- Email: [email protected]
Nanofluids spray heat transfer enhancement
International Journal of Heat and Mass Transfer 94 (2016) 104–118
Contents lists available at ScienceDirect
International Journal of Heat and Mass Transfer journal homepage: www.elsevier.com/locate/ijhmt
Nanofluids spray heat transfer enhancement Shou-Shing Hsieh ⇑, Hao-Hsiang Liu, Yi-Fan Yeh Department of Mechanical and Electromechanical Engineering, National Sun Yat-Sen University, Kaohsiung 80424, Taiwan, ROC
a r t i c l e
i n f o
Article history: Received 15 June 2015 Received in revised form 3 November 2015 Accepted 14 November 2015
Keywords: Nanofluids Spray cooling Volume fraction Transient cooling curve Boiling
a b s t r a c t Spray cooling experiments were carried out to study the effect of seven (7) different types of nanofluids on heat transfer enhancement. Three different concentrations of 0.04%, 0.07%, and 0.1% by volume of Ag, Al, Al2O3, Fe3O4, SiO2, TiO2 and multi-walled carbon nanotubes (MWCNTs) dispersed with deionized (DI) water were tested, and transient as well as steady spray boiling experiments were conducted over a copper flat plate heater 4 cm2 in size with a 1 cm thickness. The spray was issued by a 270 lm full-cone nozzle with a spray height of 30 mm and a spray mass flux of 1.5 103 kg/cm2 s. Both a transient cooling curve and steady boiling curve were obtained. The results revealed that the average heat transfer coefficient (HTC), as well as the associated critical heat flux (CHF), are significantly enhanced, and the enhancement ratio can be up to 1.7 (HTC) and 1.84 (CHF), respectively, corresponding to the DI water as the nanofluids’ volume fraction increased from 0.04% to 0.1%. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Spray cooling, either non-boiling or boiling, is a very powerful and effective means to remove heat from hot surfaces with a low surface superheat and low mass flux [1–4]. This occurs when liquid is forced through a small orifice/or nozzle into dispersed fine droplets on the surface [5]. The enhancement for a large heat removal comes from the higher convective heat transfer coefficient (HTC). It allows for phase changes in high temperatures and high heat flux applications, such as the cooling of electronic devices, nuclear power generation, cryogenics and steel making processes. A quite detailed review of spray cooling heat transfer was reported by Kim [2]. The HTC during spray cooling is governed not only by the temperature difference between the spray and the wall but by the characteristics of the spray itself, which include droplet size, liquid type, spray velocity, spray angle, spray height, and the target surface temperature/wettability. If the thermal property of the liquid changes, it results in a heat transfer enhancement (e.g. with nanoparticle additions). Over the past few decades, nanofluids, which are most likely liquids containing suspensions of nanoparticles (6100 nm), have been reported and proven to have the potential to enhance heat transfer either in conduction due to their resultant higher thermal conductivity or in convection due to the inclination to break the hydrodynamic/thermal boundary layer especially for laminar convection, which makes them very attractive as the most effective ⇑ Corresponding author. Tel.: +886 (07) 5252000x4215; fax: +886 (07) 5254215. E-mail address: [email protected] (S.-S. Hsieh). http://dx.doi.org/10.1016/j.ijheatmasstransfer.2015.11.061 0017-9310/Ó 2015 Elsevier Ltd. All rights reserved.
heat transfer fluids in many applications. However, reported documents and open citations show that the possible mechanism and the thermal conductivity increase, leading to heat transfer enhancement with volume/mass fraction increase/or decrease of the nanofluids, are still controversial [6] and need further work/ examination in detail. There are many ways to enhance the heat transfer with nanofluids. Nanofluid liquid spray is one of them that has been studied for the effects that the mass concentration/or volume fraction of the nanoparticles have on the heat transfer coefficients (HTC) of the base fluids, DI water in most cases. In fact, the application of nanofluids in spray cooling for electronic devices is an emerging area of research [7]. Nanofluid spray cooling with boiling can cause a buildup of a thin porous layer of nanoparticles on the heater’s surface, which may significantly improve the surface wettability and, consequently, result in a CHF increase; however, the layer may be responsible for a decrease in the boiling heat transfer coefficient as the nanoparticle volume/mass fraction increases due to the nanoparticle deposit on the heater’s surface. Obviously, the related heat transfer mechanisms are not yet completely understood [8]. Further systematic experimental studies need to be conducted. Water has been the most commonly used working fluid with nanoparticles so far, due to its ease of use and great capability to suspend the most nanoparticles. Phase change heat transfer during single drop impacts on a hot solid surface was explored using distilled water and TiO2–water nanofluid. The Weber number (We) was in the range of 25–239, and it was found that TiO2 nanofluids improved the boiling heat transfer at a low wall superheat as reported by Okawa et al. [9].
105
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Nomenclature Cp CHF dj d32 H h h k kcu m q00 Re x, y, z T Tj uj uo
specific heat, kJ/kg K critical heat flux, W/cm2 spray nozzle diameter, lm 3 2 Sauter mean diameter, R di /R di , lm spray height, mm heat transfer coefficient, W/m2K average heat transfer coefficient, W/m2 K thermal conductivity, W/m K thermal conductivity of the copper plate, W/m K mass flux, kg/cm2s heat flux, W/cm2 Reynolds number coordinates, m temperature, °C spray exit temperature, °C spray exit velocity, m/s impact velocity, m/s
However, at a high wall superheat, the opposite result occurs. Chun et al. [7] reported that a rapid quench cooling curve was obtained with nanofluids. Kwark et al. [10] reported that the deposition of Al2O3 nanoparticle film on the hot surface can increase the CHF as well as heat flux during nucleate boiling. Abu-Nada and Oztop [11] examined the heat transfer performance of Al2O3 nanofluids numerically and found an increase in the HTC as compared to pure DI water. Duursma et al. [12] found that the heat removal rate for an experimental study of nanofluid drops in spray cooling was not significantly different from that of its base fluid. Recently, Jackson et al. [13] have demonstrated that nanofluids produce a significantly higher HTC during instantaneous (30 ms) droplet impingement than water and, moreover, the HTC increases as the surface wettability increases. In the foregoing discussion, it was found that a firm conclusion regarding the effect of nanofluids on heat transfer enhancement has not yet been reached especially as to whether the increase in concentration of the nanofluid can cause an associated nucleate boiling heat transfer increase. A fundamental study to broaden our understanding of the underlying mechanism of the nanofluid heat transport phenomenon is essentially necessary. To this end, this study aims to explore and extensively study as well as to provide as useful a document as possible for seven (7) different types of nanofluids (Ag, Al, Al2O3, Fe3O4, SiO2, TiO2, MWCNT) on heat transfer enhancement with different volume fractions (0.04–0.1%). Both transient and steady-state cooling are experimentally investigated.
2. Preparation of nanofluid In the study, Ag, Al, Al2O3, Fe3O4, SiO2, TiO2 and a multi-walled carbon nanotube, listed in Table 1, were supplied by Yong-Zhen Techno Material Co., Taiwan. Deionized (DI) water, with related property data listed in Table 2, was used as a base fluid. The nanofluids of different volume fractions were prepared by dispersing
We
Weber number
Greek symbols density of liquid, kg/m3 surface tension, N/m viscosity of liquid, N s/m2 / volume fraction of nanoparticle DP pressure drop across the nozzle, Pa
q r l
Subscript c cu j w o sat
spray liquid layer copper nozzle exit target surface impact saturation
different quantities of the above-stated nanoparticles in DI water. The solution was sonicated continuously with an ultrasonic vibrator (D9NX-DC200H, DELTA NEW INSTRUMENT Co. Ltd.) for 24 h to ensure proper homogenization of nanoparticles to obtain a stable and uniform colloidal solution. Although there is a small increase in temperature while in the sonic bath during the solution preparation, we would keep it to reach the ambient temperature before it is used (i.e., 28 °C). There was no surfactant used in the experiment. Some evaporation of the nanofluid may have occurred due to the temperature rise during sonication. In order to avoid any significant loss of DI water, a glass cover was placed on the solution bath. Each of the nanofluids tested had volume fractions of 0.04%, 0.07% and 0.1%. The properties of the nanofluids are presented in Table 3 and shown in Fig. 1(a)–(d). Generally, the properties, such as Cp, k, l and q of the nanofluids under study, had a larger value than that of DI water except for the surface tension (r), which was measured and correlated in a definite form (as shown in Fig. 2(a)) within ±20% uncertainty of the experimental data (see Fig. 2(b)). Due to instrument limitations, the nanofluids under study, except for Fe3O4 nanofluid (of magnetic nature), were first inspected and
Table 2 Working medium thermal properties (DI water at 28 °C and 1 atm). Properties
Distilled water
Average molecular weight (kg/kg mole) Critical temperature (°C) Saturation temperature (°C) Density of liquid (kg/m3) Heat of vaporization (kJ/kg) Thermal conductivity of liquid (W/m K) Specific heat of liquid (kJ/kg K) Thermal diffusivity of liquid (m2/s) Surface tension of liquid (N/m) Viscosity (Ns/m2)
18.16 374.2 99.9 996 2256.7 0.616 4.22 1.440 10–7 0.07275 8.9 10–4
Table 1 Nanoparticles parameter (at 28 °C and 1 atm). Nanoparticles
Ag
MCNT
TiO2
SiO2
Al
Al2O3
Fe3O4
Average dimension in water
10–50 (nm)
10–30 (nm) in diameter 10–15 (lm) in length
Surface ratio (m2/g) Density (kg/m3) Specific heat capacity (kJ/kg K) Thermal conductivity (W/m K)
80–105 140 0.235 429
200–300 350 0.45 235
10–30 (nm) 55–85 130 0.7 13
10–25 (nm) 230–280 70 0.91 1.4
10–50 (nm) 40–65 230 0.89 204
5–30 (nm) 108–112 75 0.5 39
10–20 (nm) 160–170 5080 3.85 90
106
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Table 3 Nanofluids parameter (at 28 °C and 1 atm). Concentration (vol%)
Thermal conductivity, knf (W/m K)
Density, qnf (kg/m3)
Heat capacity, Cp,nf (kJ/kg-K)
Viscosity, lnf (N-s/m2)
Ag
0.10 0.07 0.04
0.616184 0.616129 0.616073
996.11 996.14 996.17
4.21961 4.21972 4.21984
8.4021 104 8.4015 104 8.4008 104
MCNT
0.10 0.07 0.04
0.625037 0.622052 0.619219
996.14 996.15 996.18
4.21962 4.21974 4.21985
8.4011 104 8.4007 104 8.4004 104
Al
0.10 0.07 0.04
0.616183 0.616128 0.616073
996.12 996.15 996.17
4.21967 4.21977 4.21987
8.4021 104 8.4015 104 8.4008 104
Al2O3
0.10 0.07 0.04
0.616173 0.616121 0.616069
996.11 996.14 996.16
4.21963 4.21974 4.21985
8.4021 104 8.4015 104 8.4008 104
TiO2
0.10 0.07 0.04
0.616152 0.616107 0.616061
996.11 996.14 996.17
4.21965 4.21975 4.21986
8.4021 104 8.4015 104 8.4008 104
SiO2
0.10 0.07 0.04
0.616178 0.616125 0.616071
996.11 996.14 996.16
4.21967 4.21977 4.21987
8.4021 104 8.4015 104 8.4008 104
Fe3O4
0.10 0.07 0.04
0.616181 0.616126 0.616072
996.61 996.49 996.36
4.21996 4.21997 4.21999
8.4021 104 8.4015 104 8.4008 104
Fig. 1. Thermal properties ratio of nanofluids vs DI water: (a) Density (b) Specific heat (c) Viscosity (d) Thermal conductivity.
characterized before/after becoming colloidal solutions via Transmission Electron Microscopy (TEM), as shown in Fig. 3, which is the major technique used to verify single particles’ dimensions and to identify agglomerations of particles. Although some particles remained dispersed, there were several that formed larger
agglomerates and settled out of the liquid for each nanofluid. Regarding the images of Fe3O4 nanofluids, based on PastorizaGallego et al. [14], it was found that most of them were lumped into a big agglomerate. The size of each type of nanofluid is also listed in Table 1.
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
107
temperature of the nozzle was measured with the aid of a T-type thermocouple attached to the tube surface, just opposite the nozzle. One more T-type thermocouple was used to monitor the spray environment’s temperature. The pressure difference across the spray nozzle was measured by a pressure transducer, along with a pressure regulator which was placed opposite the nozzle. 3.3. Nanofluid flow loop The experimental setup consisted of an open spray system and copper target surface heater. The heat transfer performance of different volume fractions of seven different nanofluids was studied and compared to DI water under the same operating conditions. The nanofluid flow loop was an open circulate and designed so that the liquid entered the test section at the design flow rate in terms of the pressure drop and temperature. Special care was taken to warrant that the concentration of the nanofluid remained constant during the experiment for each case. 3.4. Spray parameter
Fig. 2. (a) Correlation of surface tension (rnf/rbf) on (qbf/qbf) and (a/amax) (b) r⁄(experiment) vs r⁄(correlation).
3. Experiment Fig. 4(a) shows the schematic layout of the experimental setup. It includes a nanofluid delivery system, target surface heater, open spray system, and data acquisition. 3.1. Copper surface heater An oxygen-free copper block was chosen for the study due to its high thermal conductivity and machinability. A 20 20 mm heated plate was used as the target surface with a 10 mm thickness that had an enlarged bottom for embedding fifteen (15) 150 W cartridge heaters. Fig. 4(b) shows the detailed configuration of the test heater. Three Type-T thermocouples with a diameter of 80 lm were inserted 10 mm deep into the copper block at a distance of 1 mm, 3 mm, and 5 mm for the test surface and placed in the center of the impact regime of the target to determine the surface temperature. Thermal grease from Dow Corning was applied in order to make good thermal contact between the thermocouples and copper block. The total heat supplied to the heater was measured by using a voltmeter system. Heat flux was regulated gradually by changing the voltage input through a variac transformer, and AC power was used. The target surface was kept as smooth as possible via a proper emery paper used to polish the surface before each test run. 3.2. Spray system A 270 lm full-cone spray nozzle was installed. The spray height was fixed at 30 mm, as shown in Fig. 4(a). The spray nozzle was oriented downward and perpendicular to the target’s surface. The
There were several variables and parameters that governed the spray flow and heat transfer behavior. The average impact velocity (11.3 m/s) of spray droplets impinged on the target surface was calculated from the correlation developed by Tsai [15]. The spray mass flux (1.5 103 kg/cm2 s) at the target’s surface was obtained from the pressure drop (103 kPa) between the up/down streams of the spray nozzle. To calculate the Reynolds Numbers studied herein, the exit velocity (30 m/s) from the nozzle and the nozzle diameters were used. For the Weber number, the d32 and the above-stated impact velocity were applied. The d32 was found from the work of Ghodbane and Holman [16]. In addition, the spray height and the diameter (dj) were the geometric variables. For the present study they were all fixed (H = 30 mm, and dj = 270 lm). The spray exit temperature was 28 °C for all the cases under study, and the heat input was in the range of 80 W/cm2 to 375 W/cm2. 3.5. Data acquisition and uncertainty estimate Experiments were performed under the operating and working conditions listed in Table 4. Based on the order of the Weber number (We) and Ohnesorge number (Oh) under study, the spray impinging onto the target’s surface was more likely to spread rather than splash. Upon the spray impinging on the surface, power was applied to the heater via adjusting the input of the voltage. The power supply was controlled through a PC with proper software to precisely secure the input voltage and the associated input amps. The related heat flux (q00 ) was calculated using the total power generated by the power supply after deducting theheat loss to the surrounding area and then divided by the heater’s surface area. Moreover, the heat flux (q00 ) can be calculated at the heater’s surface based on 1-D Fourier’s law for double-checking. For transient cooling, results from both approaches were examined for the extreme cases through a 2-D inverse heat conduction (IHC) technique with finite difference methods by using a measured temperature as the input to inversely calculate the surface’s heat flux and temperature. It was found that the deviation in the surface temperature/surface heat flux was less than 3% and 5%, respectively. An average heat transfer coefficient for the spray cooling = q00 /DT where DT = T T . was determined for each run from h w sat The primary uncertainties of the experiments were due to uncertainties in the spray flow rate and the heater’s surface temperature and heat flux measurements. Through the deviation error propagation formula, the resultant uncertainty estimate for flow rate, heater surface temperature, and surface heat flux were found to be
108
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Fig. 3. Nanofluids TEM pictures (a) Ag (b) Al2O3 (c) MCNT (d) SiO2 (e) TiO2 (f) Al (g) Fe3O4 (Pastoriza-Gallego et al. [14]).
±2%, ±3%, and ±5%, respectively. Detailed uncertainty of the associated parameters/variables is listed in Table 5.
4. Results and discussion For the present spray cooling, a highly subcooled stream (72 °C) is impinged on the heater’s surface and the associated
experiments conducted are essentially unsteady in nature. Transient cooling curves, as well as the use of quasi-steady boiling curves from the cooling curves and steady-state boiling curves, are obtained for seven different nanofluids with three volume fractions of 0.04 vol%, 0.07 vol%, and 0.1 vol% for a spray mass of 1.5 103 kg/cm2 s at a spray height of 30 mm under the heat input of 80 W/cm2 to 375 W/cm2. Both transient and steady state HTC are extracted from the unsteady/steady boiling curves. The
109
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Fig. 4. Schematic of (a) flow loop (b) heater details.
Table 4 Experimental parameter. Ag
MCNT
Al
Al2O3
TiO2
SiO2
Fe3O4
DI water
283.39 273.48 268.77
300.96 280.28 269.54
287.19 271.63 259.87
310.58 304.27 292.69
290.78 275.89 268.77
293.91 268.61 263.87
309.61 304.97 292.17
258.93
Reynolds number (Re) 0.1 vol% 0.07 vol% 0.04 vol%
9786.44 9787.44 9788.55
9787.90 9788.47 9789.11
9786.54 9787.53 9788.55
9786.44 9787.44 9788.45
9786.44 9787.44 9788.55
9786.44 9787.44 9788.45
9791.35 9790.87 9790.41
9787.81
Ohnesorge number (Oh) 0.1 vol% 0.07 vol% 0.04 vol%
1.72 103 1.69 10–3 1.67 10–3
1.77 10–3 1.71 10–3 1.68 10–3
1.73 10–3 1.68 10–3 1.65 10–3
1.81 10–3 1.78 10–3 1.75 10–3
1.74 10–3 1.70 10–3 1.67 10–3
1.75 10–3 1.67 10–3 1.66 10–3
1.79 10–3 1.78 10–3 1.75 10–3
1.64 10–3
Heat spreader area (cm2) Nozzle to heat spreader distance (mm) Nozzle diameter, dj (lm) Pressure difference, DP (kPa) Volumetric flow rate, Q (l/min) Mass flux, G (kg/cm2 s) Droplet diameter, d32 (lm)
4 30 270 103.4 0.36 14.9 104 131.9
Weber number (We) 0.1 vol% 0.07 vol% 0.04 vol%
corresponding Leidenfrost temperature (LFT) and onset of nucleate boiling (ONB) are also secured for nanofluids and compared to those with the base fluid-DI water as well. 4.1. Transient cooling curve/HTC Fig. 5(a–h) show the cooling curves of the heated surface cooled by Ag, Al, Al2O3, MWCNT, Fe3O4, SiO2, and TiO2 nanofluids with three different volume fractions each of 0.04 vol%, 0.07 vol%, and 0.1 vol%, respectively, and compared to that of DI water. The corresponding unsteady boiling curves were also obtained and plotted. The heater surface was heated to 400 °C and then cooled by means
of the seven nanofluid sprays and DI water. Basically, the cooling rates of the nanofluid sprays were obviously more rapid than that of DI water for the first 10 s. For instance, it took less than 8 s as the heater surface temperature dropped to 160 °C or so for all nanofluids studied, which is a little less (20–25%) than that of DI water (about 10 s). For the heat removal rate, generally all the nanofluids studied had a considerably higher q00 (left ordinate in Fig. 5) reaching about 350 W/cm2 for Al2O3 and Fe3O4, as shown nanofluids in Fig. 5(b) and (c). However, the DI water was about 160 W/cm2, see Fig. 5(h), which is about two times lower. The effect of the volume fraction of the nanofluids can clearly be noted.
110
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Table 5 Measurement uncertainty for relevant parameters/variables. Parameters/variables
System error (%)
Random error (%)
Combined error (%)
Test surface
Length (L) Width (W) High (H) Area (A)
±0.05 ±0.05 ±0.1 ±0.1
±0.93 ±0.86 ±3.42 ±1.79
±0.93 ±0.86 ±3.42 ±1.79
Test chamber
Length (L) Width (W) High (H) Distance between thermocouples (Dx)
±0.02 ±0.02 ±0.03 ±0.2
±1.02 ±1.24 ±1.11 ±1.12
±1.02 ±1.24 ±1.11 ±1.14
Measurement parameters and variables
Surface temperature (Tw) Power (P) Heat flux (q00 ) Mass flux (G) Pressure drop (DP) Surface tension (r)
±3.67 ±2 ±2.1 ±0.1 ±0.06 ±0.1
±7.16 ±2.87 ±4.66 ±1.79 ±1.45 ±1.27
±8.04 ±3.49 ±5.11 ±1.79 ±1.45 ±1.27
Dimensionless group/heat transfer data
Weber number (We) Reynolds number (Re) Heat transfer coefficient (h)
±0.93 ±0.1 ±5.77
±4.29 ±1.79 ±11.82
±4.39 ±1.79 ±13.15
As the volume fraction increased, the heat removal rate consistently increased for all seven nanofluids. Moreover, as usual, the present spray cooling curve also experienced traditional boiling regimes; namely, film boiling, transition boiling, nucleate boiling, and single phase convection. However, due to the presence of nanoparticles in the fluid, there was no noticeable shift to the left in the cooling curves as the volume fraction of the nanofluid increased from 0.07 vol% to 0.1 vol%, which resulted in an early vanishing of the films’ boiling. This is perhaps because the nanoparticles in the fluid deteriorate the vapor film on the heater’s surface and change the morphology and wettability of the surface [17]. In addition to the cooling curves and transient heat removal, the CHF as well as the LFT can be extracted and are listed in Table 6, which suggests that the nanofluids have a consistently higher CHF from 227 W/cm2 to 350 W/cm2 with a corresponding higher LFT for each nanofluid with a different volume fraction. Fig. 6(a)–(h) show the plots of the transient HTC during spray cooling. Time dependent HTCs are calculated based on Newton’s cooling law with time dependent heater surface temperatures. For all the cases studied, it seems, as per Fig. 6(a)–(h), that the HTC approaches, fluctuates, and becomes a consistent value for each nanofluid, as well as for the DI water, as the experiment reaches the steady state of about 600 s. Again, it is expected that the time dependent HTC also experiences the entire boiling regimes as Fig. 5 shows. Peak HTCs are found at a value of about 3 W/cm2 K for both Al2O3 and Fe3O4 nanofluids at a higher volume fraction of 0.1 vol%, respectively among all seven nanofluids. The remainder of the five nanofluids seems to have an equivalent peak value in the average of about 2.5 W/cm2 K; while the DI water had a relatively low value of 1.6 W/cm2 K. Following the results above, the heat transfer enhancement ratio at the peak, in fact, a CHF, is nearly twofold as compared to that of DI water. 4.2. Steady-state boiling curve/HTC Fig. 7(a)–(h) display steady-state boiling curves for seven nanofluids and DI water. Comparisons were made as often as possible depending on available data. Basically, the curves shown of the present nanofluids consistently shift to the left as compared to those of the DI water results, indicating a superior thermal performance to the DI water because the slopes of these profiles (q00 vs DT) became higher as the nanofluid concentration increased. The
change in the slopes was mainly caused by the decrease in the degree of superheat (i.e. ONB) due to the effect of the nanoparticles on the nanofluid’s viscosity and surface tension. For all seven nanofluids, 0.1 vol% had the highest heat transfer rate, followed by 0.07 vol%, and 0.04 vol% in sequence. Compared with the previous investigators’ works [19–23], the present results seem to be superior as evidenced by Fig. 7(a)–(g), except for Fig. 7(c), although they may have different volume fractions. Further, the present DI water results were also compared with those of Jia and Qiu’s [24] work. It is evidenced that there is a noticeable dependence of the thermal performance of the nanofluids on the nanoparticle concentration as long as the nanoparticles can be prevented from sticking to the heater’s surface. Besides, the relevant ONB and CHF were clearly identified, and the values were tabulated in Table 6. Again, like the transient heat removal rate, both Al2O3 and Fe3O4 nanofluids with a 0.1 vol% had the best heat transfer performance (375 W/cm2, a little bit higher than transient results) as far as the CHF is concerned. Furthermore, the associated nucleate boiling heat transfer coefficients (HTCs) were also extracted and plotted against DT(Tw Tsat), as depicted in Fig. 8(a)–(h). For all the cases studied herein, the values of HTC at CHF are in the range of 2.0 (0.04 vol%) to 2.7 (0.1 vol%) W/cm2 K, see Fig. 8(a)–(g) for details. The Al2O3 nanofluids exhibit the highest HTC among the remaining six nanofluids. From the present results shown, such an enhancement is clearly not related to the increased thermal conductivity of the nanofluids. In fact, except for the MWCNT depicted in Fig. 1(d), the increase in thermal conductivity with the nanofluid’s volume fraction increase is quite small. Therefore, it seems that the enhancement of the HTC is much greater than that due to the increase in the thermal conductivity, which may attribute the enhancement to a decrease in the thermal boundary layer thickness. This indicates that the presence of nanoparticles in the spray appears to influence the heat transfer significantly beyond what we expected, due to the increased thermal conductivity alone. Following Fig. 1(c)–(d), the enhancement of the viscosity of the nanofluids with increasing nanoparticle concentration was higher than that of thermal conductivity. In addition to the nanofluid viscosity increase, the nanofluid surface tension, however, seemed to decrease as the nanofluid volume fractions increased, which would result in a higher wettability of the spray. Consequently, it increased the heat transfer. Moreover, the size and shape of the nanoparticle may also affect the heat transfer enhancement.
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Fig. 5. Nanofluids and DI water cooling curve (a) Ag (b) Al (c) Al2O3 (d) MCNT (e) Fe3O4 (f) SiO2 (g) TiO2 (h) DI Water.
111
112
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Table 6 Nanofluids and DI Water boiling curve (steady/transient). CHF (steady) (W/cm2)
CHF (transient) (W/cm2)
ONB (steady) (°C)
ONB (transient) (°C)
LFT* (°C)
0.1 vol% 0.07 vol% 0.04 vol%
175.6 340.3 330.6 321.5
167.7 329.5 315.6 301.5
5.6 4.2 4.5 5.1
4.5 2.9 3.2 3.6
204.8 265.1 266.8 269.3
Al
0.1 vol% 0.07 vol% 0.04 vol%
323.5 309.8 285.6
311.4 270.8 255.5
4.4 4.8 5.3
3.3 3.4 3.7
262.1 270.6 274.2
MCNT
0.1 vol% 0.07 vol% 0.04 vol%
338.6 324.7 304.5
323.7 289.5 267.9
4.3 4.6 4.9
3.2 3.4 3.5
264.5 271.1 272.6
Al2O3
0.1 vol% 0.07 vol% 0.04 vol%
371.5 350.1 340.6
354.6 341.8 329.6
3.8 4.2 4.4
2.1 2.4 2.6
263.2 267.5 270.3
SiO2
0.1 vol% 0.07 vol% 0.04 vol%
345.3 328.3 310.5
331.8 320.4 309.7
4.1 4.4 4.8
2.4 2.9 3.2
265.8 269.5 273.1
TiO2
0.1 vol% 0.07 vol% 0.04 vol%
319.8 312.2 290.1
304.7 290.1 271.2
4.6 4.9 5.3
3.5 3.7 4.1
267.4 270.5 273.1
Fe3O4
0.1 vol% 0.07 vol% 0.04 vol%
366.5 345.3 332.9
350.5 303.2 290.7
3.9 4.2 4.5
2.3 2.5 2.8
261.2 264.5 268.3
DI water Ag
*
Leidenfrost temperature.
Fig. 9 displays comparisons on boiling curves between the steady state and transient with different nanofluid volume fractions, see Fig. 9(a)–(g) and DI water. Also included in each figure is the corresponding LFT. Basically, the deviation for each corresponding case (steady vs transient) in between is not large at the same volume fraction. Like Figs. 5 and 7, the two results from the cooling curve and steady-state measurements seem to be equal with a quite small deviation especially for 0.1 vol% nanofluids. All the nanofluids’ data exhibited a superior heat transfer to that of DI water. LFT was also identified and extracted, as listed in Table 6. Incidentally, ONB was also listed in Table 6. Although the values of the ONB of the transient and the corresponding steady-state seem a little bit different, it was found that the ONB of the nanofluids was lower than that of DI water which indicates the former had a better heat transfer performance than that of the latter. Moreover, the value of the ONB became smaller as the volume fraction of the nanofluid increased. All these parameters like CHF (steady), CHF (transient), and ONB data show that the corresponding transient values are a little bit lower than those of steady state, see Table 6 for details. As stated previously, the highly sub-cooled water spray was impinged on a heater surface and generated a boundary layer. Such a boundary layer may interfere with the vapor bubbles departing from the surface at lower superheats. This causes the CHF of the nanofluids to shift to higher superheats (60°–70 °C) as compared to that of pure DI water (35 °C). Also, shown in Fig. 9 are the associated occurrences of LFT which were extracted from the cooling curves for each nanofluid with different volume fractions and DI water. It was suggested that the LFT for DI water was lower (205 °C) than generally observed for nanofluids and the LFT became higher as the nanofluid volume fraction increased. The corresponding value can also be found in Table 6. These all can be explained by the fact that the vapor film breakage for nanofluid spray cooling for the present study is faster than that of DI water due to the presence of nanoparticle deposition. The deposited nanoparticles on the heater surface can enhance the rewetting phenomena which resulted in the LFT shifting towards the right. Consequently, the results for LFT for
nanofluids were about 57-70 °C higher than that of DI water (see Table 6). Such results can also be found in the cooling curves, shown in Fig. 6, in which the shift for film boiling to transition is earlier for the nanofluids as compared to that of DI water. 4.3. Heat transfer enhancement and related correlation The normalized steady-state HTCs at CHF based on DI water are plotted against the volume fraction, as illustrated in Fig. 10. The enhancement ratio increases from 1.3 to 1.9 as the volume fraction increases from 0.04 vol% to 0.1 vol% for all seven nanofluids. Fig. 11 shows that the nucleate boiling heat transfer coefficient of the nanofluid increases with the increase in heat flux. As the heat flux increases, more bubbles are generated and the chances of bubble penetration become lowand contact with the thermal boundary layer becomes less possible. Meanwhile, when the nanofluid volume fraction increases, the nucleate boiling heat transfer coefficient also increases, due to an increase in surface area of the nanofluid. Furthermore, the increase in the volume fraction of the nanofluid decreases the ONB (see Table 6). The heat transfer enhancement and the nucleate boiling coefficient are observed for seven nanofluids. Finally, a functional relationship among h, q00 , and / is developed for different nanofluids and the DI water; the dimensionless nucleate boiling heat transfer coefficient, h, is correlated in terms of dimensionless q00 and dimensionless / into a composite power-law form of h/hCHF = 1.1(q00 /q00 CHF)0.49(///max)0.23 as depicted in Fig. 12(a). The exponent extracted on q00 shown is nearly the same for that of DI water alone. The discrepancy in the predicted h and the measured h seems quite small with 95% of the data within a ±10% deviation band, as presented in Fig. 12 (b). The effect of the volume fraction seems clearly noted based on the power of the term of (///max). For pure DI water, the term of (///max) is absent and the dependency of the (q00 /q00 CHF) is greater than 0.5; while for the nanofluids this dependency becomes less (0.49) and the contribution due to the nanofluid can be found from the power of the term of (///max), which is found to be about 0.23 in this study, as expected.
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
113
Fig. 6. Transient HTC distribution for nanofluids and DI water (a) Ag (b) Al (c) Al2O3 (d) MCNT (e) Fe3O4 (f) SiO2 (g) TiO2 (h) DI Water. (See above-mentioned reference for further information.)
114
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Fig. 7. Nanofluids and DI water steady boiling curve (a) Ag (b) Al (c) Al2O3 (d) MCNT (e) Fe3O4 (f) SiO2 (g) TiO2 (h) DI Water.
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Fig. 8. Nanofluids and DI water h – DT curve (a) Ag (b) Al (c) Al2O3 (d) MCNT (e) Fe3O4 (f) SiO2 (g) TiO2 (h) DI Water.
115
116
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Fig. 9. Nanofluids and DI water boiling curve (steady/transient mode) (a) Ag (b) Al (c) Al2O3 (d) MCNT (e) Fe3O4 (f) SiO2 (g) TiO2 (h) DI Water.
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Fig. 10. hnano vs hDI
water
117
at the corresponding CHF.
Fig. 12. (a) h–q00 correlation for nanofluids and DI water (b) h⁄(correlation) vs h⁄(experiment).
Fig. 11. Nanofluid and DI water heat transfer coefficient vs different concentration.
5. Conclusion The spray cooling characteristics of nanofluids at low concentrations (60.1 vol%) were experimentally studied with a copper flat heater 4 cm2 in size. Experiments were performed at atmospheric pressure using seven different nanofluids. The main purpose of this study was aimed at examining the effect of the volume fraction of a nanoparticle. Based on the experimental results, the following remarks are drawn: 1. Both the nanofluid spray cooling curve (transient) and steady-state boiling data are obtained. The boiling characteristics of ONB, CHF, as well as LFT from the cooling curves are extracted and discussed. 2. Regarding the nanofluid preparation, a correlation was first made possible for surface tension data for seven nanofluids, which has the form of rnf/rbf = 0.83(qnf/qbf)0.26 (///max)0.17. 3. The enhancement of the heat transfer coefficient for both nucleate boiling and critical heat flux was observed for all seven nanofluids of 0.04 vol%, 0.07 vol%, and 0.1 vol% compared to that of DI water. Generally, the heat transfer performance increases as the volume fraction increases. 4. Among the seven nanofluids, Al2O3 performances are at the best heat transfer rate at the volume fraction of 0.1 vol%, up to 375 W/cm2, due to it having the smallest size of nanoparticles and a small surface tension.
5. The heat transfer enhancement ratio in this study was found to be up to 1.7 for nucleate boiling HTC and 1.9 for the HTC at CHF, and the dimensionless nucleate boiling HTC was correlated into a composite power-law form in terms of dimensionless q00 and /, which has the following form of h/hCHF = 1.1 (q00 /q00 max)0.49(///max)0.23 when / is in the range of 0.04% 6 / 6 0.1%.
References [1] L.C. Chow, M.S. Sehmbey, M.R. Pais, High heat flux spray cooling, Annu. Rev. Heat Transfer 8 (1997) 291–318. [2] J. Kim, Spray Cooling Heat Transfer: The State of the Art, Int. Heat Fluid Flow 28 (4) (2007) 753–767. [3] M. Visaria, I. Mudawar, Theoretical and experimental study of the effects of spray inclination on two phase spray cooling and critical heat flux, Int. J. Heat Mass Transfer 51 (9–10) (2008) 2398–2410. [4] K. Nitin, R.K. Sunil, P.M.V. Subbarao, Experimental study of non-boiling heat transfer from a horizontal surface by water sprays, Exp. Thermal Fluid Sci. 32 (2) (2007) 571–579. [5] J.L. Xie, Z.W. Gan, T.N. Wong, F. Duan, S.C.M. Yu, Y.H. Wu, Thermal effect on a pressure swirl nozzle in spray cooling, Int. J. Heat Mass Transfer 73 (2014) 130–140. [6] Nanofluids-MIT, web.mit.edu/nanoengineering (12/13/2014). [7] S.Y. Chun, I.C. Bang, Y.J. Choo, C.H. Song, Heat transfer characteristics of si and sic nanofluids during a rapid quenching and nanoparticles deposition effects, Int. J. Heat Mass Transfer 54 (5–6) (2011) 1217–1223. [8] J. Barber, D. Brutin, L. Tadrist, A review on boiling heat transfer enhancement with nanofluids, Nanoscale Res. Lett. 6 (2011) 280. [9] T. Okawa, K. Nagano, T. Hirano, Boiling heat transfer during single nanofluid drop impacts onto a hot wall, Exp. Thermal Fluid Sci. 36 (2012) 78–85. [10] S.M. Kwark, R. Kumar, G. Moreno, J. Yoo, S.M. You, Pool boiling characteristics of low concentration nanofluids, Int. J. Heat Mass Transfer 53 (5–6) (2010) 972–981.
118
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
[11] E. Abu-Nada, H.F. Oztop, Numerical analysis of Al2O3/water nanofluids natural convection in a wavy walled cavity, Numer. Heat Transfer Part A: App.: Int. J. Comput. Methodol. 59 (5) (2011) 403–419. [12] S. Duursma, K. Sefiane, A. Kennedy, Experimental studies of nanofluid droplets in spray cooling, Heat Transfer Eng. 30 (13) (2009) 1108–1120. [13] R.G. Jackson, M. Kahani, N. Karwa, A. Wu, R. Lamb, R. Taylor, G. Rosengarten, Effect of surface wettability on carbon nanotube water-based nanofluid droplet impingement heat transfer, J. Phys. Conf. Ser. 525 (2014) 012024. [14] M.J. Pastoriza-Gallego, M. Pérez-Rodríguez, C. Gracia-Fernández, M.M. Piñeiro, Study of viscoelastic properties of magnetic nanofluids: and insight into their internal structure, Soft Matter 9 (48) (2013) 11690–11698. [15] H.-C. Tsai, An experimental study of jet impingement and spray cooling (Ph.D. thesis), National Sun Yat-Sen University, Kaohsiung, 2006. [16] M. Ghodbane, J.P. Holman, Experimental study of spray cooling with freon113, Int. J. Heat Mass Transfer 34 (4–5) (1991) 1163–1174. [17] A. Bolukbasi, D. Ciloglu, Pool boiling heat transfer characteristics of vertical cylinder quenched by SiO2–water nanofluids, Int. J. Therm. Sci. 50 (6) (2011) 1013–1021.
[18] I.L. Ferreira, J.E. Spinelli, J.C. Pires, A. Garcia, The effect of melt temperature profile on the transient metal/mold heat transfer coefficient during solidification, Mater. Sci. Eng., A 408 (1–2) (2005) 317–325. [19] S.S. Hsieh, H.Y. Leu, H.H. Liu, Spray cooling characteristics of nanofluids for electronic power devices, Nanoscale Res. Lett. 10 (2015) 139. [20] H.S. Ahn, M.H. Kim, The boiling phenomenon of alumina nanofluid near critical heat flux, Int. J. Heat Mass Transfer 62 (2013) 718–728. [21] H. Kim, H.S. Ahn, M.H. Kim, On the Mechanism of Pool Boiling Critical Heat Flux Enhancement in Nanofluids, J. Heat Transfer 132 (6) (2010). 061501-1-11. [22] K.J. Park, D. Jung, S.E. Shim, Nucleate boiling heat transfer in aqueous solutions with carbon nanotubes up to critical heat fluxes, Int. J. Multiph. Flow 35 (6) (2009) 525–532. [23] C. Gerardi, J. Buongiorno, L. Hu, T. McKrell, Infrared thermometry study of nanofluid pool boiling phenomena, Nanoscale Res. Lett. 6 (2011) 232. [24] W. Jia, H.H. Qiu, Experimental investigation of droplet dynamics and heat transfer in spray cooling, Exp. Thermal Fluid Sci. 27 (7) (2003) 829–838.
Contents lists available at ScienceDirect
International Journal of Heat and Mass Transfer journal homepage: www.elsevier.com/locate/ijhmt
Nanofluids spray heat transfer enhancement Shou-Shing Hsieh ⇑, Hao-Hsiang Liu, Yi-Fan Yeh Department of Mechanical and Electromechanical Engineering, National Sun Yat-Sen University, Kaohsiung 80424, Taiwan, ROC
a r t i c l e
i n f o
Article history: Received 15 June 2015 Received in revised form 3 November 2015 Accepted 14 November 2015
Keywords: Nanofluids Spray cooling Volume fraction Transient cooling curve Boiling
a b s t r a c t Spray cooling experiments were carried out to study the effect of seven (7) different types of nanofluids on heat transfer enhancement. Three different concentrations of 0.04%, 0.07%, and 0.1% by volume of Ag, Al, Al2O3, Fe3O4, SiO2, TiO2 and multi-walled carbon nanotubes (MWCNTs) dispersed with deionized (DI) water were tested, and transient as well as steady spray boiling experiments were conducted over a copper flat plate heater 4 cm2 in size with a 1 cm thickness. The spray was issued by a 270 lm full-cone nozzle with a spray height of 30 mm and a spray mass flux of 1.5 103 kg/cm2 s. Both a transient cooling curve and steady boiling curve were obtained. The results revealed that the average heat transfer coefficient (HTC), as well as the associated critical heat flux (CHF), are significantly enhanced, and the enhancement ratio can be up to 1.7 (HTC) and 1.84 (CHF), respectively, corresponding to the DI water as the nanofluids’ volume fraction increased from 0.04% to 0.1%. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Spray cooling, either non-boiling or boiling, is a very powerful and effective means to remove heat from hot surfaces with a low surface superheat and low mass flux [1–4]. This occurs when liquid is forced through a small orifice/or nozzle into dispersed fine droplets on the surface [5]. The enhancement for a large heat removal comes from the higher convective heat transfer coefficient (HTC). It allows for phase changes in high temperatures and high heat flux applications, such as the cooling of electronic devices, nuclear power generation, cryogenics and steel making processes. A quite detailed review of spray cooling heat transfer was reported by Kim [2]. The HTC during spray cooling is governed not only by the temperature difference between the spray and the wall but by the characteristics of the spray itself, which include droplet size, liquid type, spray velocity, spray angle, spray height, and the target surface temperature/wettability. If the thermal property of the liquid changes, it results in a heat transfer enhancement (e.g. with nanoparticle additions). Over the past few decades, nanofluids, which are most likely liquids containing suspensions of nanoparticles (6100 nm), have been reported and proven to have the potential to enhance heat transfer either in conduction due to their resultant higher thermal conductivity or in convection due to the inclination to break the hydrodynamic/thermal boundary layer especially for laminar convection, which makes them very attractive as the most effective ⇑ Corresponding author. Tel.: +886 (07) 5252000x4215; fax: +886 (07) 5254215. E-mail address: [email protected] (S.-S. Hsieh). http://dx.doi.org/10.1016/j.ijheatmasstransfer.2015.11.061 0017-9310/Ó 2015 Elsevier Ltd. All rights reserved.
heat transfer fluids in many applications. However, reported documents and open citations show that the possible mechanism and the thermal conductivity increase, leading to heat transfer enhancement with volume/mass fraction increase/or decrease of the nanofluids, are still controversial [6] and need further work/ examination in detail. There are many ways to enhance the heat transfer with nanofluids. Nanofluid liquid spray is one of them that has been studied for the effects that the mass concentration/or volume fraction of the nanoparticles have on the heat transfer coefficients (HTC) of the base fluids, DI water in most cases. In fact, the application of nanofluids in spray cooling for electronic devices is an emerging area of research [7]. Nanofluid spray cooling with boiling can cause a buildup of a thin porous layer of nanoparticles on the heater’s surface, which may significantly improve the surface wettability and, consequently, result in a CHF increase; however, the layer may be responsible for a decrease in the boiling heat transfer coefficient as the nanoparticle volume/mass fraction increases due to the nanoparticle deposit on the heater’s surface. Obviously, the related heat transfer mechanisms are not yet completely understood [8]. Further systematic experimental studies need to be conducted. Water has been the most commonly used working fluid with nanoparticles so far, due to its ease of use and great capability to suspend the most nanoparticles. Phase change heat transfer during single drop impacts on a hot solid surface was explored using distilled water and TiO2–water nanofluid. The Weber number (We) was in the range of 25–239, and it was found that TiO2 nanofluids improved the boiling heat transfer at a low wall superheat as reported by Okawa et al. [9].
105
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Nomenclature Cp CHF dj d32 H h h k kcu m q00 Re x, y, z T Tj uj uo
specific heat, kJ/kg K critical heat flux, W/cm2 spray nozzle diameter, lm 3 2 Sauter mean diameter, R di /R di , lm spray height, mm heat transfer coefficient, W/m2K average heat transfer coefficient, W/m2 K thermal conductivity, W/m K thermal conductivity of the copper plate, W/m K mass flux, kg/cm2s heat flux, W/cm2 Reynolds number coordinates, m temperature, °C spray exit temperature, °C spray exit velocity, m/s impact velocity, m/s
However, at a high wall superheat, the opposite result occurs. Chun et al. [7] reported that a rapid quench cooling curve was obtained with nanofluids. Kwark et al. [10] reported that the deposition of Al2O3 nanoparticle film on the hot surface can increase the CHF as well as heat flux during nucleate boiling. Abu-Nada and Oztop [11] examined the heat transfer performance of Al2O3 nanofluids numerically and found an increase in the HTC as compared to pure DI water. Duursma et al. [12] found that the heat removal rate for an experimental study of nanofluid drops in spray cooling was not significantly different from that of its base fluid. Recently, Jackson et al. [13] have demonstrated that nanofluids produce a significantly higher HTC during instantaneous (30 ms) droplet impingement than water and, moreover, the HTC increases as the surface wettability increases. In the foregoing discussion, it was found that a firm conclusion regarding the effect of nanofluids on heat transfer enhancement has not yet been reached especially as to whether the increase in concentration of the nanofluid can cause an associated nucleate boiling heat transfer increase. A fundamental study to broaden our understanding of the underlying mechanism of the nanofluid heat transport phenomenon is essentially necessary. To this end, this study aims to explore and extensively study as well as to provide as useful a document as possible for seven (7) different types of nanofluids (Ag, Al, Al2O3, Fe3O4, SiO2, TiO2, MWCNT) on heat transfer enhancement with different volume fractions (0.04–0.1%). Both transient and steady-state cooling are experimentally investigated.
2. Preparation of nanofluid In the study, Ag, Al, Al2O3, Fe3O4, SiO2, TiO2 and a multi-walled carbon nanotube, listed in Table 1, were supplied by Yong-Zhen Techno Material Co., Taiwan. Deionized (DI) water, with related property data listed in Table 2, was used as a base fluid. The nanofluids of different volume fractions were prepared by dispersing
We
Weber number
Greek symbols density of liquid, kg/m3 surface tension, N/m viscosity of liquid, N s/m2 / volume fraction of nanoparticle DP pressure drop across the nozzle, Pa
q r l
Subscript c cu j w o sat
spray liquid layer copper nozzle exit target surface impact saturation
different quantities of the above-stated nanoparticles in DI water. The solution was sonicated continuously with an ultrasonic vibrator (D9NX-DC200H, DELTA NEW INSTRUMENT Co. Ltd.) for 24 h to ensure proper homogenization of nanoparticles to obtain a stable and uniform colloidal solution. Although there is a small increase in temperature while in the sonic bath during the solution preparation, we would keep it to reach the ambient temperature before it is used (i.e., 28 °C). There was no surfactant used in the experiment. Some evaporation of the nanofluid may have occurred due to the temperature rise during sonication. In order to avoid any significant loss of DI water, a glass cover was placed on the solution bath. Each of the nanofluids tested had volume fractions of 0.04%, 0.07% and 0.1%. The properties of the nanofluids are presented in Table 3 and shown in Fig. 1(a)–(d). Generally, the properties, such as Cp, k, l and q of the nanofluids under study, had a larger value than that of DI water except for the surface tension (r), which was measured and correlated in a definite form (as shown in Fig. 2(a)) within ±20% uncertainty of the experimental data (see Fig. 2(b)). Due to instrument limitations, the nanofluids under study, except for Fe3O4 nanofluid (of magnetic nature), were first inspected and
Table 2 Working medium thermal properties (DI water at 28 °C and 1 atm). Properties
Distilled water
Average molecular weight (kg/kg mole) Critical temperature (°C) Saturation temperature (°C) Density of liquid (kg/m3) Heat of vaporization (kJ/kg) Thermal conductivity of liquid (W/m K) Specific heat of liquid (kJ/kg K) Thermal diffusivity of liquid (m2/s) Surface tension of liquid (N/m) Viscosity (Ns/m2)
18.16 374.2 99.9 996 2256.7 0.616 4.22 1.440 10–7 0.07275 8.9 10–4
Table 1 Nanoparticles parameter (at 28 °C and 1 atm). Nanoparticles
Ag
MCNT
TiO2
SiO2
Al
Al2O3
Fe3O4
Average dimension in water
10–50 (nm)
10–30 (nm) in diameter 10–15 (lm) in length
Surface ratio (m2/g) Density (kg/m3) Specific heat capacity (kJ/kg K) Thermal conductivity (W/m K)
80–105 140 0.235 429
200–300 350 0.45 235
10–30 (nm) 55–85 130 0.7 13
10–25 (nm) 230–280 70 0.91 1.4
10–50 (nm) 40–65 230 0.89 204
5–30 (nm) 108–112 75 0.5 39
10–20 (nm) 160–170 5080 3.85 90
106
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Table 3 Nanofluids parameter (at 28 °C and 1 atm). Concentration (vol%)
Thermal conductivity, knf (W/m K)
Density, qnf (kg/m3)
Heat capacity, Cp,nf (kJ/kg-K)
Viscosity, lnf (N-s/m2)
Ag
0.10 0.07 0.04
0.616184 0.616129 0.616073
996.11 996.14 996.17
4.21961 4.21972 4.21984
8.4021 104 8.4015 104 8.4008 104
MCNT
0.10 0.07 0.04
0.625037 0.622052 0.619219
996.14 996.15 996.18
4.21962 4.21974 4.21985
8.4011 104 8.4007 104 8.4004 104
Al
0.10 0.07 0.04
0.616183 0.616128 0.616073
996.12 996.15 996.17
4.21967 4.21977 4.21987
8.4021 104 8.4015 104 8.4008 104
Al2O3
0.10 0.07 0.04
0.616173 0.616121 0.616069
996.11 996.14 996.16
4.21963 4.21974 4.21985
8.4021 104 8.4015 104 8.4008 104
TiO2
0.10 0.07 0.04
0.616152 0.616107 0.616061
996.11 996.14 996.17
4.21965 4.21975 4.21986
8.4021 104 8.4015 104 8.4008 104
SiO2
0.10 0.07 0.04
0.616178 0.616125 0.616071
996.11 996.14 996.16
4.21967 4.21977 4.21987
8.4021 104 8.4015 104 8.4008 104
Fe3O4
0.10 0.07 0.04
0.616181 0.616126 0.616072
996.61 996.49 996.36
4.21996 4.21997 4.21999
8.4021 104 8.4015 104 8.4008 104
Fig. 1. Thermal properties ratio of nanofluids vs DI water: (a) Density (b) Specific heat (c) Viscosity (d) Thermal conductivity.
characterized before/after becoming colloidal solutions via Transmission Electron Microscopy (TEM), as shown in Fig. 3, which is the major technique used to verify single particles’ dimensions and to identify agglomerations of particles. Although some particles remained dispersed, there were several that formed larger
agglomerates and settled out of the liquid for each nanofluid. Regarding the images of Fe3O4 nanofluids, based on PastorizaGallego et al. [14], it was found that most of them were lumped into a big agglomerate. The size of each type of nanofluid is also listed in Table 1.
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
107
temperature of the nozzle was measured with the aid of a T-type thermocouple attached to the tube surface, just opposite the nozzle. One more T-type thermocouple was used to monitor the spray environment’s temperature. The pressure difference across the spray nozzle was measured by a pressure transducer, along with a pressure regulator which was placed opposite the nozzle. 3.3. Nanofluid flow loop The experimental setup consisted of an open spray system and copper target surface heater. The heat transfer performance of different volume fractions of seven different nanofluids was studied and compared to DI water under the same operating conditions. The nanofluid flow loop was an open circulate and designed so that the liquid entered the test section at the design flow rate in terms of the pressure drop and temperature. Special care was taken to warrant that the concentration of the nanofluid remained constant during the experiment for each case. 3.4. Spray parameter
Fig. 2. (a) Correlation of surface tension (rnf/rbf) on (qbf/qbf) and (a/amax) (b) r⁄(experiment) vs r⁄(correlation).
3. Experiment Fig. 4(a) shows the schematic layout of the experimental setup. It includes a nanofluid delivery system, target surface heater, open spray system, and data acquisition. 3.1. Copper surface heater An oxygen-free copper block was chosen for the study due to its high thermal conductivity and machinability. A 20 20 mm heated plate was used as the target surface with a 10 mm thickness that had an enlarged bottom for embedding fifteen (15) 150 W cartridge heaters. Fig. 4(b) shows the detailed configuration of the test heater. Three Type-T thermocouples with a diameter of 80 lm were inserted 10 mm deep into the copper block at a distance of 1 mm, 3 mm, and 5 mm for the test surface and placed in the center of the impact regime of the target to determine the surface temperature. Thermal grease from Dow Corning was applied in order to make good thermal contact between the thermocouples and copper block. The total heat supplied to the heater was measured by using a voltmeter system. Heat flux was regulated gradually by changing the voltage input through a variac transformer, and AC power was used. The target surface was kept as smooth as possible via a proper emery paper used to polish the surface before each test run. 3.2. Spray system A 270 lm full-cone spray nozzle was installed. The spray height was fixed at 30 mm, as shown in Fig. 4(a). The spray nozzle was oriented downward and perpendicular to the target’s surface. The
There were several variables and parameters that governed the spray flow and heat transfer behavior. The average impact velocity (11.3 m/s) of spray droplets impinged on the target surface was calculated from the correlation developed by Tsai [15]. The spray mass flux (1.5 103 kg/cm2 s) at the target’s surface was obtained from the pressure drop (103 kPa) between the up/down streams of the spray nozzle. To calculate the Reynolds Numbers studied herein, the exit velocity (30 m/s) from the nozzle and the nozzle diameters were used. For the Weber number, the d32 and the above-stated impact velocity were applied. The d32 was found from the work of Ghodbane and Holman [16]. In addition, the spray height and the diameter (dj) were the geometric variables. For the present study they were all fixed (H = 30 mm, and dj = 270 lm). The spray exit temperature was 28 °C for all the cases under study, and the heat input was in the range of 80 W/cm2 to 375 W/cm2. 3.5. Data acquisition and uncertainty estimate Experiments were performed under the operating and working conditions listed in Table 4. Based on the order of the Weber number (We) and Ohnesorge number (Oh) under study, the spray impinging onto the target’s surface was more likely to spread rather than splash. Upon the spray impinging on the surface, power was applied to the heater via adjusting the input of the voltage. The power supply was controlled through a PC with proper software to precisely secure the input voltage and the associated input amps. The related heat flux (q00 ) was calculated using the total power generated by the power supply after deducting theheat loss to the surrounding area and then divided by the heater’s surface area. Moreover, the heat flux (q00 ) can be calculated at the heater’s surface based on 1-D Fourier’s law for double-checking. For transient cooling, results from both approaches were examined for the extreme cases through a 2-D inverse heat conduction (IHC) technique with finite difference methods by using a measured temperature as the input to inversely calculate the surface’s heat flux and temperature. It was found that the deviation in the surface temperature/surface heat flux was less than 3% and 5%, respectively. An average heat transfer coefficient for the spray cooling = q00 /DT where DT = T T . was determined for each run from h w sat The primary uncertainties of the experiments were due to uncertainties in the spray flow rate and the heater’s surface temperature and heat flux measurements. Through the deviation error propagation formula, the resultant uncertainty estimate for flow rate, heater surface temperature, and surface heat flux were found to be
108
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Fig. 3. Nanofluids TEM pictures (a) Ag (b) Al2O3 (c) MCNT (d) SiO2 (e) TiO2 (f) Al (g) Fe3O4 (Pastoriza-Gallego et al. [14]).
±2%, ±3%, and ±5%, respectively. Detailed uncertainty of the associated parameters/variables is listed in Table 5.
4. Results and discussion For the present spray cooling, a highly subcooled stream (72 °C) is impinged on the heater’s surface and the associated
experiments conducted are essentially unsteady in nature. Transient cooling curves, as well as the use of quasi-steady boiling curves from the cooling curves and steady-state boiling curves, are obtained for seven different nanofluids with three volume fractions of 0.04 vol%, 0.07 vol%, and 0.1 vol% for a spray mass of 1.5 103 kg/cm2 s at a spray height of 30 mm under the heat input of 80 W/cm2 to 375 W/cm2. Both transient and steady state HTC are extracted from the unsteady/steady boiling curves. The
109
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Fig. 4. Schematic of (a) flow loop (b) heater details.
Table 4 Experimental parameter. Ag
MCNT
Al
Al2O3
TiO2
SiO2
Fe3O4
DI water
283.39 273.48 268.77
300.96 280.28 269.54
287.19 271.63 259.87
310.58 304.27 292.69
290.78 275.89 268.77
293.91 268.61 263.87
309.61 304.97 292.17
258.93
Reynolds number (Re) 0.1 vol% 0.07 vol% 0.04 vol%
9786.44 9787.44 9788.55
9787.90 9788.47 9789.11
9786.54 9787.53 9788.55
9786.44 9787.44 9788.45
9786.44 9787.44 9788.55
9786.44 9787.44 9788.45
9791.35 9790.87 9790.41
9787.81
Ohnesorge number (Oh) 0.1 vol% 0.07 vol% 0.04 vol%
1.72 103 1.69 10–3 1.67 10–3
1.77 10–3 1.71 10–3 1.68 10–3
1.73 10–3 1.68 10–3 1.65 10–3
1.81 10–3 1.78 10–3 1.75 10–3
1.74 10–3 1.70 10–3 1.67 10–3
1.75 10–3 1.67 10–3 1.66 10–3
1.79 10–3 1.78 10–3 1.75 10–3
1.64 10–3
Heat spreader area (cm2) Nozzle to heat spreader distance (mm) Nozzle diameter, dj (lm) Pressure difference, DP (kPa) Volumetric flow rate, Q (l/min) Mass flux, G (kg/cm2 s) Droplet diameter, d32 (lm)
4 30 270 103.4 0.36 14.9 104 131.9
Weber number (We) 0.1 vol% 0.07 vol% 0.04 vol%
corresponding Leidenfrost temperature (LFT) and onset of nucleate boiling (ONB) are also secured for nanofluids and compared to those with the base fluid-DI water as well. 4.1. Transient cooling curve/HTC Fig. 5(a–h) show the cooling curves of the heated surface cooled by Ag, Al, Al2O3, MWCNT, Fe3O4, SiO2, and TiO2 nanofluids with three different volume fractions each of 0.04 vol%, 0.07 vol%, and 0.1 vol%, respectively, and compared to that of DI water. The corresponding unsteady boiling curves were also obtained and plotted. The heater surface was heated to 400 °C and then cooled by means
of the seven nanofluid sprays and DI water. Basically, the cooling rates of the nanofluid sprays were obviously more rapid than that of DI water for the first 10 s. For instance, it took less than 8 s as the heater surface temperature dropped to 160 °C or so for all nanofluids studied, which is a little less (20–25%) than that of DI water (about 10 s). For the heat removal rate, generally all the nanofluids studied had a considerably higher q00 (left ordinate in Fig. 5) reaching about 350 W/cm2 for Al2O3 and Fe3O4, as shown nanofluids in Fig. 5(b) and (c). However, the DI water was about 160 W/cm2, see Fig. 5(h), which is about two times lower. The effect of the volume fraction of the nanofluids can clearly be noted.
110
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Table 5 Measurement uncertainty for relevant parameters/variables. Parameters/variables
System error (%)
Random error (%)
Combined error (%)
Test surface
Length (L) Width (W) High (H) Area (A)
±0.05 ±0.05 ±0.1 ±0.1
±0.93 ±0.86 ±3.42 ±1.79
±0.93 ±0.86 ±3.42 ±1.79
Test chamber
Length (L) Width (W) High (H) Distance between thermocouples (Dx)
±0.02 ±0.02 ±0.03 ±0.2
±1.02 ±1.24 ±1.11 ±1.12
±1.02 ±1.24 ±1.11 ±1.14
Measurement parameters and variables
Surface temperature (Tw) Power (P) Heat flux (q00 ) Mass flux (G) Pressure drop (DP) Surface tension (r)
±3.67 ±2 ±2.1 ±0.1 ±0.06 ±0.1
±7.16 ±2.87 ±4.66 ±1.79 ±1.45 ±1.27
±8.04 ±3.49 ±5.11 ±1.79 ±1.45 ±1.27
Dimensionless group/heat transfer data
Weber number (We) Reynolds number (Re) Heat transfer coefficient (h)
±0.93 ±0.1 ±5.77
±4.29 ±1.79 ±11.82
±4.39 ±1.79 ±13.15
As the volume fraction increased, the heat removal rate consistently increased for all seven nanofluids. Moreover, as usual, the present spray cooling curve also experienced traditional boiling regimes; namely, film boiling, transition boiling, nucleate boiling, and single phase convection. However, due to the presence of nanoparticles in the fluid, there was no noticeable shift to the left in the cooling curves as the volume fraction of the nanofluid increased from 0.07 vol% to 0.1 vol%, which resulted in an early vanishing of the films’ boiling. This is perhaps because the nanoparticles in the fluid deteriorate the vapor film on the heater’s surface and change the morphology and wettability of the surface [17]. In addition to the cooling curves and transient heat removal, the CHF as well as the LFT can be extracted and are listed in Table 6, which suggests that the nanofluids have a consistently higher CHF from 227 W/cm2 to 350 W/cm2 with a corresponding higher LFT for each nanofluid with a different volume fraction. Fig. 6(a)–(h) show the plots of the transient HTC during spray cooling. Time dependent HTCs are calculated based on Newton’s cooling law with time dependent heater surface temperatures. For all the cases studied, it seems, as per Fig. 6(a)–(h), that the HTC approaches, fluctuates, and becomes a consistent value for each nanofluid, as well as for the DI water, as the experiment reaches the steady state of about 600 s. Again, it is expected that the time dependent HTC also experiences the entire boiling regimes as Fig. 5 shows. Peak HTCs are found at a value of about 3 W/cm2 K for both Al2O3 and Fe3O4 nanofluids at a higher volume fraction of 0.1 vol%, respectively among all seven nanofluids. The remainder of the five nanofluids seems to have an equivalent peak value in the average of about 2.5 W/cm2 K; while the DI water had a relatively low value of 1.6 W/cm2 K. Following the results above, the heat transfer enhancement ratio at the peak, in fact, a CHF, is nearly twofold as compared to that of DI water. 4.2. Steady-state boiling curve/HTC Fig. 7(a)–(h) display steady-state boiling curves for seven nanofluids and DI water. Comparisons were made as often as possible depending on available data. Basically, the curves shown of the present nanofluids consistently shift to the left as compared to those of the DI water results, indicating a superior thermal performance to the DI water because the slopes of these profiles (q00 vs DT) became higher as the nanofluid concentration increased. The
change in the slopes was mainly caused by the decrease in the degree of superheat (i.e. ONB) due to the effect of the nanoparticles on the nanofluid’s viscosity and surface tension. For all seven nanofluids, 0.1 vol% had the highest heat transfer rate, followed by 0.07 vol%, and 0.04 vol% in sequence. Compared with the previous investigators’ works [19–23], the present results seem to be superior as evidenced by Fig. 7(a)–(g), except for Fig. 7(c), although they may have different volume fractions. Further, the present DI water results were also compared with those of Jia and Qiu’s [24] work. It is evidenced that there is a noticeable dependence of the thermal performance of the nanofluids on the nanoparticle concentration as long as the nanoparticles can be prevented from sticking to the heater’s surface. Besides, the relevant ONB and CHF were clearly identified, and the values were tabulated in Table 6. Again, like the transient heat removal rate, both Al2O3 and Fe3O4 nanofluids with a 0.1 vol% had the best heat transfer performance (375 W/cm2, a little bit higher than transient results) as far as the CHF is concerned. Furthermore, the associated nucleate boiling heat transfer coefficients (HTCs) were also extracted and plotted against DT(Tw Tsat), as depicted in Fig. 8(a)–(h). For all the cases studied herein, the values of HTC at CHF are in the range of 2.0 (0.04 vol%) to 2.7 (0.1 vol%) W/cm2 K, see Fig. 8(a)–(g) for details. The Al2O3 nanofluids exhibit the highest HTC among the remaining six nanofluids. From the present results shown, such an enhancement is clearly not related to the increased thermal conductivity of the nanofluids. In fact, except for the MWCNT depicted in Fig. 1(d), the increase in thermal conductivity with the nanofluid’s volume fraction increase is quite small. Therefore, it seems that the enhancement of the HTC is much greater than that due to the increase in the thermal conductivity, which may attribute the enhancement to a decrease in the thermal boundary layer thickness. This indicates that the presence of nanoparticles in the spray appears to influence the heat transfer significantly beyond what we expected, due to the increased thermal conductivity alone. Following Fig. 1(c)–(d), the enhancement of the viscosity of the nanofluids with increasing nanoparticle concentration was higher than that of thermal conductivity. In addition to the nanofluid viscosity increase, the nanofluid surface tension, however, seemed to decrease as the nanofluid volume fractions increased, which would result in a higher wettability of the spray. Consequently, it increased the heat transfer. Moreover, the size and shape of the nanoparticle may also affect the heat transfer enhancement.
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Fig. 5. Nanofluids and DI water cooling curve (a) Ag (b) Al (c) Al2O3 (d) MCNT (e) Fe3O4 (f) SiO2 (g) TiO2 (h) DI Water.
111
112
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Table 6 Nanofluids and DI Water boiling curve (steady/transient). CHF (steady) (W/cm2)
CHF (transient) (W/cm2)
ONB (steady) (°C)
ONB (transient) (°C)
LFT* (°C)
0.1 vol% 0.07 vol% 0.04 vol%
175.6 340.3 330.6 321.5
167.7 329.5 315.6 301.5
5.6 4.2 4.5 5.1
4.5 2.9 3.2 3.6
204.8 265.1 266.8 269.3
Al
0.1 vol% 0.07 vol% 0.04 vol%
323.5 309.8 285.6
311.4 270.8 255.5
4.4 4.8 5.3
3.3 3.4 3.7
262.1 270.6 274.2
MCNT
0.1 vol% 0.07 vol% 0.04 vol%
338.6 324.7 304.5
323.7 289.5 267.9
4.3 4.6 4.9
3.2 3.4 3.5
264.5 271.1 272.6
Al2O3
0.1 vol% 0.07 vol% 0.04 vol%
371.5 350.1 340.6
354.6 341.8 329.6
3.8 4.2 4.4
2.1 2.4 2.6
263.2 267.5 270.3
SiO2
0.1 vol% 0.07 vol% 0.04 vol%
345.3 328.3 310.5
331.8 320.4 309.7
4.1 4.4 4.8
2.4 2.9 3.2
265.8 269.5 273.1
TiO2
0.1 vol% 0.07 vol% 0.04 vol%
319.8 312.2 290.1
304.7 290.1 271.2
4.6 4.9 5.3
3.5 3.7 4.1
267.4 270.5 273.1
Fe3O4
0.1 vol% 0.07 vol% 0.04 vol%
366.5 345.3 332.9
350.5 303.2 290.7
3.9 4.2 4.5
2.3 2.5 2.8
261.2 264.5 268.3
DI water Ag
*
Leidenfrost temperature.
Fig. 9 displays comparisons on boiling curves between the steady state and transient with different nanofluid volume fractions, see Fig. 9(a)–(g) and DI water. Also included in each figure is the corresponding LFT. Basically, the deviation for each corresponding case (steady vs transient) in between is not large at the same volume fraction. Like Figs. 5 and 7, the two results from the cooling curve and steady-state measurements seem to be equal with a quite small deviation especially for 0.1 vol% nanofluids. All the nanofluids’ data exhibited a superior heat transfer to that of DI water. LFT was also identified and extracted, as listed in Table 6. Incidentally, ONB was also listed in Table 6. Although the values of the ONB of the transient and the corresponding steady-state seem a little bit different, it was found that the ONB of the nanofluids was lower than that of DI water which indicates the former had a better heat transfer performance than that of the latter. Moreover, the value of the ONB became smaller as the volume fraction of the nanofluid increased. All these parameters like CHF (steady), CHF (transient), and ONB data show that the corresponding transient values are a little bit lower than those of steady state, see Table 6 for details. As stated previously, the highly sub-cooled water spray was impinged on a heater surface and generated a boundary layer. Such a boundary layer may interfere with the vapor bubbles departing from the surface at lower superheats. This causes the CHF of the nanofluids to shift to higher superheats (60°–70 °C) as compared to that of pure DI water (35 °C). Also, shown in Fig. 9 are the associated occurrences of LFT which were extracted from the cooling curves for each nanofluid with different volume fractions and DI water. It was suggested that the LFT for DI water was lower (205 °C) than generally observed for nanofluids and the LFT became higher as the nanofluid volume fraction increased. The corresponding value can also be found in Table 6. These all can be explained by the fact that the vapor film breakage for nanofluid spray cooling for the present study is faster than that of DI water due to the presence of nanoparticle deposition. The deposited nanoparticles on the heater surface can enhance the rewetting phenomena which resulted in the LFT shifting towards the right. Consequently, the results for LFT for
nanofluids were about 57-70 °C higher than that of DI water (see Table 6). Such results can also be found in the cooling curves, shown in Fig. 6, in which the shift for film boiling to transition is earlier for the nanofluids as compared to that of DI water. 4.3. Heat transfer enhancement and related correlation The normalized steady-state HTCs at CHF based on DI water are plotted against the volume fraction, as illustrated in Fig. 10. The enhancement ratio increases from 1.3 to 1.9 as the volume fraction increases from 0.04 vol% to 0.1 vol% for all seven nanofluids. Fig. 11 shows that the nucleate boiling heat transfer coefficient of the nanofluid increases with the increase in heat flux. As the heat flux increases, more bubbles are generated and the chances of bubble penetration become lowand contact with the thermal boundary layer becomes less possible. Meanwhile, when the nanofluid volume fraction increases, the nucleate boiling heat transfer coefficient also increases, due to an increase in surface area of the nanofluid. Furthermore, the increase in the volume fraction of the nanofluid decreases the ONB (see Table 6). The heat transfer enhancement and the nucleate boiling coefficient are observed for seven nanofluids. Finally, a functional relationship among h, q00 , and / is developed for different nanofluids and the DI water; the dimensionless nucleate boiling heat transfer coefficient, h, is correlated in terms of dimensionless q00 and dimensionless / into a composite power-law form of h/hCHF = 1.1(q00 /q00 CHF)0.49(///max)0.23 as depicted in Fig. 12(a). The exponent extracted on q00 shown is nearly the same for that of DI water alone. The discrepancy in the predicted h and the measured h seems quite small with 95% of the data within a ±10% deviation band, as presented in Fig. 12 (b). The effect of the volume fraction seems clearly noted based on the power of the term of (///max). For pure DI water, the term of (///max) is absent and the dependency of the (q00 /q00 CHF) is greater than 0.5; while for the nanofluids this dependency becomes less (0.49) and the contribution due to the nanofluid can be found from the power of the term of (///max), which is found to be about 0.23 in this study, as expected.
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
113
Fig. 6. Transient HTC distribution for nanofluids and DI water (a) Ag (b) Al (c) Al2O3 (d) MCNT (e) Fe3O4 (f) SiO2 (g) TiO2 (h) DI Water. (See above-mentioned reference for further information.)
114
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Fig. 7. Nanofluids and DI water steady boiling curve (a) Ag (b) Al (c) Al2O3 (d) MCNT (e) Fe3O4 (f) SiO2 (g) TiO2 (h) DI Water.
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Fig. 8. Nanofluids and DI water h – DT curve (a) Ag (b) Al (c) Al2O3 (d) MCNT (e) Fe3O4 (f) SiO2 (g) TiO2 (h) DI Water.
115
116
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Fig. 9. Nanofluids and DI water boiling curve (steady/transient mode) (a) Ag (b) Al (c) Al2O3 (d) MCNT (e) Fe3O4 (f) SiO2 (g) TiO2 (h) DI Water.
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
Fig. 10. hnano vs hDI
water
117
at the corresponding CHF.
Fig. 12. (a) h–q00 correlation for nanofluids and DI water (b) h⁄(correlation) vs h⁄(experiment).
Fig. 11. Nanofluid and DI water heat transfer coefficient vs different concentration.
5. Conclusion The spray cooling characteristics of nanofluids at low concentrations (60.1 vol%) were experimentally studied with a copper flat heater 4 cm2 in size. Experiments were performed at atmospheric pressure using seven different nanofluids. The main purpose of this study was aimed at examining the effect of the volume fraction of a nanoparticle. Based on the experimental results, the following remarks are drawn: 1. Both the nanofluid spray cooling curve (transient) and steady-state boiling data are obtained. The boiling characteristics of ONB, CHF, as well as LFT from the cooling curves are extracted and discussed. 2. Regarding the nanofluid preparation, a correlation was first made possible for surface tension data for seven nanofluids, which has the form of rnf/rbf = 0.83(qnf/qbf)0.26 (///max)0.17. 3. The enhancement of the heat transfer coefficient for both nucleate boiling and critical heat flux was observed for all seven nanofluids of 0.04 vol%, 0.07 vol%, and 0.1 vol% compared to that of DI water. Generally, the heat transfer performance increases as the volume fraction increases. 4. Among the seven nanofluids, Al2O3 performances are at the best heat transfer rate at the volume fraction of 0.1 vol%, up to 375 W/cm2, due to it having the smallest size of nanoparticles and a small surface tension.
5. The heat transfer enhancement ratio in this study was found to be up to 1.7 for nucleate boiling HTC and 1.9 for the HTC at CHF, and the dimensionless nucleate boiling HTC was correlated into a composite power-law form in terms of dimensionless q00 and /, which has the following form of h/hCHF = 1.1 (q00 /q00 max)0.49(///max)0.23 when / is in the range of 0.04% 6 / 6 0.1%.
References [1] L.C. Chow, M.S. Sehmbey, M.R. Pais, High heat flux spray cooling, Annu. Rev. Heat Transfer 8 (1997) 291–318. [2] J. Kim, Spray Cooling Heat Transfer: The State of the Art, Int. Heat Fluid Flow 28 (4) (2007) 753–767. [3] M. Visaria, I. Mudawar, Theoretical and experimental study of the effects of spray inclination on two phase spray cooling and critical heat flux, Int. J. Heat Mass Transfer 51 (9–10) (2008) 2398–2410. [4] K. Nitin, R.K. Sunil, P.M.V. Subbarao, Experimental study of non-boiling heat transfer from a horizontal surface by water sprays, Exp. Thermal Fluid Sci. 32 (2) (2007) 571–579. [5] J.L. Xie, Z.W. Gan, T.N. Wong, F. Duan, S.C.M. Yu, Y.H. Wu, Thermal effect on a pressure swirl nozzle in spray cooling, Int. J. Heat Mass Transfer 73 (2014) 130–140. [6] Nanofluids-MIT, web.mit.edu/nanoengineering (12/13/2014). [7] S.Y. Chun, I.C. Bang, Y.J. Choo, C.H. Song, Heat transfer characteristics of si and sic nanofluids during a rapid quenching and nanoparticles deposition effects, Int. J. Heat Mass Transfer 54 (5–6) (2011) 1217–1223. [8] J. Barber, D. Brutin, L. Tadrist, A review on boiling heat transfer enhancement with nanofluids, Nanoscale Res. Lett. 6 (2011) 280. [9] T. Okawa, K. Nagano, T. Hirano, Boiling heat transfer during single nanofluid drop impacts onto a hot wall, Exp. Thermal Fluid Sci. 36 (2012) 78–85. [10] S.M. Kwark, R. Kumar, G. Moreno, J. Yoo, S.M. You, Pool boiling characteristics of low concentration nanofluids, Int. J. Heat Mass Transfer 53 (5–6) (2010) 972–981.
118
S.-S. Hsieh et al. / International Journal of Heat and Mass Transfer 94 (2016) 104–118
[11] E. Abu-Nada, H.F. Oztop, Numerical analysis of Al2O3/water nanofluids natural convection in a wavy walled cavity, Numer. Heat Transfer Part A: App.: Int. J. Comput. Methodol. 59 (5) (2011) 403–419. [12] S. Duursma, K. Sefiane, A. Kennedy, Experimental studies of nanofluid droplets in spray cooling, Heat Transfer Eng. 30 (13) (2009) 1108–1120. [13] R.G. Jackson, M. Kahani, N. Karwa, A. Wu, R. Lamb, R. Taylor, G. Rosengarten, Effect of surface wettability on carbon nanotube water-based nanofluid droplet impingement heat transfer, J. Phys. Conf. Ser. 525 (2014) 012024. [14] M.J. Pastoriza-Gallego, M. Pérez-Rodríguez, C. Gracia-Fernández, M.M. Piñeiro, Study of viscoelastic properties of magnetic nanofluids: and insight into their internal structure, Soft Matter 9 (48) (2013) 11690–11698. [15] H.-C. Tsai, An experimental study of jet impingement and spray cooling (Ph.D. thesis), National Sun Yat-Sen University, Kaohsiung, 2006. [16] M. Ghodbane, J.P. Holman, Experimental study of spray cooling with freon113, Int. J. Heat Mass Transfer 34 (4–5) (1991) 1163–1174. [17] A. Bolukbasi, D. Ciloglu, Pool boiling heat transfer characteristics of vertical cylinder quenched by SiO2–water nanofluids, Int. J. Therm. Sci. 50 (6) (2011) 1013–1021.
[18] I.L. Ferreira, J.E. Spinelli, J.C. Pires, A. Garcia, The effect of melt temperature profile on the transient metal/mold heat transfer coefficient during solidification, Mater. Sci. Eng., A 408 (1–2) (2005) 317–325. [19] S.S. Hsieh, H.Y. Leu, H.H. Liu, Spray cooling characteristics of nanofluids for electronic power devices, Nanoscale Res. Lett. 10 (2015) 139. [20] H.S. Ahn, M.H. Kim, The boiling phenomenon of alumina nanofluid near critical heat flux, Int. J. Heat Mass Transfer 62 (2013) 718–728. [21] H. Kim, H.S. Ahn, M.H. Kim, On the Mechanism of Pool Boiling Critical Heat Flux Enhancement in Nanofluids, J. Heat Transfer 132 (6) (2010). 061501-1-11. [22] K.J. Park, D. Jung, S.E. Shim, Nucleate boiling heat transfer in aqueous solutions with carbon nanotubes up to critical heat fluxes, Int. J. Multiph. Flow 35 (6) (2009) 525–532. [23] C. Gerardi, J. Buongiorno, L. Hu, T. McKrell, Infrared thermometry study of nanofluid pool boiling phenomena, Nanoscale Res. Lett. 6 (2011) 232. [24] W. Jia, H.H. Qiu, Experimental investigation of droplet dynamics and heat transfer in spray cooling, Exp. Thermal Fluid Sci. 27 (7) (2003) 829–838.