e nd

mi

, Iranepar

a r t i c l e i n f o

Article history:Received 25 October 2013

a b s t r a c t

channels, electronic devices, phase change materials, and so on[1,2].

Here, some studies on the natural convection in cavities usingnanouids are reviewed briey. First, the numerical works are re-viewed. Ho et al. [3] in a numerical work investigated the effects of

stimations in anin high solid vol-d numerically theof a squareupto 10%)rk were sim

the boundary conditions considered by Ho et al. [3]. gsimulated the natural convection of ve different waternanouids (volume fractions upto 20%) containing Cu, Ag, CuO,Al2O3, and TiO2 nanoparticles in a cavity where one of the sidewalls was heated using a heater mounted on the wall. The cavityangle was varied between 0 and 90. They indicated that withincreasing the concentration, the heat transfer rate increases. Thisincrease was more considerable for the particles with higher ther-mal conductivity (i.e. Ag and Cu). In a similar work to the study of

Corresponding authors. Tel./fax: +98 511 8816840.E-mail addresses: zeinali@ferdowsi.um.ac.ir (S.Z. Heris), omid.mahian@gmail.

com (O. Mahian).

International Journal of Heat and Mass Transfer 73 (2014) 231238

Contents lists availab

International Journal of H

.ethe eld of nanouids as an advanced type of liquids with sur-prising properties have led to a renewed interest in the study ofnatural convection in enclosures where a nanouid is the workinguid. The satisfactory effects of using nanouids on the heattransfer enhancement are proven in many applications such assolar cells, solar collectors, nuclear reactors, automobiles, micro-

ent thermophysical models may lead to the eopposite trend for the Nusselt number especiallyume fractions. Abu-Nada and Oztop [4] examineeffects of inclination angle (between 0 and 120)lled with Cu/water nanouid (volume fractionsheat transfer. The boundary conditions in this wohttp://dx.doi.org/10.1016/j.ijheatmasstransfer.2014.01.0710017-9310/ 2014 Elsevier Ltd. All rights reserved.cavityon theilar to

t [5]based1. Introduction

Natural convection ow in enclosures is one of the classicproblems that is important due to its application in many thermalengineering devices such as solar collectors, energy storagesystems, electronic devices, and so on. Recent developments in

using different models to calculate the viscosity and thermal con-ductivity on the natural convection heat transfer inside a squarecavity. The bottom and top walls were insulated while one of theside walls was assumed cold and another one hot. They usedalumina/water nanouids as the working uid where the volumeconcentrations do not exceed 4%. They concluded that using differ-Received in revised form 30 January 2014Accepted 30 January 2014Available online 4 March 2014

Keywords:Natural convectionInclination angleNanouidsExperimental studyAn experimental study is conducted to investigate the effects of inclination angle on the natural convec-tion of nanouids inside a cubic cavity with the side size of 10 cm. One of the surfaces of the cavity is keptin cold temperature and another one (opposite side) in hot temperature while the other four surfaces areinsulated. The mixtures of three different types of nanoparticles including Al2O3, TiO2, and CuO withinturbine oil (TO) are used as the heat transfer uid. The heat transfer in the cavity is investigated in threeinclination angles with respect to the horizontal position including 0, 45 and 90where the weight frac-tions of nanoparticles are 0.2%, 0.5%, and 0.8%. The Nusselt number results are presented for the threetypes of nanouids, and different angles of inclination, Rayleigh number, and weight fraction of nanopar-ticles. The results reveal that the turbine oil has the highest Nusselt number in any inclination angle ofthe cavity compared to the nanouids. Also, it is found that at the inclination angle of 90, and the weightfraction 0.2%, the application of TiO2 particles results in the maximum Nusselt number while for weightfraction of 0.8%, the maximum Nusselt number is associated with the CuO nanopowders.

2014 Elsevier Ltd. All rights reserved.A comparative experimental study on thtransfer of different metal oxide nanopowoil inside an inclined cavity

Saeed Zeinali Heris a,, Masoumeh Borhani Pour a, OaDepartment of Chemical Engineering, Ferdowsi University of Mashhad, Mashhad, IranbYoung Researchers and Elite Club, Mashhad Branch, Islamic Azad University, Mashhadc Fluid Mechanics, Thermal Engineering and Multiphase Flow Research Lab. (FUTURE), DTechnology Thonburi, Bangmod, Bangkok 10140, Thailand

journal homepage: wwwatural convection heaters suspended in turbine

d Mahian b,c,, Somchai Wongwises c

tment of Mechanical Engineering, Faculty of Engineering, King Mongkuts University of

le at ScienceDirect

eat and Mass Transfer

l sevier .com/locate / i jhmt

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

HeaHo et al. [3], Abu-Nada et al. [6] investigated the effects of uncer-tainties in thermophysical properties on the heat transfer ofAl2O3water and CuOwater nanouids (volume fractions upto9%) in the cavity. They concluded that the average Nusselt numberhas a higher sensitivity to viscosity compared to thermal conduc-tivity at high Rayleigh numbers, hence, a suitable viscosity modelshould be selected for such conditions.

Ghasemi and Aminossadati [7] analysed numerically the ow ofCu, Al2O3 and TiO2 nanoparticles suspended in water in a cavitywhere the top and bottom walls are insulated, right wall is keptcold and the left wall is subjected to a periodic heat ux. Theyfound that using Cu and TiO2 (volume fractions upto 20%) leadsto the maximum and minimum heat removal from the heat source,respectively. Shahi et al. [8] considered the ow of Cu/water nano-uid (volume fractions upto 5%) in a square cavity where the bot-tom wall is subjected a constant heat ux, while the cooling of thecavity is conducted by entering a nanouid ow from the left walland exiting from the right wall. They concluded that an increase inthe volume fraction increases the average Nusselt number in thecavity. Lin and Violi [9] simulated the effects of particle size onthe natural convection ow of Al2O3/water nanouid (volume frac-tion upto 5%) in a cavity. Their results indicated that by decreasingthe nanoparticle size from 250 nm to 5 nm, the heat transfer rateincreases.

Kahveci [10] conducted a work similar to Ref. [5], with thisdifference that the cavity can rotate from 0 to 90. The author

Nomenclature

A heat transfer areacp specic heat (kJ/kg K)dp diameter of the alumina particles (nm)g gravitational acceleration (m/s2)h surface-averaged heat transfer coefcient (W/m2 K)k thermal conductivity (W/m K)l length of cavity (m)Nunf average Nusselt numberq00 surface-averaged heat ux from the hot wall (W/m2)Prnf Prandtl numberRanf Rayleigh numberT temperature (C)

232 S.Z. Heris et al. / International Journal ofrevealed that with the increase of the Rayleigh number, the incli-nation angle in which the maximum heat transfer rate occurs,changes from 45 to 30. He also found that using Ag nanopartil-ces results in the maximum heat transfer rate, while the heattransfer rate is minimum for TiO2 nanoparticles. Corcione [11] ob-tained the optimum particle loading for the nanouid ow in acavity with different aspect ratios. The author concluded thatwith the decrease of nanoparticle size, the level of optimal vol-ume fraction increases. Mahmoudi et al. [12] studied the mixedconvection ow of Cu/water nanouid in a cavity where fourdifferent congurations were considered based on the inlet andoutlet ow. In another work, Mahmoudi et al. [13] solved the nat-ural convection problem where the thickness of the left wall isconsidered.

The interested readers can refer to other numerical works are asfollows. The effects of temperature dependent models on the nat-ural convection were considered by Abu-Nada [14], the naturalconvection of nanouids in a cavity using the lattice Boltzmannmethod was studied by Lee and Yang [15], investigation of thermo-phoresis and Brownian motion effects on natural convection byHaddad et al. [16], double-diffusive natural convection of nanouidwas studied by Parvin et al. [17], heatline analysis of nanouid owin natural convection has been conducted by Basak and Chamkhah[18], the effect of nanoparticle shape was investigated by Ooi andPopov [19], free convection in a complicated cavity was studiedby Nasrin and Alim [20], and nally the free convection in a cavitylled with nanouids was considered by Garoosi et al. [21] whereseveral pairs of coolers and heaters were installed inside the cavity.A comprehensive review on the natural convection of nanouids incavities with different structures is performed by Haddad et al.[22].

However, also some experimental works on the natural convec-tion of nanouids in cavities are observed in the literature.

Wen and Ding [23] during an experimental work on the natu-ral convection of TiO2/water nanouid between two paralleldisks, concluded that using nanouids leads to the decrease ofthe heat transfer rate in the enclosure. They used the nanouidsat low volume concentrations (not more than 0.57%). Nnanna[24] tested Al2O3water nanouids with volume fractions upto8% in a partially heated cavity. It was found that using nanouidswith concentrations upto 2% results in the increase in the heattransfer rate while for volume fractions greater than 2%, duethe unfavorable effects of viscosity the heat transfer ratedecreases. Li and Peterson [25] sought for the possible reasonsbehind the heat transfer detraction due to using Al2O3/waternanouids (volume fractions upto 6%) in a vertical cylindrical cav-ity. In separate experiments, they could nd that the Brownian

Greek symbolsa thermal diffusivity (m2/s)b volumetric thermal expansion coefcient (1/K)l dynamic viscosity (kg/m s)q density (g/cm3)

Subscriptsbf base uidc cold wallh hot wallm mean valuenf nanouidp nanoparticles

t and Mass Transfer 73 (2014) 231238motion and thermophoresis phenomena affect the heat transferrate. In a nice work, Ho et al. [26] examined the natural convec-tion of Al2O3/water nanouid (volume fractions upto 4%) in thevertical cavities with three different aspect ratios. They concludedthat increasing the particle loading (more than about 1%) leads tothe decrease of the average Nusselt number especially at lowerRayleigh numbers.

As it is seen, although extensive numerical studies have beenperformed on the natural convection in a cavity using nanouids,but there has been little experimental works in this eld. This isdue to the difculties in the experimental study of such ows. Also,in most of the studies the base uids were water or ethylene glycol.

The present work aims to provide a comprehensive investiga-tion on the effects of using three different turbine oil (TO) basednanouids including Al2O3, TiO2, and CuO nanoparticles in a cubiccavity. After preparing the suspensions of nanouids, the effects ofthe inclination angle, Rayleigh number, and volume fraction ofnanoparticles on the Nusselt number are investigated. Utilizingthree different types of nanoparticles let to nd out the effects ofthermal conductivity, heat capacity, density and thermal expan-sion of nanoparticles on the heat transfer rate.

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

2. Nanouid preparation

Nanoparticles of Al2O3 (20 nm), TiO2 (20 nm), and CuO (60 nm)have been purchased from the NIPPON AEROSIL company, Japanwhile the turbine oil is provided by Behran Oil company, Iran.The thermophysical properties of nanoparticles and turbine oilare presented in Table 1. To prepare a given concentration of thenanouids, a specied quantity of the nanoparticles is weighed,then it is added to the turbine oil little by little. Simultaneously,by using a magnetic stirrer the suspension is mixed well for3045 min. After mixing, the samples are inserted inside an ultra-sonic vibrator (Misonix Inc. XL2020) for about 1 h to have a

homogenous mixture and breaking down the agglomeration be-tween the nanoparticles.

The samples are provided in three different concentrationsincluding 0.2 wt.%, 0.5 wt.%, and 0.8 wt.%. By converting theconcentrations in terms of volume fraction, one can nd that theconcentrations do not exceed 0.5 vol.%. This reveals that low con-centrations of nanouids used in the present work. Fig. 1 showsthe photographs of the prepared samples and the pure turbineoil. It should be noted that the samples were stable during the testsand no sedimentation is observed with naked eyes, but; in general,CuO/TO nanouids were stable for a shorter time in comparisonwith TiO2/TO and Al2O3/TO nanouids. The lower level of stability

Table 1Properties of nanoparticles and turbine oil in room temperature.

Nanoparticle type Mean diameter (nm) Thermal expansion coefcient (1/K) Density (kg/m3) Heat capacity (J/kg K) Thermal conductivity (W/m K)

TiO2 20 0.9 105 4250 686.2 8.95CuO 60 1.8 105 6400 535.6 76.5Al2O3 20 0.85 105 3970 765 40Turbine oil 64 105 875 2000 0.133

S.Z. Heris et al. / International Journal of Heat and Mass Transfer 73 (2014) 231238 233Fig. 1. Photographs of the samples: (a) turbine oil, (b) Al2O3/ TO, (c) TiO2/ TO, and (d) CuO/TO.

sc470Highlight

sc470Highlight

sc470Highlight

of CuO/TO nanouids may be due to the higher density and biggersize of its particles compared to other ones.

3. Experimental set up and procedure

The schematic view and photograph of the experimental set upis shown in Fig. 2.

The experimental set up is composed of a test cell, insulatormaterials, nanouid tank, DC power supply, and electrical panel.The test cell is a cubic cavity with the side size of 10 cm. One ofthe walls is cooled down by cooling water while another wall atthe opposite side is heated by a heating element with the maxi-mum power of 1000 W. The power of heating element is regulatedby varying the voltage of DC power supply. The cold and hot wallsare made of copper with a thickness of 2 mm, and other walls aremade of steel and the thickness of 2 mm. All walls are insulatedwith a ceramic ber insulating blanket with a thickness of50 mm. PT 100 thermocouples with an accuracy of 0.1 C are usedto measure the surface temperature of the hot and cold walls. Thethermocouples are calibrated in a constant temperature bathequipped with a digital thermometer with an accuracy of 0.1 C.After about 3 h the system reaches the steady state conditions sothat no change is observed in the temperatures. The temperaturesare recorded in the steady state conditions. After each test, the

234 S.Z. Heris et al. / International Journal of HeaFig. 2. (a) Schematic view of the inclined cavity, and (b) photograph of theexperimental set up.enclosure is cleaned well. The tests are repeated 34 times to besure about the repeatability of the data.

4. Data reduction

The steady state heat ux applied on the hot wall is obtained bymultiplying the voltage (V) and current (I) as follows:

q00 VI 1Therefore the heat transfer coefcient (h) and the Nusselt number(Nu) can be determined as:

h q00

ATh Tc 2

Nu hlknf

3

In which Th and Tc are the temperatures of hot and cold walls,respectively. Also, l is the size of cavity, and knf is the thermal con-ductivity of nanouid.

Also, the Rayleigh number is computed by:

Ra gq2nf cPbnf Th Tcl

knflnf

3

4

The thermophysical properties of the nanouid are calculated at themean temperature of the hot and cold walls (Th + Tc/2) as follows.Thermal conductivity is obtained by HamiltonCrosser equations[27]:

Knf kp n 1kf n 1/kp kf kp n 1kf /kp kf kf 5

Density (qnf), heat capacity (Cnf) and thermal expansion coefcient(bnf) are calculated by Khanafer and Vafai [28]:

qnf qf 1 / qp/ 6

Cnf qf Cf 1 / qpCp/

qnf7

bnf 1 /qbf /qbp

qnf8

The viscosity is determined by the Brinkman relation [29]:

lnf lf

1 /2:59

As mentioned the volume concentrations used in the present workdo not exceed 0.5%, therefore, as it is conrmed by Ho et al. [3] andAbu-Nada et al. [6] the uncertainties in thermophysical modelshave no effect on the reported results in low concentrations.

5. Results and discussion

In this section, the effects of concentration, inclination angle,and nanoparticle type in different Rayleigh numbers on the Nusseltnumber are investigated.

5.1. Effects of concentration

Fig. 3 shows the variations of the Nusselt number with the Ray-leigh number for different concentrations of TiO2/TO, Al2O3/TO,and CuO/TO nanouids and inclination angles of 0, 45, and 90. As

t and Mass Transfer 73 (2014) 231238it is observed, at all inclination angles for TiO2/TO nanouid, theNusselt number decreases with an increase in the weight fractionof nanoparticles. However, for the two other nanouids there is

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

HeatS.Z. Heris et al. / International Journal ofno a specied trend for the changes in the Nusselt number. The fol-lowing relation can help to describe the reasons behind thechanges in the Nusselt number [26]:

hnfhf

bnfbf

!nqnfqf

!2nCp;nfCp;f

n knfkf

1n lnflf

!n10

The above relation can be developed in terms of the Nusseltnumber as follows:

NunfNuf

bnfbf

!nqnfqf

!2nCp;nfCp;f

n knfkf

n lnflf

!n11

or:

NunfNuf

brnqr2nCp;rnkrnlrn 12

where n is considered equal to 1/3, and the subscript r indicates theratio of the properties. With an increase in the concentration ofnanouids, the density, thermal conductivity and viscosity ratios

Fig. 3. Effects of concentration on the Nussand Mass Transfer 73 (2014) 231238 235increase while the heat capacity and thermal expansion coefcientratios decrease. However, it should be considered that the power ofn for the thermal conductivity and viscosity ratios in the relation isnegative. Therefore, the Nusselt number increases with particleloading due to the increase of density ratio. On the other hand,the Nusselt number decreases with particle loading due to the de-crease of thermal conductivity, viscosity, heat capacity and thermalexpansion coefcient ratios. The outcome of these changes is thedecrease of the Nusselt number for the TiO2/TO nanouid. ForAl2O3/TO and CuO/TO nanouids it is not predictable that whichconcentration yields the minimumNusselt number, as shown in g-ure. More experiments are needed to know the reasons behindthese unusual changes. However, the physical and chemical speci-cations of nanoparticles such as the shape, size, heat absorption andthe level of surface activity should be considered as possiblereasons. Moreover, CuO and Al2O3 particles have higher thermalconductivity compared to TiO2 particles. This may affect the ther-mophoresis phenomenon which is occurring in the cavity. Thermo-phoresis is sensitive to the changes in the temperature gradient andhence the variations of the thermal conductivity. In all cases, the

elt number for different nanoparticles.

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

Nusselt number increases with an increase in the Rayleigh numberdue to the increases in the buoyancy force. The rate of increase inthe Nusselt number for nanouids is lower than the pure turbineoil. This may happen because of the agglomeration and probablesedimentation of nanoparticles on the hot plate which deterioratesthe heat transfer.

5.2. Effects of inclination angle

Now, the aim is to focus on the effects of inclination angle onthe Nusselt number for the three types of nanouids. As it is shownin Fig. 4, for all types of nanouids and any concentration, the Nus-selt number is minimized at the inclination angle of 0 where thehot plate is placed on the ground. This happens because in this casethe heat transfer is deteriorated due to the lower velocity of parti-cles that is due to the gravitational effects.

For TiO2 particles, the maximum heat transfer occurs at theangle of 90. For Al2O3 and CuO the heat transfer is highest at90 in the weight fractions of 0.8% and 0.5%. However, for 0.2%particle loading, the maximum heat transfer happens at 45 dueto the higher migration velocity of particles in this angle comparedto 90.

5.3. Effects of nanoparticle type

Finally, the effects of using different nanoparticles in the cavityon the Nusselt number are investigated. Fig. 5 shows the variationof the Nusselt number with the Rayleigh number at the inclinationangle of 90. As it is seen, for the weight fraction of 0.2%, the use ofTiO2 particles leads to the maximum Nusselt number in the cavity,although its thermal conductivity is the smallest. By increasing theweight fraction of nanoparticles from0.2% to 0.8%, Cuo/To nanouidshows a higher Nusselt number compared to TiO2/TO nanouid.Also, it is perceived that for the weight fraction of 0.8%, the magni-tude of the Nusselt number for Al2O3/TO nanouid is higher thanTiO2/TO nanouid at lower Rayleigh numbers.

Referring to Eq. (11) and with the development of it for two par-ticles, one can write:

Nunf p2Nunf P1

bnf ;p2bnf ;p1

!nqnf ;p2qnf ;p1

!2nCp;nf ;p2Cp;nf ;p1

n knf ; p1knf ; p2

n lnf ;p1lnf ;p2

!n

13

where the subscripts p1and p2 indicates the particle type #1 andparticle type #2. To explain the reasons behind the changes of the

236 S.Z. Heris et al. / International Journal of Heat and Mass Transfer 73 (2014) 231238Fig. 4. Effects of inclination angle on the Nusselt number for different nanoparticles.

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

sc470Highlight

2uids is 90. However, the tests on the two other nanouids indicate

for the angle of 90, in the weight fraction of 0.2%, using TiO /TO

e Nu

HeatNusselt number, besides the above relation, the effects of size,-shape, heat absorption, and Brownian motion of nanoparticlesshould be considered. Also, the van der Waals forces between thenanoparticles are another important factor. As the van der Waalsforces increase, the agglomeration of nanoparticles increases whichit can reduce the effect of buoyancy force and consequently theNusselt number decreases. Therefore, more experiments in micro-scopic and macroscopic scales should be done to nd out thereasons behind the changes in the Nusselt number for differentnano materials.

6. Uncertainty analysis

Uncertainty analysis is performed to determine the percentagesof error in each experiment. In the following, the uncertainty anal-ysis for important parameters such as the heat transfer coefcient,Nusselt number, and Rayleigh number is conducted in details.

The maximum possible error (EMi) in the measurement ofparameter M is determined by:

EMi XiM@M@Xi

Exi 14

in which Xi is an independent parameter that can be determined di-rectly through the measurements. M is the d that is calculatedthrough X. EXi is obtained by dividing the measurement accuracyof the device that is used for measuring a parameter to the mini-mum value of the measured quantity during the tests.

Another form of the above equation is as follows:

MaxEM X1M@M@X1

Ex1

2 X2

M@M@X2

Ex2

2 Xn

M@M@Xn

Exn

2" #1=Fig. 5. Effects of different nanoparticles on th

S.Z. Heris et al. / International Journal of6.1. Calculation of errors

In this section, EXi for all measured parameters are calculated.

The mass of nanoparticle : Em 0:0011:752 5:71 104

The temperature : EThTc 0:122:8

4:93 103

The length of cavity : El 0:0010:1 0:01

The surface area of the cavity : EA E2l E2lh i1=2

0:014

Voltage : EV 125 0:04that at the low concentration (i.e. 0.2 wt%), the maximum heattransfer occurs at the inclination angle of 45. Also, it is found thatCurrent : EI 0:0010:648 1:54 103

6.2. Maximum error in the estimation of heat transfer parameters

First, the maximum error in the estimation of heat transfer coef-cient is determined as follows:

MaxEh E2V E2I EA2 EThTc2h i1=2

0:0426

Therefore, the maximum error in the measurement of heat transfercoefcient is 4.26%. In a similar way, the uncertainties in the esti-mation of the Nusselt number and the Rayleigh number are 7.43%and 9.45%, respectively.

7. Conclusions

A comparative experimental study is conducted to examine theeffects of metal oxide nanopowders including TiO2, CuO, and Al2O3suspended in turbine oil on the natural convection ow inside atitled cube cavity. Three inclination angles of 0, 45, and 90 andthree weight fractions of 0.2%, 0.5%, and 0.8% are investigated inthe work. The results show that for any inclination angle and theRayleigh number, the Nusselt number is higher for turbine oil com-pared to the nanouids. For TiO2/TO nanouids, with increasingthe inclination angle from 0 to 90, the Nusselt number increases.In other words, the optimum inclination angle for TiO2/TO nano-

sselt number for the inclination angle of 90.

and Mass Transfer 73 (2014) 231238 2372

nanouid results in the maximum Nusselt number while for0.8% wt.% the maximum Nusselt number is caused by the CuO/TO nanouid ow. It was concluded that besides some factors suchas shape, size, heat absorption, Brownian motion, and physical andchemical properties of the nanoparticles, future experimental stud-ies are needed to know the possible reasons behind the changes inthe Nusselt number for different nano materials.

Acknowledgments

The authors acknowledge Iranian Nano Technology Initiativefor the nancial support of this study. The third author wishes tothank Prof. Somchai Wongwises for the supports during hisresearch at FUTURE, King Mongkuts University of TechnologyThonburi, Thailand. The fourth author wishes to thank the NationalScience and Technology Development Agency, the ThailandResearch Fund and the National Research University Project forthe support.

sc470Highlight

sc470Highlight

References

[1] R. Saidur, K.Y. Leong, H.A. Mohammad, A review on applications and challengesof nanouids, Renewable Sustainable Energy Rev. 15 (3) (2011) 16461668.

[2] O. Mahian, A. Kianifar, S.A. Kalogirou, I. Pop, S. Wongwises, A review of theapplications of nanouids in solar energy, Int. J. Heat Mass Transfer 57 (2013)582594.

[3] C.J. Ho, M.W. Chen, Z.W. Li, Numerical simulation of natural convection ofnanouid in a square enclosure: Effects due to uncertainties of viscosity andthermal conductivity, Int. J. Heat Mass Transfer 51 (1718) (2008) 45064516.

[4] E. Abu-Nada, H.F. Oztop, Effects of inclination angle on natural convection inenclosures lled with Cuwater nanouid, Int. J. Heat Fluid Flow 30 (4) (2009)669678.

[5] E. Byk gt, Natural convection of water-based nanouids in an inclinedenclosure with a heat source, Int. J. Therm. Sci. 48 (11) (2009) 20632073.

[6] E. Abu-Nada, Z. Masoud, H.F. Oztop, A. Campo, Effect of nanouid variableproperties on natural convection in enclosures, Int. J. Therm. Sci. 49 (3) (2010)479491.

[7] B. Ghasemi, S.M. Aminossadati, Periodic natural convection in a nanouid-lled enclosure with oscillating heat ux, Int. J. Therm. Sci. 49 (1) (2010) 19.

[8] M. Shahi, A.H. Mahmoudi, F. Talebi, Numerical study of mixed convectivecooling in a square cavity ventilated and partially heated from the belowutilizing nanouid, Int. Commun. Heat Mass Transfer 37 (2) (2010) 201213.

[9] K.C. Lin, A. Violi, Natural convection heat transfer of nanouids in a verticalcavity: effects of non-uniform particle diameter and temperature on thermalconductivity, Int. J. Heat Fluid Flow 31 (2) (2010) 236245.

[10] K. Kahveci, Buoyancy driven heat transfer of nanouids in a tilted enclosure, J.Heat Transfer 132 (6) (2010) 112.

[11] M. Corcione, Heat transfer features of buoyancy-driven nanouids insiderectangular enclosures differentially heated at the sidewalls, Int. J. Therm. Sci.49 (9) (2010) 15361546.

[12] A.H. Mahmoudi, M. Shahi, F. Talebi, Effect of inlet and outlet location on themixed convective cooling inside the ventilated cavity subjected to an externalnanouid, Int. Commun. Heat Mass Transfer 37 (8) (2010) 11581173.

[13] A.H. Mahmoudi, M. Shahi, A.H. Raouf, Modeling of conjugated heat transfer ina thick walled enclosure lled with nanouid, Int. Commun. Heat MassTransfer 38 (1) (2011) 119127.

[14] E. Abu-Nada, RayleighBnard convection in nanouids: effect of temperaturedependent properties, Int. J. Therm. Sci. 50 (9) (2011) 17201730.

[15] F.-H. Lai, Y.-T. Yang, Lattice Boltzmann simulation of natural convection heattransfer of Al2O3/water nanouids in a square enclosure, Int. J. Therm. Sci. 50(10) (2011) 19301941.

[16] Z. Haddad, E. Abu-Nada, H.F. Oztop, A. Mataoui, Natural convection innanouids: are the thermophoresis and Brownian motion effects signicantin nanouid heat transfer enhancement?, Int J. Therm. Sci. 57 (2012) 152162.

[17] S. Parvin, R. Nasrin, M. Alim, N. Hossain, Double-diffusive natural convection ina partially heated enclosure using a nanouid, Heat Transfer Asian Res. 41(6) (2012) 484497.

[18] T. Basak, A.J. Chamkha, Heatline analysis on natural convection for nanouidsconned within square cavities with various thermal boundary conditions, Int.J. Heat Mass Transfer 55 (2122) (2012) 55265543.

[19] E.H. Ooi, V. Popov, Numerical study of inuence of nanoparticle shape on thenatural convection in Cuwater nanouid, Int. J. Therm. Sci. 65 (2013) 178188.

[20] R. Nasrin, M.A. Alim, Free convective ow of nanouid having twonanoparticles inside a complicated cavity, Int. J. Heat Mass Transfer 63(2013) 191198.

[21] F. Garoosi, G. Bagheri, F. Talebi, Numerical simulation of natural convection ofnanouids in a square cavity with several pairs of heaters and coolers (HACs)inside, Int. J. Heat Mass Transfer 67 (2013) 362376.

[22] Z. Haddada, H.F. Oztop, E. Abu-Nada, A. Mataoui, A review on naturalconvective heat transfer of nanouids, Renewable Sustainable Energy Rev.16 (7) (2012) 53635378.

[23] D. Wen, Y. Ding, Formulation of nanouids for natural convective heat transferapplications, Int. J. Heat Fluid Flow 26 (2005) 855864.

[24] A.G.A. Nnanna, Experimental model of temperature-driven nanouid, J. HeatTransfer 129 (2007) 697704.

[25] C.H. Li, G.P. Peterson, Experimental studies of natural convection heat transferof Al2O3/DI water nanoparticle suspensions (nanouids), Adv. Mech. Eng.(2010) (Article ID 742739).

[26] C.J. Ho, W.K. Liu, Y.S. Chang, C.C. Lin, Natural convection heat transfer ofalumina-water nanouid in vertical square enclosures: an experimental study,Int. J. Therm. Sci. 49 (8) (2010) 13451353.

[27] R.L. Hamilton, O.K. Crosser, Thermal conductivity of heterogeneous twocomponent systems, I&EC Fundam. 1 (1962) 182191.

[28] K. Khanafer, K. Vafai, A critical synthesis of thermophysical characteristics ofnanouids, Int. J. Heat Mass Transfer 54 (2011) 44104428.

[29] H.C. Brinkman, The viscosity of concentrated suspensions and solutions, J.Chem. Phys. 20 (1952) 571.

238 S.Z. Heris et al. / International Journal of Heat and Mass Transfer 73 (2014) 231238

A comparative experimental study on the natural convection heat transfer of different metal oxide nanopowders suspended in turbine oil inside an inclined cavity1 Introduction2 Nanofluid preparation3 Experimental set up and procedure4 Data reduction5 Results and discussion5.1 Effects of concentration5.2 Effects of inclination angle5.3 Effects of nanoparticle type

6 Uncertainty analysis6.1 Calculation of errors6.2 Maximum error in the estimation of heat transfer parameters

7 ConclusionsAcknowledgmentsReferences