)
5 0.01% vof Al2O3-MWCNT (8:2) 200
0.01%vof Al2O3-MWCNT (7:3)
4 0
50 100 150 200 250 300 350 400 450 50 100 150 200 250 300 350 400
Re Re

(a) (a)

8 1400
Al2O3 3mm channel height
1200

Pressure drop (Pa)

7 Al2O3 1mm channel height
1000
6 Al2O3 0.5mm channel height
800
Nu

5 600
water
4 0.01%vof Al2O3 400
0.01%vof Al2O3-MWCNT (9:1)
3 0.01%vof Al2O3-MWCNT (8:2) 200
0.01%vof Al2O3-MWCNT (7:3) 0
2 50 100 150 200 250 300 350 400
50 100 150 200 250 300 350 400
Re Re

(b) (b)

6 1400
Al2O3-MWCNT (9:1) 3mm
5.5 1200
Pressure drop (Pa)

Al2O3-MWCNT (9:1) 1mm

5 1000
Al2O3-MWCNT (9:1) 0.5mm
4.5
800
Nu

4 water 600
3.5 0.01%vof Al2O3
0.01%vof Al2O3-MWCNT (9:1) 400
3
0.01%vof Al2O3-MWCNT (8:2) 200
2.5 0.01%vof Al2O3-MWCNT (7:3)
2 0
50 100 150 200 250 300 350 400 50 100 150 200 250 300 350 400
Re Re

(c) (c)

Fig. 4. Variation of Nusselt number with Reynolds number for channel height of (a) 3 mm Fig. 5. Variation of pressure drop with Reynolds number for (a) water, (b) Al2O3 nano-
(Ar = 3), (b) 1 mm, (Ar = 1), (c) 0.5 mm (Ar = 2). fluid, (c) hybrid nanofluid.

number. By addition of nanoparticles, pressure drop increases due to 0.25

dual effects increasing viscosity and density. The maximum value of
pressure drop is observed for minichannel of aspect ratio 2 due to
0.2
lowest hydraulic diameter for all the studied fluids. Its maximum value water
is 1009.2 N/m2 for DI water, 1202.5 for Al2O3 and 1202.6 N/m2 for 0.01%vof Al2O3
friction factor

Al2O3-MWCNT hybrid nanofluid at Reynold number of 378 for 0.5 mm 0.15 0.01%vof Al2O3-MWCNT
depth minichannel. For all the Al2O3-MWCNT hybrid nanofluids with
different nanoparticles fraction (9:1, 8:2 and 7:2), there is no significant 0.1
changes for pressure drop. Addition of MWCNT nanoparticles in Al2O3
nanofluid does not significant effect on pressure drop for same average 0.05
particle size.
Variation of friction factor with Reynolds number is shown in Fig. 6.
0
It seems that the friction factor decreases with increase in Reynolds 0 100 200 300 400 500
number, but there is an increase in friction factor with the addition of
Re
nanoparticles. Friction factor for Al2O3 nanofluid and Al2O3-MWCNT
hybrid nanofluid is almost same. Friction factor has its maximum value
Fig. 6. Variation of friction factor with Reynolds number for 3 mm channel height
0.22 for DI water, 0.227 for Al2O3 nanofluid and 0.223 for Al2O3-
(Ar = 3) for different fluids.
MWCNT hybrid nanofluid. The decrease in friction factor with Reynolds
is due to because of laminar fluid flow in a pipe friction factor is in-
versely proportional to Reynolds number and the increase in friction for the aspect ratio of 2, and it decreases with the increase of Reynold
factor with the addition of nanoparticles was due to increase in pressure number.
drop. There was a decrease in friction factor with the decrease in hy- Velocity profile of water and 0.01%vof Al2O3-MWCNT hybrid na-
draulic diameter. Figs. 7 and 8 show the effect of friction factor on nofluid along the length of the channel (centreline) on all the 3 heat
aspect ratio for different working fluid. The friction factor is minimum sinks of having channel height of 3 mm, 1 mm, 0.5 mm have been
shown in Fig. 9. It can be seen that there is a decrease in developing

245
V. Kumar, J. Sarkar International Communications in Heat and Mass Transfer 91 (2018) 239–247

water 3mm channel height

0.25 309 water 1mm channel height
308 water 0.5mm channel height
water 3mm channel height hybrid 3mm channel height
0.2 307 hybrid 1mm channel height
water 1mm channel height hybrid 0.5mm channel height
306

Temperature (K)
friction factor

water 0.5mm channel height

0.15 305
304
0.1 303
302
0.05 301
300
0 299
0 50 100 150 200 250 300 350 400 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035
Re Axial distance (m)

Fig. 7. Variation of friction factor with Reynolds number for base fluid (water). Fig. 10. Temperature profile of water and 0.01% hybrid nanofluid in 3 mm, 1 mm,
0.5 mm channel height along the length of the channel.

0.25 And there is slight increase in outlet temperature by the addition of

0.01% vof Al2O3-MWCNT 1mm
channel height
nanoparticles because of the slight decrease in specific heat capacity.
0.2 0.01% vof Al2O3-MWCNT 3mm
channel height 6. Conclusions
friction factor

0.15 0.01% vof Al2O3-MWCNT 0.5mm

channel height Numerical analysis of nanofluid and hybrid nanofluid flowing in
0.1 minichannel heat sink has been done by using two-phase mixture model
to study the effects of various operating and design parameters. An
0.05 experimental study has been also conducted on minichannel heat sink
to validate the numerical model. From the results and discussion, the
following conclusions can be made:
0
0 50 100 150 200 250 300 350 400
Re ➢ The maximum heat transfer coefficient improvements of nearly 4%
for 0.01%vof Al2O3/water nanofluid, and 9%, 12% and 15.6% for
0.01% volume fraction of Al2O3-MWCNT in the ratio of (9:1), (8:2),
Fig. 8. Variation of friction factor with Reynolds number for Al2O3-MWCNT hybrid na- and (7:3) are observed.
nofluid. ➢ Heat transfer coefficient increases with increase in mixing ratio of
MWCNT nanoparticles fraction mainly due to high thermal con-
ductivity.
0.4
➢ The heat transfer coefficient increases by 30–35% with the decrease
in hydraulic diameter of the channel from 0.0015 m to 0.001 m.
0.35 ➢ Pressure drop increases with the decrease in hydraulic diameter and
with an addition of nanoparticles. Friction factor increases with
increase in channel height.
u (m/sec)

0.3
➢ Hydrodynamic developing length increases with increase in hy-
0.25 water 3mm channel height draulic diameter and with the addition of nanoparticles.
hybrid 3mm channel heigth ➢ In comparison to the single-phase approach, the proposed mixture
water 0.5mm channel height
model revealed a better agreement with the experimental data.
0.2 hybrid 0.5mm channel height
water 1mm channel height
hybrid 1mm channel height References
0.15
0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 [1] S.U.S. Choi, J.A. Eastman, Enhancing thermal conductivity of fluids with nano-
Axial ditance (m) particles, in: D.A. Singer, H.P. Wang (Eds.), Developments and Applications of Non-
Newtonian Flows, vol. 231, ASME, New York, 1995, pp. 99–105.
[2] J. Sarkar, P. Ghosh, A. Adil, A review on hybrid nanofluids: recent research, de-
Fig. 9. Velocity profile of water and 0.01% hybrid nanofluid in 3 mm, 1 mm, 0.5 mm velopment and applications, Renew. Sust. Energ. Rev. 43 (2015) 164–177.
channel height along the length of the channel. [3] A.K. Tiwari, P. Ghosh, J. Sarkar, Particle concentration levels of various nanofluids
in plate heat exchanger for best performance, Int. J. Heat Mass Transf. 89 (2015)
1110–1118.
length with decrease in hydraulic diameter of the channels. And there is [4] Z. Nikkhah, A. Karimipour, M.R. Safaei, P.F. Tehrani, M. Goodarzib, M. Dahari,
S. Wongwises, Forced convective heat transfer of water/functionalized multi-walled
a slight increase in the developing length by the addition of hybrid carbon nanotube nanofluids in a microchannel with oscillating heat flux and slip
nanoparticles because of the increase in Reynolds number for same inlet boundary condition, Int. Commun. Heat Mass Transfer 68 (2015) 69–77.
velocity, which thickens the hydrodynamic boundary layer. The tem- [5] J.Y. Jung, H.S. Oh, H.Y. Kwak, Forced convective heat transfer of nanofluids in
microchannels, Int. J. Heat Mass Transf. 52 (2009) 466–472.
perature profile of water and 0.01%vof Al2O3-MWCNT hybrid nano-
[6] C.J. Ho, L.C. Wei, Z.W. Li, An experimental investigation of forced convective
fluid along the length of the channel (centreline) on all the 3 heat sinks cooling performance of a microchannel heat sink with Al2O3/water nanofluids,
of having channel height of 3 mm, 1 mm, 0.5 mm are shown in Fig. 10. Appl. Therm. Eng. 30 (2010) 96–103.
[7] B. Rimbault, C.T. Nguyen, N. Galanis, Experimental investigation of CuO-water
There is increase in outlet temperature with decrease in hydraulic
nanofluid flow and heat transfer inside a microchannel heat sink, Int. J. Therm. Sci.
diameter of the channels as the distance between the center plane of the 84 (2014) 275–292.
channel and wall decreases with decrease in the hydraulic diameter. [8] S.M. Peyghambarzadeh, S.H. Hashemabadi, A.R. Chabi, M. Salimi, Performance of

246
V. Kumar, J. Sarkar
water based CuO and Al2O3 nanofluids in a Cu–Be alloyheat sink with rectangular
microchannels, Energy Convers. Manag. 86 (2014) 28–38.
[9] N. Ahammed, L.G. Asirvatham, S. Wongwises, Thermoelectric cooling of electronic
devices with nanofluids in a multiport minichannel heat exchanger, Exp. Thermal
Fluid Sci. 74 (2016) 81–90.
[10] E. Manay, B. Sahin, The effect of microchannel height on performance of nano-
fluids, Int. J. Heat Mass Transf. 95 (2016) 307–320.
[11] M.K. Aliabadi, M. Sahamiyan, Performance of nanofluid flow in corrugated mini-
channels heat sink (CMCHS), Energy Convers. Manag. 108 (2016) 297–308.
[12] Y. Wang, Z. Chen, X. Ling, An experimental study of the latent functionally thermal
fluid with micro-encapsulated phase change material particles flowing in micro-
channels, Appl. Therm. Eng. 105 (2016) 209–216.
[13] P. Selvakumar, S. Suresh, Use of Al2O3–Cu/water hybrid nanofluid in an electronic
heat sink, IEEE Trans. Compon. Packag. Manuf. Technol. 2 (2012) 1600–1607.
[14] H.A. Mohammed, P. Gunnasegaran, N.H. Shuaib, Heat transfer in rectangular mi-
crochannels heat sink using nanofluids, Int. Commun. Heat Mass Transfer 37 (2010)
1496–1503.
[15] R. Nimmagadda, K. Venkatasubbaiah, Conjugate heat transfer analysis of micro-
channel using novel hybrid nanofluids (Al2O3 + Ag/Water), Eur. J. Mech. B. Fluids
52 (2015) 19–27.
[16] Y. Yue, S.K. Mohammadian, Y. Zhang, Analysis of performances of a manifold
microchannel heat sink with nanofluids, Int. J. Therm. Sci. 89 (2015) 305–313.
[17] G.D. Xia, R. Liu, J. Wang, M. Du, The characteristics of convective heat transfer in
microchannel heat sinks using Al2O3 and TiO2 nanofluids, Int. Commun. Heat Mass
Transfer 76 (2016) 256–264.
[18] Y.T. Yang, H.W. Tang, W.P. Ding, Optimization design of micro-channel heat sink
using nanofluid by numerical simulation coupled with genetic algorithm, Int.
Commun. Heat Mass Transfer 72 (2016) 29–38.
[19] M. Kalteh, A. Abbassi, M. Saffar-Avval, J. Harting, Eulerian–Eulerian two-phase
numerical simulation of nanofluid laminar forced convection in a microchannel,
Int. J. Heat Fluid Flow 32 (2011) 107–116.
[20] M. Kalteh, A. Abbassi, M. Saffar-Avval, A. Frijns, A. Darhuber, J. Harting,
Experimental and numerical investigation of nanofluid forced convection inside a
wide microchannel heat sink, Appl. Therm. Eng. 36 (2012) 260–268.
[21] M.K. Moraveji, R.M. Ardehali, CFD modeling (comparing single and two-phase
approaches) on thermal performance of Al2O3/water nanofluid in minichannel heat
sink, Int. Commun. Heat Mass Transfer 44 (2013) 157–164.
[22] P. Naphon, L. Nakharintr, Numerical investigation of laminar heat transfer of na-
nofluid-cooled mini-rectangular fin heat sinks, J. Eng. Phys. Thermophys. 88 (2015)
666–675.
[23] A. Esmaeilnejad, H. Aminfar, M.S. Neistanak, Numerical investigation of forced
convection heat transfer through microchannels with non-Newtonian nanofluids,
Int. J. Therm. Sci. 75 (2014) 76–86.
[24] T. Ambreen, M.H. Kim, Comparative assessment of numerical models for nanofluids'
laminar forced convection in micro and mini channels, Int. J. Heat Mass Transf. 115
(2017) 513–523.
247
International Communications in Heat and Mass Transfer 91 (2018) 239–247
[25] M. Khoshvaght-Aliabadi, Z. Arani, F. Rahimpour, Influence of Al2O3–H2O nanofluid
on performance of twisted minichannels, Adv. Powder Technol. 27 (2016)
1514–1525.
[26] M. Amani, P. Amani, A. Kasaeian, O. Mahian, Y. Wei-Mon, Two-phase mixture
model for nanofluid turbulent flow and heat transfer: effect of heterogeneous dis-
tribution of nanoparticles, Chem. Eng. Sci. 167 (2017) 135–144.
[27] K.M. Shirvan, M. Mamourian, S. Mirzakhanlari, R. Ellahi, Numerical investigation
of heat exchanger effectiveness in a double pipe heat exchanger filled with nano-
fluid: a sensitivity analysis by response surface methodology, Powder Technol. 313
(2017) 99–111.
[28] M.H. Toosi, M. Siavashi, Two-phase mixture numerical simulation of natural con-
vection of nanofluid flow in a cavity partially filled with porous media to enhance
heat transfer, J. Mol. Liq. 238 (2017) 553–569.
[29] R.D. Selvakumar, S. Dhinakaran, Forced convective heat transfer of nanofluids
around a circular bluff body with the effects of slip velocity using a multi-phase
mixture model, Int. J. Heat Mass Transf. 106 (2017) 816–828.
[30] R. Nimmagadda, K. Venkatasubbaiah, Experimental and multiphase analysis of
nanofluids on the conjugate performance of micro-channel at low Reynolds num-
bers, Heat Mass Transf. 53 (2017) 2099–2115.
[31] R. Nimmagadda, K. Venkatasubbaiah, Two-phase analysis on the conjugate heat
transfer performance of microchannel with Cu, Al, SWCNT, and hybrid nanofluids,
J. Therm. Sci. 9 (2017) 041011.
[32] B.C. Pak, Y. Cho, Hydrodynamic and heat transfer study of dispersed fluids with
submicron metallic oxide particle, Exp. Heat Transfer 11 (1998) 151–170.
[33] Y. Xuan, W. Roetzel, Conceptions for heat transfer correlation of nanofluids, Int. J.
Heat Mass Transf. 43 (19) (2000) 3701–3707.
[34] G.R. Batchelor, Brownian diffusion of particles with hydrodynamic interaction, J.
Fluid Mech. 74 (1) (1976) 1–29.
[35] R.L. Hamilton, O.K. Crosser, Thermal conductivity of heterogeneous two-compo-
nent systems, Ind. Eng. Chem. Fundam. 1 (3) (1962) 187–191.
[36] B. Takabi, S. Salehi, Augmentation of the heat transfer performance of a sinusoidal
corrugated enclosure by employing hybrid nanofluid, Adv. Mech. Eng. 2014 (2014)
147059.
[37] R. Lotfi, Y. Saboohi, A.M. Rashidi, Numerical study of forced convective heat
transfer of nanofluids: comparison of different approaches, Int. Commun. Heat Mass
Transfer 37 (2010) 74–78.
[38] M.N. Labib, M.J. Nine, H. Afrianto, H. Chung, H. Jeong, Numerical investigation on
effect of base fluids and hybrid nanofluid in forced convective heat transfer, Int. J.
Therm. Sci. 71 (2013) 163–171.
[39] M.S. Mojarrad, Z. Keshavarz, A. Shokouhi, Nanofluids thermal behavior analysis
using a new dispersion model along with single-phase, Heat Mass Transf. 49 (9)
(2013) 1333–1343.
[40] L. Schiller, A. Naumann, A drag coefficient correlation, Z. Ver. Dtsch. Ing. 77 (1935)
318–320.
[41] S.J. Kline, F.A. McClintock, Describing uncertainties in single-sample experiments,
Mech. Eng. 75 (1953) 3–8.