International Communications in Heat and Mass Transfer 38 (2011) 1474–1478

Contents lists available at ScienceDirect

International Communications in Heat and Mass Transfer

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

Investigation of turbulent convective heat transfer and pressure drop of TiO2/water

nanofluid in circular tube☆
A.R. Sajadi, M.H. Kazemi ⁎
School of Mechanical Engineering, College of Engineering, University of Tehran, Tehran, Iran

a r t i c l e i n f o a b s t r a c t

Available online 19 July 2011 Turbulent heat transfer behavior of titanium dioxide/water nanofluid in a circular pipe was investigated
experimentally where the volume fraction of nanoparticles in the base fluid was less than 0.25%. The
Keywords: experimental measurements have been carried out in the fully-developed turbulent regime for various
Turbulent flow volumetric concentrations. The results indicated that addition of small amounts of nanoparticles to the base
Nanofluid fluid augmented heat transfer remarkably. There was no much effect on heat transfer enhancement with
Convective heat transfer
increasing the volume fraction of nanoparticles. The measurements also showed that the pressure drop of
Convection correlation
nanofluid was slightly higher than that of the base fluid and increased with increasing the volume
concentration. In this paper, experimental results have been compared with the existing correlations for
nanofluid convective heat transfer coefficient in turbulent regime. Finally, a new correlation of the Nusselt
number will be presented using the results of the experiments with titanium dioxide nanoparticles dispersed
in water.
© 2011 Elsevier Ltd. All rights reserved.

1. Introduction the radiators and in heat exchangers. In all of these applications, the
regime of fluid flow is generally turbulent in which higher heat transfer
The poor thermal properties of fluids act as a main barrier to the is achievable. Based on numerical or experimental analyses of previous
growth of energy-efficient heat exchanger. During the past decades, studies, with increasing the particle volume concentration of nanofluids
many efforts have been devoted to the enhancement of heat transfer. the turbolant heat transfer improved [7–10]. On the other hand,
One of the possible techniques for improving heat transfer is adding Fotukian and Nasr [8] studied turbulent convective heat transfer
millimeter- or micrometer-sized particles in fluids. In recent years, performance of very dilute (less than 0.24% volume) CuO/water
nanofluids have been introduced as an ideal candidate for enhancing nanofluid. In their study, increasing the volume fraction of CuO particles
heat transfer [1,2]. As a result of the small size of nanoparticles, a little in nanofluid had negligible effect on the heat transfer enhancement.
pressure drop is observed in the fluid. In this situation the fluid A few studies have reported the influence of TiO2 on the heat
behaves like a pure fluid or single phase liquid [1]. transfer of nanofluids. Duangthongsuck and Wongwises [12,13]
Various studies have been conducted on the performances of showed that using different predictive thermophysical models for
nanofluids convective heat transfer in laminar flow [3–5]. They nanofluid has no effect on the predicted values of the convective heat
concluded that nanofluids provide heat transfer enhancement in transfer coefficient of TiO2/water nanofluids. Murshed et al. [14]
comparison with their corresponding base fluids. Mansour et al. [6] measured the thermal conductivity of suspended TiO2 nanoparticles
experimentally studied the mixed convection of water–Al2O3 mixture in deionized water. The results showed that the thermal conductivity
inside an inclined tube. Their results indicated that a higher particle of nanofluid increased remarkably with increasing the volume
volume concentration clearly induced a decrease of the Nusselt fraction of nanoparticles. He et al. [15] studied static thermal
number for the horizontal inclination. On the other hand, they conductivity, heat transfer and flow behavior of stable aqueous TiO2
showed that for the vertical one, the Nusselt number remains nearly nanofluid with different particle sizes and concentrations. They found
constant with an increase of particle volume concentration from 0 to that the convective heat transfer coefficient increased with increasing
4%. nanoparticle concentration. They showed that the heat transfer
Nanofluids have valuable industrial applications including heating improvement was more significant in turbulent flow regime.
systems through the hydronic coils, cooling automotive engines through The previous work of the authors did not examine the effect of the
dilute particle concentration on heat transfer and pressure drop of
☆ Communicated by W.J. Minkowycz.
TiO2/water based nanofluids. Since this behavior has not been already
⁎ Corresponding author. reported, experiments with very dilute TiO2 nanoparticles were
E-mail address: kazemi_mh@[Link] (M.H. Kazemi). planned to see if the same behavior was observed. As a result, this

0735-1933/$ – see front matter © 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/[Link].2011.07.007
 A.R. Sajadi, M.H. Kazemi / International Communications in Heat and Mass Transfer 38 (2011) 1474–1478 1475

nanoparticles were mixed with distilled water by a mixer for ten

Nomenclature
minutes. Then, ultrasonic cleaner (model UP400S-Hielcher) was
used to disperse nanoparticles for thirty minutes.
A tube cross section area, m 2
Cp specific heat capacity, kJ/kgK
2.2. Experimental apparatus and procedure
D diameter of the tube, m
h heat transfer coefficient, W/m 2K
An experimental apparatus was built to study the flow and
k thermal conductivity, W/mK
convective heat transfer features in a tube. As shown schematically
L length of the tube, m
in Fig. 1, the experimental system mainly includes a reservoir tank, a
Nu Nusselt number
pump, a flow loop, a test section, a cooler and a steam supplier tank.
Pe Peclet number
The transparent plastic reservoir tank with capacity of 6 L was
Pr Prandtl number
manufactured to reserve the nanofluid and monitor the sedimentation
Re Reynolds number
rate and the height of nanofluid. The test section is a straight copper
T temperature, K
tube with the inner diameter, thickness and length of about 5, 0.675
u mean fluid velocity, m 2/s
and 1800 mm, respectively. Five K-type thermocouples were mounted
on the copper tube wall at equal intervals to measure the wall
temperature. Two other K-type thermocouples were inserted at the
Greek symbols
entrance and exit of the test section to measure the bulk temperature.
α thermal diffusivity, m 2/s
The first 50 cm of the copper tube was thermally isolated from its
ε roughness, m
begining using fiberglass to minimize the heat loss and to guarantee
μ viscosity, Pa s
hydrodynamically fully-developed condition. The flow rate was
ρ density, kg/m 3
controlled with two adjusting valves, one at the end of the test section
υ kinematics viscosity, m 2/s
and the other at the by-pass line. The cooler includes a shell and tube
ϕ nanoparticle volume fraction
heat exchanger which was used to reduce the temperature of the
nanofluid at the inlet of the test section. The 50-liter steam supplier
tank contains water as well as a 8 KW element heater to generate fully
Subscripts saturated vapor. The next 120 cm of test section is surrounded by
b bulk saturated vapor to reach constant water temperature. In order to
nf nanofluid minimize the heat loss from the steam tank supplier to the
w water surrounding area, the whole tank was thermally isolated with a
p solid nanoparticle fiberglass cover. A differential pressure transducer (manufactured by
exp. obtained experimentally Endress Hauser) with an uncertainty of ±1 Pa was employed for
ν laminar sublayer measuring the pressure loss along the test section tube. A 1-liter glass
in inlet vessel with a drain valve was utilized to calculate the flow rate. A
out outlet stopwatch with accuracy of ±0.01 s was employed to measure the
time required for filling the vessel.

2.3. Data reduction

paper aims at studying the heat transfer of TiO2 water-based
nanofluids flow inside the tubes with constant wall temperature The heat transfer performance of nanofluid was defined in terms of
and developing heat transfer correlation based on the experimental the convective heat transfer coefficient as follows:
results. The effect of the dilute particles concentration on heat transfer



and pressure drop of TiO2/water based nanofluids will be investigated. ρCp ⋅ A⋅ u Tbout −Tbin
nf
hnf ð exp:Þ = ð1Þ
2. Experimental procedures π⋅ D⋅ LðTw −Tb ÞLMTD

2.1. Nanofluid preparation where (Tw − Tb)LMTD is the logarithmic mean temperature difference,
and Tw is the wall temperature that is the average of six measured
Preparing stable suspension of nanoparticles in the base liquid
was the first step in applying nanofluids for the heat transfer
enhancement. In this study, TiO2 nanoparticles with an average
diameter of about 30 nm were prepared at Sharif University Physics
laboratory. Thermophysical properties of these nanoparticles calcu-
lated at 65 °C are shown in Table 1. Nanofluid was obtained by
dispersing TiO2 nanoparticles in deionized water as a base fluid. The
nanofluid with five different nanoparticle volume concentrations
(0.05%, 0.1%, 0.15%, 0.20%, and 0.25%) was prepared and utilized to
study the heat transfer enhancement. The proper amount of TiO2

Table 1
Thermophysical properties of TiO2 nanoparticles.

Mean diameter Density Thermal conductivity Specific heat

nm kg/m3 kJ/kgK W/mK

30 4170 11.8 711 Fig. 1. Schematic diagram of experimental set-up.

1476 A.R. Sajadi, M.H. Kazemi / International Communications in Heat and Mass Transfer 38 (2011) 1474–1478

Fig. 2. Comparison of experimental Nusselt number with data obtained by Dittus–

Boelter equation vs Reynolds number.
Fig. 4. Nusselt number for water and TiO2/water nanofluid vs Reynolds number at
different volume concentration.

temperatures on tube wall at different positions. The convective heat

transfer coefficient is usually expressed in the form of Nusselt number
(Nu) as: The viscosity of nanofluid could be estimated with the existing
relations for the two-phase mixture:
hnf ð exp:Þ⋅ D
Nunf ð exp:Þ = ð2Þ
knf μnf = μw ð1 + 2:5ϕÞ: ð8Þ

where D is the tube diameter, and knf is the nanofluid thermal The effective thermal conductivity of solid–liquid mixtures was
conductivity. Traditionally, Nu is related to the Reynolds number and introduced by Yu and Choi [16] according to the following equation:
the Prandtl number defined as:
2
 3
kp + 2kw + 2 kp −kw ð1 + βÞ3 ϕ
uD Knf = 4
 5kw ð9Þ
Renf = ð3Þ
υnf kp + 2kw −2 kp −kw ð1 + βÞ3 ϕ

υnf
Prnf = ð4Þ where β is the ratio of the nanolayer thickness to the original particle
αnf
radius and kp is the equivalent thermal conductivity of the equivalent
particle.
knf The rheological and physical properties of the nanofluid were
αnf =
 ð5Þ
ρCp calculated at the mean fluid temperature. Then, the Nusselt number
nf and convective heat transfer coefficient at different concentrations
were calculated.
where υnf is the nanofluid kinematic viscosity and αnf is the nanofluid
thermal diffusivity. 3. Result and discussion
Three main parameters involved in calculating heat transfer rate of
the nanofluid are heat capacity, viscosity, and thermal conductivity, 3.1. Heat transfer results
which may be quite different from those of the pure fluid. (ρCp)nf for
nanofluids is defined as follows: Before measuring the convective heat transfer coefficient of



 nanofluids, the reliability and accuracy of the experimental system
ρCp = ð1−ϕÞ ρCp + ϕ ρCp p ð6Þ was investigated using water as the working fluid. The results of this
nf w
study are compared with the calculated values obtained from the
well-known Dittus–Boelter equation. As can be seen in Fig. 2, a good
ρnf = ϕρp + ð1−ϕÞρw : ð7Þ coincidence between the experimental results and the calculated
values for water revealed that the experimental data were in a good
agreement (within ±10%) with the prediction of the correlation.
Fig. 3 presents the ratio of convective heat transfer coefficient in
nanofluid to that of pure water as a function of Reynolds number. As
can be seen, a significant augmentation of heat transfer could be
obtained by suspending a small amount of nanoparticles in water. The
suspended particles increased the thermal conductivity of the two-
phase mixture. Furthermore, chaotic movement of these particles
accelerated energy exchange process in the fluid.
As can be seen in Fig. 3, at higher concentration levels, no sensible
increase in the heat transfer of nanofluid was obtained. As an example,
at Reynolds number around 5000, the value of hnf/hw was 1.19 and 1.22
for 0.05% and 0.25% concentrations, respectively, which showed 3%
improvement for 0.20% increase in the concentration. It is worth to
notice that the concentration of 0.05% and 0.25% was the lowest and the
Fig. 3. The ratio of heat transfer coefficient of nanofluid to that of pure water vs highest amount of volume fraction in this experiment, respectively. In
Reynolds number at different concentration of nanoparticles. addition, by increasing Reynolds number for all concentrations, the
 A.R. Sajadi, M.H. Kazemi / International Communications in Heat and Mass Transfer 38 (2011) 1474–1478 1477

Fig. 5. Comparison of experimental Nusselt number with existing convective heat transfer correlations at 0.2% volume fraction.

ratio of heat transfer coefficient of nanofluid to that of the base fluid Nu = 0:085Re
0:71
Pr
0:35
ð12Þ
decreased.
Fig. 4 depicts Nusselt number variations versus Reynolds number for

0:6886 0:001 0:9238 0:4
different volume fractions of nanoparticles. Based on this figure, Nu for Nu = 0:0059 1:0 + 7:6286ϕ Pep Renf Prnf : ð13Þ
nanofluid was greater than that of water. This result confirms the claim
by Xuan and Li [8] that Dittus–Boelter correlation underestimates the Pak and Cho [7] and Maiga et al. [10] showed that for nanofluids,
heat transfer coefficients of nanofluids. Nu is a function of Reynolds number and Prandtl number, and the use
Increasing the concentration of nanoparticles in a small range of a correlation of the Dittus was adequate to represent the forced
studied in this work had no influence on the heat transfer enhancement. convection heat transfer in nanofluids. However, in their correlations
The latter fact was against the observation of Xuan and Li [8] who any dependence on the particle volumetric concentration was not
reported a large increase in the heat transfer of a nanofluid with considered. Xuan and Li [8] presented a correlation in which Nu was a
augmentation of the nanoparticles volume fraction. However, the result function of concentration, particle size (through a particle Peclet
of our study was consistent with the experimental results reported by number), Reynolds number and Prandtl number.
Fotukian and Nasr [11]. In Fig. 5, a comparison between the experimental data and the
proposed correlations is shown for 0.2% volume fraction of TiO2
3.2. Comparison of results with proposed correlations dispersed in water. From this figure, it is clearly observed that except
Maiga et al. correlation, which overestimates the Nusselt number, the
Nusselt number for the fully developed turbulent flow for TiO2/ other correlations underestimate the Nusselt number of nanofluids.
water is compared with the correlation proposed by Dittus–Boelter, This behavior was observed for all volume fractions of nanofluids
Pack and Cho, Maiga et al., and Xuan and Li.,which are shown here investigated in this work.
respectively:
3.3. New Nusselt number correlation
0:23 0:4
Nu = 0:023Re Pr ð10Þ
From our experimental observation, we noticed that increasing the
nanoparticles concentration had no influence on the heat transfer
0:8 0:5 enhancement in turbulent regime in the range of concentrations
Nu = 0:021Re Pr ð11Þ
studied. Therefore, the Nu was not a function of ϕ. In the present work,
a new correlation was derived by careful analyzing of the data

Fig. 6. Comparison of the experimental values with the values obtained from the Fig. 7. The ratio of experimental pressure drop of nanofluid to that of pure water along
proposed correlation, Eq. (14). the test tube versus Reynolds number at different volume concentrations.
1478 A.R. Sajadi, M.H. Kazemi / International Communications in Heat and Mass Transfer 38 (2011) 1474–1478

Table 2 6. Conclusion
The experimental data for convective heat transfer and pressure drop.

ϕ(%) Re hw hnf hnf/hw dPnf/dPw The following conclusions have been drawn from the present
study:
0.05 5547 4568.51 5436.9 1.19 1.13
0.05 10308 7453.6 8795.25 1.18 1.11 1. By suspending a small amount of TiO2 nanoparticles, heat transfer
0.05 15943 10439.2 12631.42 1.21 1.08
0.05 19821 12318.3 14289.23 1.16 1.09
coefficient of nanofluids increased. At a Reynolds number of 5000,
0.05 26004 15451.7 17305.12 1.12 1.08 for 0.25% volume fraction of TiO2, an increase of about 22% in the
0.05 30411 17498.2 18723.07 1.07 1.07 heat transfer coefficient was occurred in comparison with pure
0.1 5933 4802.2 5762.64 1.20 1.2 water.
0.1 11076 7869.8 9286.36 1.18 1.18
2. Increasing the nanoparticles concentration had not considerable
0.1 14730 9850.8 11722.45 1.19 1.17
0.1 20352 12575.2 14209.98 1.13 1.19 effect on the heat transfer enhancement in the range of
0.1 25271 1511.4 16622.1 1.10 1.16 concentration studied in this work.
0.1 30007 17312 18349.66 1.06 1.14 3. The rate of the heat transfer coefficient enhancement of nanofluid
0.15 5329 4436.4 5368.04 1.21 1.26 to that of pure water decreased with increasing the Reynolds
0.15 9852 7177.5 8541.11 1.19 1.23
0.15 15333 10143 11664.45 1.15 1.22
number.
0.15 19948 12380 14608.4 1.18 1.20 4. The pressure drop of nanofluid increased with increasing the
0.15 25541 15236 16759.6 1.11 1.18 volume fraction of nanoparticles. The maximum pressure drop was
0.15 30242 17420 18291 1.05 1.20 about 25% greater than that of pure water which was occurred in
0.2 5281 4408 5377.76 1.22 1.31
the highest volume fraction of nanofluid (0.25%) at Reynolds
0.2 12324 8545.9 10255.1 1.20 1.28
0.2 15308 10131 11650.66 1.15 1.30 number of 5000.
0.2 20463 12629 14144.48 1.12 1.28 5. The experimental data were in disagreement with existing
0.2 25631 15278 16805.8 1.1 1.26 correlations for Nu of nanofluids developed in previous studies.
0.2 30031 17322 18534.54 1.07 1.26 Therefore, a new Nu correlation for single phase liquid was
0.25 5579 4587.9 5597.24 1.22 1.37
0.25 9736 7107.1 8386.38 1.18 1.35
developed. This new equation was a function of the Reynolds
0.25 17409 11149 12821.35 1.15 1.36 number and Prandtl number.
0.25 20359 12575 14587 1.16 1.31
0.25 24472 14739 16507.68 1.12 1.33 References
0.25 29312 16988 18347 1.08 1.30
[1] S.U.S. Choi, Enhancing thermal conductivity of fluids with nanparticles, Develepments
and Applications of Non-Newtonian Flows, FED-vol. 231/MD-Vol. 66, ASME, 1995,
pp. 99–105.
[2] S. Lee, S.U.S. Choi, S. Li, J.A. Eastman, Measuring thermal conductivity of fluids
obtained for all volume fractions of nanofluids illustrated in Fig. 6. This containing oxide nanoparticles, J. Heat Transfer 121 (1999) 280–289.
correlation is a function of Re and Pr numbers as follows: [3] D. Wen, Y. Ding, Experimental investigation into convective heat transfer of
nanofluids at the entrance region under laminar flow condition, Int. J. Heat Mass
0:71 0:35
Transfer 47 (24) (2004) 5181–5188.
Nu = 0:067Re Pr + :0005Re: ð14Þ [4] S.Z. Heris, [Link]. Etemad, Experimental inverstigation of Oxide Nanofluids laminar flow
convective heat transfer, Int. Commun. Heat Mass Transfer 33 (4) (2006) 529–535.
[5] K.S. Hwang, S.P. Jang, S.U.S. Choi, Flow and convective heat transfer characteristics of
The above equation was obtained by curve fitting of all the water-based Al2O3 nanofluids in fully developed laminar flow regime, Int. J. Heat Mass
experimental data for nanofluids. The results showed a good relation Transfer 52 (2009) 193–199.
[6] R. Ben Mansour, N. Galanis, C.T. Nguyen, Experimental study of mixed convection
between the experimental values and the calculated ones by the above
with water–Al2O3 nanofluid in inclined tube with uniform wall heat flux, Int. J. Therm.
equation. As can be seen, the majority of the data was fallen within ±8% of Sci. 50 (2011) 403–410.
the proposed equation. This equation can be used for predicting the heat [7] B. Pak, Y.I. Cho, Hydrodynamic and heat transfer study of dispersed fluids with
transfer coefficient of nanofluids with a volume concentraiton of ≤0.25% submicron metallic oxide partilice, Exp. Heat Transfer 11 (1998) 151–170.
[8] Y. Xuan, Q. Li, Heat transfer enhancement of nanofluids, Int. J. Heat Fluid Flow 21
and a Reynolds number range between 5000 and 30000. Morover, it is (2000) 58–64.
very important to note that this equation was only established with [9] Y. Xuan, W. Rotzel, Conceptions of heat transfer correlation of nanofluids, Int. J. Heat
respect to the data of TiO2/water nanofluids. Mass Transfer 43 (2000) 3701–3707.
[10] S.E.B. Maiga, C.T. Nguyen, N. Galanis, G. Roy, T. Mare, M. Coqueux, Heat transfer
enhancement in turbulent tube flow using Al2O3 nanoparticles suspension, Int. J.
3.4. Pressure drop results Numer Methods Heat Fluid Flow 16 (3) (2006) 275–292.
[11] S.M. Fotukian, M. Nasr Esfahany, Experimental study of turbulent convective heat
transfer and pressure drop of dilute Cu/Water nanofluid inside a circular tube, Int.
It is necessary to measure the pressure drop of nanofluids beside Commun. Heat Mass Transfer 37 (2010) 214–219.
the heat transfer performance in order to apply nanofluids in practical [12] W. Duangthongsuk, S. Wongwises, Effect of Thermo physical properties models on
applications. In this study, nanofluids with 0.05, 0.1, 0.15 and 0.2 vol.% the predicting of the convective heat transfer coefficient for low concentration
nanofluid, Int. J. Heat Mass Transfer 35 (2008) 1320–1326.
suspended nanoparticles were used in the pressure drop test under [13] W. Duangthongsuk, S. Wongwises, An experimental study on the heat transfer
the turbulent flow condtion. The pressure drop of nanofluids along the performance and pressure drop of Ti02–water nanofluids flowing under a turbulent
test section was measured by a Differntial pressure transducer with an flow regime, Int. J. Heat Mass Transfer 53 (2010) 334–344.
[14] S.M.S. Murshed, K.C. Leong, C. Yang, Enhanced thermal conductivity of TiO2 Water
uncertainty of ±1 Pa. Fig. 7 shows the ratio of pressure drop of
nanofluids, Int. J. Therm. Sci. 44 (2005) 367–373.
nanofluids to that of pure water as a function of Reynolds number. The [15] Y. He, Y. Jin, H. Chen, Y. Ding, D. Cang, H. Lu, Heat transfer and flow behaviour of
experimental data are reported in Table 2. As shown in Fig. 7, addition aqueous suspensions of TiO2 nanoparticles (nanofluids) flowing upward through a
of nanoparticles to the base fluid increased the pressure drop. This vertical pipe, Int. J. Heat Mass Transfer 50 (2006) 2272–2282.
[16] W. Yu, S.U.S. Choi, The role of international layers in the enhanced thermal
means that using the nanofluids at higher particle volume fraction conductivity of nanofluids: a renovated Mexwell model, J. Nanopart. Res. 5 (2003)
may create a small penalty in pressure drop. 167–171.