Enhanced flow boiling in microchannels using auxiliary channels and multiple micronozzles (II): Enhanced CHF and reduced pressure drop

International Journal of Heat and Mass Transfer 115 (2017) 264–272

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer journal homepage: www.elsevier.com/locate/ijhmt

Enhanced flow boiling in microchannels using auxiliary channels and multiple micronozzles (II): Enhanced CHF and reduced pressure drop Wenming Li a, Tamanna Alam a, Fanghao Yang b, Xiaopeng Qu a, Benli Peng a, Jamil Khan a, Chen Li a,⇑ a b

Department of Mechanical Engineering, University of South Carolina, Columbia, SC 29208, United States Princeton Plasma Physics Laboratory, Princeton, NJ 08540, United States

a r t i c l e

i n f o

Article history: Received 15 January 2017 Received in revised form 7 August 2017 Accepted 11 August 2017

Keywords: Flow boiling Microchannel Pressure drop Critical heat flux Two-phase flow instability

a b s t r a c t Enhancing critical heat flux (CHF) of flow boiling without escalating pressure drop is highly desirable in thermal management of high power-density electronic devices. Usually, an improved CHF can be achieved by restricting flow or at a higher mass velocity, leading to a higher pressure drop. In this study, compared to the two-nozzle microchannel configuration, the improved microchannel configuration as detailed in the Part (I) of this study can enhance CHF without sacrificing pressure drop. In this part, CHF is experimentally evaluated together with the pressure drop with mass flux ranging from 120 kg/m2 s to 600 kg/m2 s. Compared to the two-nozzle configuration, our study shows that CHF can be enhanced up to 32% with a
53% reduction of pressure drop at a mass flux of 325 kg/m2 s. The bubble collapse-removal process is significantly improved because more micronozzles are integrated. The enhanced pumping effect, which is created by rapid bubble collapse processes in the entire main channels, enables a more sustainable liquid supply and hence delays the CHF conditions. Moreover, two-phase flow in terms of pressure drop fluctuations is more stable owing to the effective management of bubble confinement in the entire channel. Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction Flow boiling in microchannels is one of the most promising cooling techniques for microelectronics [1]. Using latent heat by vaporization can significantly improve heat dissipation of high power density electronic devices. However, the vigorous and rapid generation of vapor through phase change leads to adverse twophase flows in microchannels due to the well-known bubble confinement effect [2], which can lead to unfavorable two-phase flow instabilities in terms of low frequency and large amplitude of wall temperature and pressure drop/flow fluctuations. These could trigger premature CHF conditions and result in high pressure drop [2]. CHF and pressure drop are two of the most critical factors in evaluating flow boiling performance in two-phase microchannel cooling. Usually, it is highly desirable to enhance CHF without escalating two-phase pressure drop. Nonetheless, it is challenging to achieve this goal that appears conflicting. In last decades, numerous techniques have been developed to improve CHF by regulating bubble slugs [2,3], suppressing flow instabilities [4,5], modifying surface properties [6–11], and promoting liquid rewetting [12–15]. A comparison of experimental CHF data has been reported ⇑ Corresponding author. E-mail address: [email protected] (C. Li). http://dx.doi.org/10.1016/j.ijheatmasstransfer.2017.08.032 0017-9310/Ó 2017 Elsevier Ltd. All rights reserved.

in our previous study [16]. The microchannels with inlet restrictors (IRs) have been proved to effectively delay CHF conditions by controlling vapor flow reversal, but the frequency of periodic bubbles/vapor slugs growth-collapse process is low [2]. Promotion of bubble collapse/removal can provide a more effective approach to regulate the two-phase flows in microchannels. Enhanced pumping effect, which refers to the high frequency periodic growth-collapse of bubbles/vapor slugs, can facilitate adequate liquid supply and eventually, delay CHF conditions. In our previous study, a novel concept was developed by rapidly collapsing confined bubbles [17]. With this concept, CHF is enhanced by improving the liquid supply to main channels through rapid bubble-collapse-induced jetting flows and the resulting enhanced pumping effect [17,18]. However, the two nozzles located at the middle of each main channel cannot remove the elongated bubbles effectively in the entire channel, particularly, in the upstream. Thus, persistent vapor slugs were observed near the inlet region during the fully developed boiling. To address this issue, an improved design has been developed to extend the effect of high frequency jetting flows on bubble collapse/removal to the entire channel [19]. It would be highly desirable to enhance CHF without escalating two-phase pressure drop. Numerous techniques have been proposed to address this issue. These include using diverging channels

W. Li et al. / International Journal of Heat and Mass Transfer 115 (2017) 264–272

265

Nomenclature a A Cc Cf D H hfg f G kd L _ m N Dp P S ST u  u W z

c

channel aspect ratio area, m2 parameter for exit effect frictional coefficient diameter, m height, m latent heat of vaporization, kJ/kg fanning friction factor, m mass flux, kg/m2 s momentum correction factor, W/m K length, m mass flow rate, kg/s number of pin fins pressure drop, Pa power, W slip coefficient transverse pitch, m velocity, m/s average velocity, m/s microchannel width, m coordinate, m

surface tension, N/m contact angle viscosity, kg/(s m) density, kg/m3 vapor quality interfacial stress, N/m2 flow area contraction ratio

h

l q v ri s

Subscripts 2/ two-phase acc accelerational c cross section CHF critical heat flux e exit eff effective exp experimental f frictional h hydraulic i inlet l liquid m mixture o outlet sat saturated v vapor

Greek symbols a void fraction b kinetic energy correction factor

[20,21], stepped fin microchannels [22] and taper channels [5,23]. However, these techniques did not consider rapidly removing the confined bubbles. Since the accelerational and frictional pressure drop would be elevated by the rapid bubble expansion, the management of bubble collapse/removal could be promising to reduce accelerational and frictional pressure drop. In our previous study, the two-nozzle configuration was shown to effectively remove the confined bubbles in microchannels [17], leading a significant reduction of pressure drop compared to the plain wall microchannel with IRs [2]. A further reduction of pressure drops owing to an efficient bubble removal has been demonstrated in our early study of the improved four-nozzle design [19] and will be further studied herein. A low collapse frequency of elongated bubbles and flow reversal due to vigorous vapor generation in conventional microchannels is the key factor resulting in an obvious increase of two-phase pressure drop compared to single phase flow at a specific flow rate [24]. Based on our previous studies in developing the two-nozzle microchannel configuration [16,17], an improved microchannel configuration with an aim at enhancing CHF without elevating pressure drop is developed in this study. The global liquid supply would be significantly improved owing to two more evenlydistributed nozzles in each main channel. Pressure drop would be effectively managed owing to increased bypass flow areas and the enhanced pumping effect induced by high frequency bubble growth-collapse processes.

where the total pressure drop (Dptotal ) was measured by two pressure transducers at the pressure-drop measuring ports; Dpdistributor , Dpplenum , Dpi and Dpo are the pressure drop of the flow distributor, pressure drop of the plenum chambers, inlet minor losses and exit minor losses, respectively. Accelerational pressure drop (Dpacc ) is given by: 2

Dpacc ¼ G

qi



!

1

ð2Þ

q2/;o

_ c ; and the density of the where mass flux is estimated from G ¼ m=A exit two-phase flow q2/;o is estimated by

q2/;o ¼ a  qv þ ð1  aÞ  ql

ð3Þ

where ql and qv denote the densities of liquid and vapor, respectively; the density of inlet flow is assumed to be the density of liquid flow (ql  qi ), and a is the void fraction, which can be estimated by

a¼1

  1  v qv 1þ S

v

ð4Þ

ql

where S is the slip ratio, which is calculated based on Chisholm corrffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi   relation: S ¼ 1  v 1  qql [25]. v

The inlet minor losses is estimated from the model developed in [26]:

Dpi ¼ ql

2. Data reduction

1

 l;i 1  b  s2  C 2c  2C c þ 2C c  K d u 2 Cc

ð5Þ

The exit minor losses is calculated using an equation developed by Abdelall [26]:

2.1. Two-phase frictional pressure drop Experimental frictional pressure drop (Dpf ;exp ) is deduced from the following equation [24]:

 v ;o 1  2C c  s þ s2 ðK d  1Þ u  ð1  s2 Þ Dpo ¼ qv 2 2

Dpf ;exp ¼ Dptotal  Dpacc  Dpi  Dpo  Dpplenum  Dpdistributor

where the parameter (C c ) is given by Geiger [27]:

ð1Þ

ð6Þ

266

W. Li et al. / International Journal of Heat and Mass Transfer 115 (2017) 264–272

Cc ¼ 1 

1s 2:08ð1  sÞ þ 0:5371

ð7Þ

 l;i and u  v ;o represent the average liquid velocities at inlet and exit, u  l;i ¼ m=ðA _  v ;o ¼ respectively. They are estimated as u l  ql Þ and u _  ve =ðAv  qv Þ. The momentum correction factor (K d ¼ 1:33) and m b ¼ 2 are used for laminar flow [24]. s is the flow area contraction ratio. The pressure drop of the plenum chambers consists of two components:

Dpplenum ¼ Dpi;plenum þ Dpo;plenum

ð8Þ

where the pressure drop of the inlet plenum is estimated from

Dpi;plenum ¼ f l;plenum  ql

 2plenum Lplenum u Dplenum 2

ð9Þ

and the pressure drop of the exit plenum is estimated from

Dpo;plenum ¼ f v ;plenum  qv

 2v ;plenum Lplenum u Dplenum 2

ð10Þ

where Lplenum is the length of the plenums; and Dplenum is the hydraulic diameter of the plenums. The friction factor is calculated as

f plenum ¼

Cf Replenum

ð11Þ

where Cf is adopted from [28]. The pressure drop of the flow distributors is given by

Dpdistributor ¼ N  f cross  ql



ST ST  Dpin

2  2 uplenum 2

ð12Þ

where N is the number of pin fins in a single row of flow distributers; fcross was estimated by Gaddis’s method [29]; and ST is the transverse pitch; and Dpin is the hydraulic diameter of pin fins. In this experimental study, the Reynolds number of vapor flow is defined as,

q ðuv  ul ÞDh Re ¼ v

lv

where

ð13Þ

lv is the dynamic viscosity of vapor.

3. Results and discussion 3.1. Calibration To assure the accuracy of pressure drop measurements, the single phase friction factor in smooth wall rectangular microchannels is calibrated using several correlations in literature, and then compared to the theoretical value [28] (as shown in Fig. 1). In Eq. (1), the pressure drop induced by the flow distributor is relative larger than other pressure drop components such as losses from plenum chambers, inlet minor and exits. To find a suitable correlation, four friction factor correlations (as listed in Table 1) for the flow distributor are compared to calculate Dpdistributor in estimating a single phase friction factor. Fig. 1 shows that the single phase friction factor of the smooth wall rectangular microchannel using Gaddis’ correlation [29] is closer to the theoretical values. Therefore, the Gaddis’ correlation is selected to estimate the flow distributor frictional pressure drop in this study. 3.2. Reduced pressure drop The measured total two-phase pressure drop is mainly determined by frictional and accelerational pressure drops [24,33]. Fig. 2 shows the measured total single phase/two-phase pressure drops versus effective heat flux and exit vapor quality, respectively. Single phase pressure drop is determined by mass flux for a given channel configuration. The slopes of pressure drop curves in twophase region dramatically increase with mass flux and exit vapor quality. The change of flow regimes as indicated by exit vapor quality would be the main reason for the sharp elevation of two-phase pressure drops. The uncertainty of measured pressure drop is ±7.1%. The uncertainty propagation is calculated using methods developed by Kline and McClintock [34]. The comparison of total pressure drops is conducted between the current design and the previous two-nozzle configuration [17] to investigate the mechanism of pressure drop reduction. Fig. 3 shows that a significant reduction of pressure drop, up to 56%, is achieved at a mass flux of 430 kg/m2 s. In conventional microchannel configurations, the elongated vapor slug is hard to be removed. Hydraulic resistance was reduced by managing vapor slug expansion rate in our previous study [24]. The overall construction of the present design is identical to the previous two-nozzle configuration, except for additional two nozzles and shared auxiliary channels [17]. With the additional two

2.2. Single phase friction factor in a smooth wall rectangular microchannel

Friction factor correlations for pin fins of inlet distributor: Gunther et al. [30] Gaddis et al. [29] Short et al. [31] Kosar et al. [32] Friction factor theoretical model: Theoretical model

0.16

f ¼

qDpDh 2LG2

ð14Þ

where Dp is the pressure drop, Dh is the hydraulic diameter, L is the passage length, and G is the mass flux. The friction factor is a function of aspect ratio for a rectangular passage. It can be determined using equation from Shah and London [28].

f ¼ 24ð1  1:3553a þ 1:9467a2  1:7012a3 þ 0:9564a4  0:2537a5 Þ=Re

Friction factor

The friction factor in terms of the pressure drop and mass flux is given by:

0.12

0.08

0.04

0.00

ð15Þ

where a is the channel aspect ratio and Re is the Reynolds number of liquid flow. One constraint for this equation is that the aspect ratio must be less than one. If the channel aspect ratio is greater than 1.0, the inverse is taken to use with Eq. (15).

0

200

400

600

800

1000

1200

Re Fig. 1. Single phase friction factor of a smooth wall rectangular microchannel array is calibrated using different correlations of friction factor. These correlations are used to estimate pressure across the fluid distributor. The results show that Gaddis’ correlation agrees better with the theoretical model [28].

267

W. Li et al. / International Journal of Heat and Mass Transfer 115 (2017) 264–272 Table 1 List of single phase friction factor correlations. Reference

Shape of pin fins

Working fluid

Correlations of friction factor

Gunther and Shaw [30]

Circular

Air

f ¼ 180 Re



0:4  0:6 dbG 1 ST

4ST SL pd2

bG ¼ SdL for in-line tube configuration Gaddis and Gnielski [29]

Circular

All fluids

bG ¼ SdT for staggered tube configuration   2

0:5 f ¼

280p

SL d

 Re

0:6



þ0:75

p C 1:6

SL ST  d2

Where C ¼ SdT for in-line/staggered arrangement with qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi SL P 12 2 SdT þ 1; d

Short et al. [31]

Circular 100 < Re < 1000 2 < ST =d < 6:41 1:83 < SL =d < 3:21 H=d ¼ 2:43 Circular

Kosar [32]

Air

Water

C ¼ Sdd for in-line/staggered arrangement with qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi SL < 12 2 SdT þ 1 d  1:3  0:78  0:55 Hfin ST f ¼ 140:4 SdL Re0:65 dLehsNL de de

f ¼ 1739 Re1:7



1:1  0:3 Hfin =de SL ST Ac Hfin =de þ1

þ 345 Red



2 

1 Hfin =de þ1

SL ST Ac

0:3

Red ¼

qf umax dhe lf

Lhs dhe ¼ 4AAmin , Awet ¼ N t Afin þ 2ðAt  N t Ac Þ wet

2

Effective heat flux (W/cm ) 120

0

50

100

150

200

250

300

Pressure drop ( kPa )

2

100 80 60 40 20 0

120 kg/m s 2 150 kg/m s 2 180 kg/m s 2 325 kg/m s 2 380 kg/m s 2 430 kg/m s 2 600 kg/m s

Two-phase

Single phase

0.0

0.2

0.4

0.6

Exit vapor quality Fig. 2. Single phase/two-phase pressure drops versus effective heat flux and exit vapor quality, respectively.

nozzles integrated, the vapor bubbles can be removed more effectively, which could further reduce pressure drop owing to increased bypass flow area. As shown in Fig. 3, the present design can significantly reduce total pressure drops compared to previous two-nozzle configuration [17]. 3.2.1. Dominance of the frictional pressure drop Fig. 4 shows a comparison of total pressure drop components versus vapor quality at a mass flux of 430 kg/m2 s in the present four-nozzle microchannel configuration. Frictional pressure drop is the dominant component of the total pressure drop, which is consistent with literature [24,35]. 3.2.2. Reduced accelerational pressure drop (Dpacc ) and exit minor loss (Dpo ) Fig. 5 shows that the exit vapor quality of the present study is lower than that of two-nozzle configuration at the same effective heat flux. Two-phase flow density would be larger at a lower exit vapor quality at the same heat flux. Eq. (2) indicates that a larger

Fig. 3. A significant reduction of pressure drop is achieved in the present configuration compared to the two-nozzle configuration [17] at 150 kg/m2 s (a) and 430 kg/m2 s (b), respectively.

two-phase density can result in larger accelerational pressure drop, but a smaller value of Dpacc (as shown in Fig. 6(a)). Besides the accelerational pressure drop and exit minor losses, other components in Eq. (1) are calculated as well. The inlet minor loss (Dpi ) is determined by mass fluxes. For example, Dpi are 1.43 kPa and

268

W. Li et al. / International Journal of Heat and Mass Transfer 115 (2017) 264–272

Pressure drop (kPa)

80

60

ΔPtotal ΔPf,exp −ΔPacc −ΔPo

40

20

0 0.10

0.15

0.20

0.25

0.30

0.35

0.40

Exit vapor quality Fig. 4. Total, frictional, accelerational pressure drops and exit minor losses versus vapor quality at a mass flux of 430 kg/m2 s in the present configuration microchannels.

Fig. 6. (a) The accelerational pressure drop (Dpacc ) and (b) exit minor losses (Dpo ) of the present study are reduced compared to those of two-nozzle configuration at a mass flux of 150 kg/m2 s and 430 kg/m2 s, respectively.

Fig. 5. (a, b) A comparison of exit vapor quality against effective heat flux is plotted at 150 kg/m2 s and 430 kg/m2 s, respectively.

11.74 kPa at mass flux of 150 kg/m2 s and 430 kg/m2 s, respectively. However, the pressure drops induced by the flow distributor and plenum chambers are very small, less than 1 kPa in this study.

3.2.3. Reduced frictional pressure drop Fig. 7 shows that the frictional pressure drop is plotted versus effective heat flux and exit vapor quality, respectively. A significant reduction of
56% and
40% are achieved for a given effective heat flux and vapor quality at a mass flux of 430 kg/m2 s, as shown in Fig. 7(b) and (d), respectively. Two major factors are believed to contribute to the pressure drop reduction. The first is the bypass effect induced by additional auxiliary channels. It has been demonstrated in our previous study that the frictional pressure drop can be reduced by bypass effect through micronozzles and auxiliary channels [17,18]. At the same exit vapor quality, the augmented flow area can result in a substantial reduction of vapor superficial velocity. For example, at a mass flux of 150 kg/m2 s, the superficial velocity of vapor flow is 97.7 m/s at an exit vapor quality of 0.49 under a heat flux of 228 W/cm2. A reduction of
22% is achieved in the present study compared to the two-nozzle microchannel configuration. At a higher mass flux of 430 kg/m2 s, the superficial velocity of vapor flow is 191 m/s at an exit vapor quality of 0.34 under a heat flux of 504 W/cm2. A reduction of
24% is achieved as well. The reduction of superficial velocity of vapor flow leads to such a significant reduction of frictional pressure drop compared to the two-nozzle microchannel configuration [17] as shown in Fig. 7(c) and (d). The other factor should be the management of bubble or vapor slug expansion, which leads to additional reduction of frictional

W. Li et al. / International Journal of Heat and Mass Transfer 115 (2017) 264–272

269

Fig. 7. The frictional pressure drops are plotted versus effective heat flux and exit vapor quality at mass flux of 150 kg/m2 s and 430 kg/m2 s, respectively. Drastic reduction of frictional pressure drop is achieved with the present configuration compared to two-nozzle configuration for a given working heat flux or exit vapor quality.

pressure drop as indicated by vapor quality and effective heat flux. In microchannel flow boiling, the rapid growth and expansion of confined bubbles toward upstream can hinder and block incoming flows, resulting in a substantial increase of pressure drop [17]. Well-managed vapor slug expansion has an effect on reducing frictional pressure drop [24]. Extended mixing in the entire channel is also achieved by increased and well-distributed jetting flows. Fig. 8 shows that confined bubble shrinks with the effect of jetting flows.

Usually, the contribution of the frictional pressure drop to the overall pressure drop is larger at a higher heat flux and vapor quality [24,35]. In this study, the reduction of frictional pressure drop is shown to be larger at a higher heat flux and vapor quality, as illustrated in Fig. 7(c) and (d). Under higher mass fluxes, the reduction of frictional pressure drop is more significant. For example, the reduction of frictional pressure drop is
26% at an exit vapor quality of 0.5 and a mass flux of 150 kg/m2 s; while such a reduction increases up to 40% at an exit vapor quality of 0.34 at a mass flux of 430 kg/m2 s. In the low mass flux scenario, as shown in Fig. 7(c), when the exit vapor quality is less than 0.4, the difference of frictional pressure drop between the present study and two-nozzle configuration is minimal. As vapor quality reaches above 0.4, the difference of frictional pressure drop becomes more significant and shows higher differences with increased vapor quality. The reduction is up to 26% at a vapor quality of around 0.5. 3.3. Two-phase flow instabilities

Fig. 8. Mixing induced by rapid bubble collapse is observed in the present study at a heat flux of 200 W/cm2 and a mass flux of 325 kg/m2 s. (Scale bar is 100 mm).

Fig. 9 shows the comparison of the Dp-G curves between the present study, the two-nozzle microchannels [17], and microchannels with IRs [2] at a heat flux of 150 W/cm2 and 250 W/cm2, respectively. The slopes of Dp-G curves for microchannels with IRs are positive, which enabled high CHF by suppressing twophase reversal flows, but at a cost of pressure drop. In the twonozzle microchannels [17], the slopes were greatly moderated compared to the microchannels with IRs and the slopes of Dp-G curves at 150 W/cm2 are even positive. In the present configuration, the slope of Dp-G curves is similar with that of the classic Dp-G curve, but not so steep.

270

W. Li et al. / International Journal of Heat and Mass Transfer 115 (2017) 264–272

as shown in Fig. 10. In section c-e, the slope of Dp-G curve is negative. In microchannels, it has been argued that the OFI corresponds to a mass flux only slightly less than the ONB. In this study, the mass flux of onset of nucleation bubble (ONB) is 650 kg/m2 s at 150.6 W/cm2, and 900 kg/m2 s at 256 W/cm2, as shown in Fig. 9. As discussed by Boure et al. [37], when @ðDpÞ j @G channelpumping

6 0, the system is susceptible to the static Ledinegg instability mainly in the parallel channels. In addition, upstream compressible volume instability can be significant at microscale. Such type of two-phase flow instabilities is characterized by low frequency and large amplitude of flows, pressure drops, and wall temperature oscillations and could result in premature CHF conditions. Microchannels with IRs [2] have been employed and demonstrated as an effective method to mitigate two-phase flow instabilities by reshaping the Dp-G curve such that the curve is rectified into positive. The flow instability of the system can be assessed by considering the negative slope of Dp-G curves. As the slope becomes steeper, the system is more likely to exhibit flow instabilities, leading to a premature CHF condition. There are several factors that affect the curve’s slope, including heat flux, saturation pressure, mass flux, and channel hydraulic diameter. This can be expressed as follows [36]:

dðDPÞ ¼ f ðPsat ; G; DT sub;i ; q00 ; Dh ; L; fluidÞ dG

Fig. 9. Comparisons of Dp-G curves of flow boiling between the present configuration, the two-nozzle configuration [17] and microchannels with IRs [2].

Fig. 10 depicts the classic Dp-G curve in microchannels [36], which is used to analyze the Ledinegg instability of the present study. Notice that the Dp-G curves (Fig. 9) of the present configuration are similar with the classic one in shape as shown in Fig. 10. For a given heat flux, the pressure drop increases with increasing mass flux in the regime of single phase flow before ONB. In the section a-b, the slope of Dp-G curve is positive. Conversely, after reaching the onset of flow instability (OFI) point, pressure drop increases with decreasing mass flux in the regime of two-phase

Superheated Single-phase Vapor Flow

ð16Þ

Using microfluidic transistors (auxiliary channel combined with multiple micronozzles), up to 78% pressure drop reduction was demonstrated in the two-nozzle microchannel configuration [17,18]. A further 51% pressure drop reduction has been realized in the present configuration. High frequency oscillations lead to more stable flow boiling in microchannels. Therefore, in twophase regime (Fig. 9), the slope of Dp-G curves is still negative, but remaining small. The microchannel integrated auxiliary channels combined with multiple micronozzles is mainly contributed to the reduced negative slopes. Fig. 11 shows the transient pressure drop of present microchannels compared to that of the two-nozzle microchannels at an effective heat flux of
300 W/cm2, for a given mass flux of 380 kg/m2 s. At such a high heat flux, pressure drop has been observed to fluctuate within 1.4% (±0.6 kPa), which confirms a stable two-phase flow achieved in the present microchannel configuration. The standard deviation of the present design is
28.9% lower than that of the two-nozzle microchannels (in Fig. 11). Also, the two-nozzle

Subcooled Single-phase Liquid Flow

Bulk Boiling

Fully e Developed Boiling

p

a f

Negative Slope d Unstable Boiling

c

b

Constant Heat Flux

ONB OFI

G Fig. 10. Classic Dp-G curve of flow boiling under constant heat flux in plain wall microchannels [17].

Fig. 11. Enhanced two-phase flow stability in the present study compared to the two-nozzle configuration [17].

W. Li et al. / International Journal of Heat and Mass Transfer 115 (2017) 264–272

271

design shows high amplitude oscillations of
0.5 kPa whereas the four-nozzle design shows lower amplitude oscillations of
0.3 kPa as illustrated in Fig. 11. High amplitude of oscillations would escalate flow boiling instabilities. Hence, enhanced two-phase flow stability is achieved in the present configuration. A better management of the confined compressible vapor bubbles, which is enabled by four evenly-distributed micro-nozzles, likely results in more stable two-phase flows. 3.4. Enhanced CHF In this study, CHF is triggered by the burnout and failure of the tested device due to the crisis of liquid supply to the heating surface. In experiments, it is determined by a spike wall temperature reading. By generating high frequency two-phase oscillations, CHF has been achieved to 1020 W/cm2 at a mass flux of 1350 kg/m2 s in the two-nozzle microchannel configuration [17]. The present configuration aims to achieve better global liquid supply and to enhance local rewetting through four evenly-distributed nozzles along the side wall of each main channel. Noticeable enhancement of CHF has been achieved as illustrated in Fig. 12. The enhancement is up to ’35% at a low mass flux of 200 kg/m2 s, and then drops to
10% at a higher mass flux of 600 kg/m2 s. The uncertainty value of CHF is ±10 W/cm2 [34]. In Fig. 13, the enhanced liquid supply in the present configuration is visualized and compared with the two-nozzle configuration [17] at mass flux of 600 kg/m2 s and heat flux of around 600 W/cm2. The present configuration can enable more sustainable global liquid supply with the high frequency jetting flows induced from auxiliary channels (Fig. 13(a–c)). The increased number of nozzles can maintain a sustainable liquid supply to the entire channel under high work heat fluxes. This is different from the persistent vapor slug near inlet section of main channel in the two-nozzle configuration [17], leading to CHF conditions (Fig. 13(d)). Equally important, the local surface rewetting in the present configuration is significantly improved. For example, Fig. 14 shows the local surface dryout and rewetting process near CHF conditions at a mass flux of 180 kg/m2 s associated with a heat flux of 264 W/cm2. The channel surface is rapidly rewetted by the jetting flows. However, we observed that the two-phase oscillations and jetting flows in the downstream are not as strong and frequent as these in the upstream at a higher mass velocity as illustrated in Figs. 5 and 6 in the part (I) of this study. Fig. 12 shows that the CHF enhancement decreases with increasing mass velocity. The enhancement drops to
10% at a mass flux of 600 kg/m2 s. The

Fig. 13. Enhanced global liquid supply in the present study compared to the twonozzle configuration at a mass flux of 600 kg/m2 s and at heat flux of
600 W/cm2. (a–c) Liquid can be effectively supplied to the main channel through jetting flows through four nozzles when the inlet section of main channel is blocked by the vapor slug. (d) Only a small area near the nozzle is rewetted by the jetting flows in the two-nozzle configuration. (Number 1–4 represents the four locations of micronozzles from inlet to outlet. All scale bars are 200 lm in this figure).

Fig. 14. The surface dryout and rewetting process was observed at a mass flux of 180 kg/m2 s and a heat flux of 264 W/cm2. (Scale bar is 100 lm).

Fig. 12. Significant enhanced CHF is achieved in the present study compared to that of the two-nozzle microchannels [17].

main reason is that the increased inertia of the incoming flow would suppress the two-phase oscillations in auxiliary channels, resulting in degraded liquid supply in the downstream of main channels. At the same mass fluxes, the fluid flow in auxiliary channels is higher than that of the two-nozzle configuration microchannels owing to the augmented areas, resulting from the increased

272

W. Li et al. / International Journal of Heat and Mass Transfer 115 (2017) 264–272

openings of micronozzles. To conclude, the four-nozzle configuration can enhance global liquid supply and local rewetting effectively and is primarily responsible for the increased CHF in the present study compared to the two-nozzle configuration [17]. 4. Conclusions In part (II), the enhanced CHF with reduced pressure drops in the present four-nozzle configuration has been experimentally studied. The key conclusions can be drawn: (a) The total measured pressure drop and frictional pressure drop are significantly reduced in the present study compared to the two-nozzle configuration. The reduction of frictional pressure drop is mainly caused by the increased bypasses owing to the integration of auxiliary channels and by the enhanced pumping effect because of effective management of bubble confinement. (b) The two-phase flow in terms of pressure drop fluctuation becomes more stable in the four-nozzle microchannel configuration compared to that of the two-nozzle microchannel configuration. (c) CHF has been significantly enhanced in the four-nozzle microchannel configuration compared to the two-nozzle microchannel configuration. In addition, we observed that such an enhancement degrades with the increase of mass velocity. Acknowledgments This work was supported by the US Department of Defense, Office of Naval Research under the Grant N000141210724 and N000141612307 (Program Officer Dr. Mark Spector). Devices were fabricated at Institute of Electronics and Nanotechnology (IEN) in Georgia Tech, which are supported by the US National Science Foundation under the Grant ECS-0335765. SEM images were taken at USC Microscopy Center.

Conflict of interest The author declares that there is no conflict of interest. References [1] S.G. Kandlikar, History, advances, and challenges in liquid flow and flow boiling heat transfer in microchannels: a critical review, J. Heat Transf. 134 (3) (2012) 034001. [2] A. Kosßar, C.-J. Kuo, Y. Peles, Suppression of boiling flow oscillations in parallel microchannels by inlet restrictors, J. Heat Transf. 128 (3) (2006) 251–260. [3] G. Wang, P. Cheng, A.E. Bergles, Effects of inlet/outlet configurations on flow boiling instability in parallel microchannels, Int. J. Heat Mass Transf. 51 (9–10) (2008) 2267–2281. [4] C.J. Kuo, Y. Peles, Local measurement of flow boiling in structured surface microchannels, Int. J. Heat Mass Transf. 50 (23–24) (2007) 4513–4526. [5] A. Kalani, S.G. Kandlikar, Combining liquid inertia with pressure recovery from bubble expansion for enhanced flow boiling, Appl. Phys. Lett. 107 (18) (2015) 181601. [6] N. Singh, V. Sathyamurthy, W. Peterson, J. Arendt, D. Banerjee, Flow boiling enhancement on a horizontal heater using carbon nanotube coatings, Int. J. Heat Fluid Flow 31 (2) (2010) 201–207. [7] A.K.M.M. Morshed, F. Yang, M. Yakut Ali, J.A. Khan, C. Li, Enhanced flow boiling in a microchannel with integration of nanowires, Appl. Therm. Eng. 32 (2012) 68–75. [8] C. Li, Z. Wang, P.-I. Wang, Y. Peles, N. Koratkar, G.P. Peterson, Nanostructured copper interfaces for enhanced boiling, Small 4 (8) (2008) 1084–1088.

[9] A. Sitar, I. Golobic, Heat transfer enhancement of self-rewetting aqueous nbutanol solutions boiling in microchannels, Int. J. Heat Mass Transf. 81 (2015) 198–206. [10] M.S. Sarwar, Y.H. Jeong, S.H. Chang, Subcooled flow boiling CHF enhancement with porous surface coatings, Int. J. Heat Mass Transf. 50 (17–18) (2007) 3649– 3657. [11] S.M. You, J.H. Kim, K.H. Kim, Effect of nanoparticles on critical heat flux of water in pool boiling heat transfer, Appl. Phys. Lett. 83 (16) (2003) 3374–3376. [12] H. O’Hanley, C. Coyle, J. Buongiorno, T. McKrell, L.-W. Hu, M. Rubner, R. Cohen, Separate effects of surface roughness, wettability, and porosity on the boiling critical heat flux, Appl. Phys. Lett. 103 (2) (2013) 024102. [13] B.S. Kim, H. Lee, S. Shin, G. Choi, H.H. Cho, Interfacial wicking dynamics and its impact on critical heat flux of boiling heat transfer, Appl. Phys. Lett. 105 (19) (2014) 191601. [14] D.E. Kim, S.C. Park, D.I. Yu, M.H. Kim, H.S. Ahn, Enhanced critical heat flux by capillary driven liquid flow on the well-designed surface, Appl. Phys. Lett. 107 (2) (2015) 023903. [15] K.-H. Chu, R. Enright, E.N. Wang, Structured surfaces for enhanced pool boiling heat transfer, Appl. Phys. Lett. 100 (24) (2012) 241603. [16] W. Li, F. Yang, T. Alam, J. Khan, C. Li, Experimental and theoretical studies of critical heat flux of flow boiling in microchannels with microbubble-excited high-frequency two-phase oscillations, Int. J. Heat Mass Transf. 88 (2015) 368– 378. [17] F. Yang, X. Dai, C.-J. Kuo, Y. Peles, J. Khan, C. Li, Enhanced flow boiling in microchannels by self-sustained high frequency two-phase oscillations, Int. J. Heat Mass Transf. 58 (1–2) (2013) 402–412. [18] F. Yang, X. Dai, C. Li, High frequency microbubble-switched oscillations modulated by microfluidic transistors, Appl. Phys. Lett. 101 (7) (2012), 073509-073504. [19] W. Li, F. Yang, T. Alam, C. Li, Enhanced flow boiling in microchannels using auxiliary channels and multiple micronozzles, in: ASME 2016 5th International Conference on Micro/Nanoscale Heat and Mass Transfer, Singapore, 2016. [20] V.S. Duryodhan, S.G. Singh, A. Agrawal, Liquid flow through a diverging microchannel, Microfluid. Nanofluid. 14 (1–2) (2012) 53–67. [21] C.T. Lu, C. Pan, Convective boiling in a parallel microchannel heat sink with a diverging cross section and artificial nucleation sites, Exp. Therm. Fluid Sci. 35 (5) (2011) 810–815. [22] K. Balasubramanian, P. Lee, L. Jin, S. Chou, C. Teo, S. Gao, Enhanced heat transfer and reduced pressure drop using stepped fin microchannels, in: Electronics Packaging Technology Conference (EPTC), 2011 IEEE 13th, IEEE, 2011, pp. 653–659. [23] S.G. Kandlikar, T. Widger, A. Kalani, V. Mejia, Enhanced flow boiling over open microchannels with uniform and tapered gap manifolds, J. Heat Transf. 135 (6) (2013) 061401. [24] F. Yang, X. Dai, Y. Peles, P. Cheng, J. Khan, C. Li, Flow boiling phenomena in a single annular flow regime in microchannels (II): reduced pressure drop and enhanced critical heat flux, Int. J. Heat Mass Transf. 68 (2014) 716–724. [25] C. Duncan, Two-Phase Flow in Pipelines and Heat Exchangers, Longman Higher Education, London, New York, 1983. [26] F.F. Abdelall, G. Hahn, S.M. Ghiaasiaan, S.I. Abdel-Khalik, S.S. Jeter, M. Yoda, D.L. Sadowski, Pressure drop caused by abrupt flow area changes in small channels, Exp. Therm. Fluid Sci. 29 (4) (2005) 425–434. [27] G.E. Geiger, Sudden Contraction Losses in Single and Two-phase Flow, University of Pittsburgh, 1964. [28] A.L.L.R.K. Shah, Laminar Flow Forced Convection in Ducts, Academic Press, 1978. [29] E.S. Gaddis, V. Gnielski, Pressure drop in horizontal cross flow across tube bundles, Int. Chem. Eng. 25 (1) (1985) 1–15. [30] A.Y. Gunther, W.A. Shaw, A general correlation of friction factors for various types of surfaces in cross flow, Trans. ASME 67 (1945) 643–660. [31] B.E. Short Jr., P.E. Raad, D.C. Price, Performance of pin fin cast aluminum coldwalls. 2. Colburn j-factor correlations, J. Thermophys. Heat Transf. 16 (3) (2002) 397–403. [32] A. Kosßar, C. Mishra, Y. Peles, Laminar flow across a bank of low aspect ratio micro pin fins, J. Fluids Eng. 127 (3) (2005) 419. [33] S.-M. Kim, I. Mudawar, Universal approach to predicting two-phase frictional pressure drop for mini/micro-channel saturated flow boiling, Int. J. Heat Mass Transf. 58 (1–2) (2013) 718–734. [34] S.J. Kline, F.A. McClintock, Describing uncertainties in single-sample experiments, Mech. Eng. 75 (1) (1953) 3–8. [35] J. Moreno Quibén, J.R. Thome, Flow pattern based two-phase frictional pressure drop model for horizontal tubes. Part I: Diabatic and adiabatic experimental study, Int. J. Heat Fluid Flow 28 (5) (2007) 1049–1059. [36] T. Zhang, T. Tong, J.-Y. Chang, Y. Peles, R. Prasher, M.K. Jensen, J.T. Wen, P. Phelan, Ledinegg instability in microchannels, Int. J. Heat Mass Transf. 52 (25– 26) (2009) 5661–5674. [37] A.E.B.J.A. Boure, L.S. Tong, Review of two-phase flow instability, Nucl. Eng. Des. 25 (2) (1973) 165–192.