Flow boiling heat transfer in a rectangular copper microchannel

Abstract

Keywords: Flow boiling; microchannel; heat transfer coefficient; flow patterns; instability

Introduction

Rapid developments in miniaturization and performance of electronics devices resulted in a significant challenge in the thermal management of these devices. The heat dissipation rate from these high performance and compact devices reached huge values and consequently the traditional cooling methods such as air cooling became ineffective. For example, as cited in ref. [1], the heat flux in applications such as high power laser, microwave devices and radars exceeds 10 MW/m 2. Microchannel heat sinks are one of the proposed methods 761

Research Article - JTEN - 2015 - 69 studies on multi-microchannel heat sinks to enable a broad understanding in the topic.

occurrence of flow reversal [17]. Balasubramanian et al. [18] investigated de-ionized water flow boiling in copper rectangular multi-microchannels with hydraulic diameter Dh = 0.342 mm, aspect ratio W/H = 0.167 and two channel lengths 25 and 20 mm. They observed dispersed bubble, intermittent (slug and elongated bubble) and annular flows in their study. In the intermittent flow regime, bubbles nucleate and grow rapidly to the channel size before detaching from the wall with many bubbles nucleating from the side walls.

Flow Patterns Flow visualization in microchannels is important due to its direct link with heat transfer characteristics and pressure drop. The commonly observed flow patterns in large diameter horizontal channels are bubbly, slug, plug, annular, stratified, annular-mist and wavy flow as reported by Thome [13]. According to Kandlikar [14], the dominant flow patterns in microchannels are isolated bubble, confined bubble/slug and annular flow. Several researchers [15-18] studied flow boiling patterns in microchannels. For example, Sobierska et al. [15] performed experiments using water in a single copper vertical rectangular microchannel with hydraulic diameter Dh = 1.2 mm and aspect ratio W/H = 0.43. The total length of the channel was 330 mm and flow visualization was conducted at five locations along the channel. They reported three flow patterns namely bubbly, slug and annular flow but they did not mention where exactly these flow patterns were observed along the channel. It is worth mentioning that bubbly flow was observed under subcooled conditions, i.e. fluid bulk temperature is below saturation. The inlet subcooling was varied from 2 to 20 K. Megahed and Hassan [16] carried out experiments with FC-72 in silicone rectangular multi-microchannels having hydraulic diameter 0.247 mm, aspect ratio W/H = 0.82, and 16 mm length. Flow visualization was conducted at three locations along the transverse direction at the middle of the heat sink and also near the channel exit. The flow patterns reported at one heat flux value of 130.6 kW/m2 and mass flux of 341 kg/m 2s demonstrated that the flow patterns observed along the transverse direction are different. Slug flow was observed in the central channels while wispy-annular flow was observed in the extreme side channels. This indicates that the flow was not uniformly distributed among the channels. Near the channel exit, three main flow patterns were observed in the central channels, namely bubbly, slug and annular flow. Chen and Garimella [17] conducted a visualization study on flow boiling of FC-77 in silicone multimicrochannels with 0.389 mm hydraulic diameter, aspect ratio 1 and 12.7 mm length over a range of heat flux ( 40 – 730 kW/m 2) and at three different mass flux (332, 446 and 570 kg/m 2s). Visualization was conducted near the channel exit at the centerline of the heat sink. They reported that bubbly flow is dominant at low heat flux while bubbles grow and merge faster as heat flux increases resulting in slug flow at intermediate heat flux. As heat flux increased further, alternating wispy-annular flow and churn flow were observed. At these conditions, when the camera was moved to the channel inlet region, flow reversal was observed. Thus, the formation of alternating wispy-annular and churn flow was attributed to the

Flow Instabilities Flow boiling instability is one of the limitations in microchannel heat exchangers. Identifying the operating conditions of stable flow boiling is very important for the reliable design of micro heat exchangers. Some researchers [19-25] investigated flow boiling instability in microchannels. Qu and Mudawar [19] investigated pressure drop and hydrodynamic instability of water flow boiling in a copper multi-microchannel heat sink having hydraulic diameter 0.35 mm, aspect ratio W/H = 0.32 and

conclusion

about the dominant heat transfer mechanism(s) and the effect of heat and mass flux. Also, flow instability is a critical issue that should be highlighted in every boiling experiment. Thus, more research is still required in order to clarify and understand these fundamental points. The objective of the present study is to investigate the effect of heat and mass flux on flow boiling characteristics of de-ionized water in a single rectangular channel made of oxygen free copper. Flow visualization and instability are also investigated.

The single microchannel used in this study was tested first by Mirmanto et al. [31] for high mass flux range 600 – 1005 kg/m2s. In this study the low mass flux range was considered. The test section was made of an oxygen-free copper block with dimensions, 1 mm width (W), 0.39 mm height (H) and 62 mm length (L) resulting in 0.56 mm hydraulic diameter and 2.56 aspect ratio (W/H). Figure 2 depicts the details of the test section. The average surface roughness (Ra) of the bottom surface of the microchannel was measured using a Zygo NewView 5000 surface profiler. The Ra value was measured near the channel inlet, the middle and the outlet and the value was respectively 0.601 μm, 0.450 and 0.387 μm. The test section was heated using one cartridge heater (Elmatic 600 W) inserted below the channel in a direction parallel to the flow. The heater has the same length as the test section to provide uniform heating. Six holes were drilled inside the copper block at 1.1 mm distance below the channel bottom to accommodate 6 K-type thermocouples to measure the local axial wall temperature. The top cover was made of a transparent polycarbonate material to enable flow visualization. The fluid inlet and outlet temperature was measured using K-type thermocouples. All thermocouples were calibrated during the present study with an uncertainty of ± 0.3 K. The pressure drop was measured directly using a differential pressure transducer (Omega PX409-015DWU) which was calibrated with an uncertainty value of ± 0.4 kPa. All data were recorded at a frequency of 1 kHz for 3 min using NIcompact modular data acquisition system with Labview software. Flow patterns were observed with a high-speed, high-resolution camera (Phantom V6) which was located on top of the test section. The camera was integrated with a stereo microscope (Huvitz HSZ-600) for better flow visualization. Data Reduction The heat loss from the test section was estimated by applying an electrical power (P) to the test section when there is no fluid inside. The temperature difference between the bottom wall and the ambient was recorded for each heating power after attaining steady state. The applied power was then plotted versus this temperature difference and the data were fitted to get an equation to calculate the heat loss (Qloss) during boiling experiments. In other words, the temperature difference between the test section bottom wall and the ambient during an actual experiment was used to assess the Qloss.. The percentage of the heat loss varied from 5.2 to 9.2% over the experimental range. The heat flux ( ) was defined as:

Experimental Setup And Data Reduction

Figure 1 depicts a schematic drawing of the experimental facility. The system consisted of a reservoir, a sub-cooler, a magnetically coupled gear pump (Micropump GA-T23, PFSB), ultrasonic flowmeter (Atrato 710-V20-D) with an accuracy of ± 1% of the reading, preheater, test section and a condenser. A separate refrigeration system (not shown in the figure) was used for cooling the condenser and the sub-cooler. The de-ionized water was degassed in the reservoir by vigorous boiling for approximately one and half hour. The noncondensable gases were released to the ambient by opening the valve located on top of the condenser. The de-gassed water was then pumped to the test section and the fluid inlet temperature was controlled using a preheater and a PID controller. Two fine mesh filters (1 mm and 15 µm) were installed in the system to remove any particles in the water.

Research Article - JTEN - 2015 - 69 The top cover was assumed to be adiabatic and the heat flux was assumed to be uniform in the axial and transverse direction of the channel on all three walls. For a fin with an adiabatic tip, the fin efficiency (η) was defined as: (3) (4) The fin efficiency is nearly 0.995 for the present test section. Single phase flow: The local single phase heat transfer coefficient and average Nusselt number Nu are calculated as:

(6) where, the channel wall temperature at axial location z corrected using the 1D heat conduction equation as given by eqn. (7) below. Dh, kl and Tf(z) are calculated by eqn. (8) below based on energy balance assuming uniform heat flux boundary condition.

(7) (8) In the above equations, the distance from the channel bottom to the thermocouple location (t) is 1.1 mm. The single phase fanning friction factor f is calculated as:

(b) Figure 2: Test section details (a) front view, (b) dimensions of the test section in mm, Mirmanto et al. [31].

where Aht is the heat transfer area, which is defined below as:

is defined by eqn. (11) below is the pressure loss due to the inlet and outlet manifolds and the sudden contraction and enlargement. (11) , and values were taken from Shah and London [32]. The flow enters and leaves the channel in a direction normal to the flow direction. Two phase flow: Since the fluid enters the channel as a sub-cooled liquid, the channel is divided into a single phase region and a two phase region. The single-phase region starts from the channel inlet to the location of zero thermodynamic quality with length Lsub. Thus, the length of the two phase region becomes Lsat = L - Lsub. The length of the single-phase region Lsub is calculated iteratively using eqns. (12-14).

(12) (13) (14) (b) Figure 3: Single-phase validation, (a) f-Re, (b) Average Nu-Re

is the apparent friction factor that was taken from Shah [33]. The values of the constant C, fFDRe, and K(∞) in eqn. (14) are given in Shah [33] for rectangular channels. The dimensionless length in eqn. (14) is defined as:

The single phase pressure drop part in the above equation is calculated based on the length of the single phase region Lsub calculated using eqns. (12-14) and the apparent friction factor defined by eqn. (14). The local two phase heat transfer coefficient was defined as:

The local pressure in the two-phase region was assumed to decrease linearly with the axial distance z and it can be calculated as:

The local saturation temperature in the above equation is calculated based on the local pressure given by eqn. (16) above. The thermodynamic mass quality was computed using eqns. (19) and (20) below.

where the net two phase pressure drop across the channel is defined as: 766

(19) ii is calculated based on the measured Pi and Ti. The experiments were conducted over a mass flux range 200 – 800 kg/m2s. In the tests, the mass flux was kept constant and the heat flux was increased gradually in small steps. The propagated uncertainty analysis was conducted according to the method explained in Coleman and Steel [34] and the average values were 6.3 % for hsp,

Results And Discussion

Boiling Incipience Instability Figure 4 depicts the fluctuations in the fluid inlet temperature, fluid outlet temperature, wall temperature and pressure drop for G = 200 kg/m2 s. Figure 4(a, b) displays the signals at zero heat flux (before applying heat to the test section) and indicates very stable signals with insignificant fluctuations. The amplitude of fluctuations were 0.10 K, 0.06 K, 0.1 K and 0.31 kPa for the fluid inlet, outlet and wall temperature and pressure drop, respectively. When the heat flux was increased to 64.9 kW/m2, boiling started and small bubbles were observed inside the channel. At this condition, the signal demonstrated large periodic fluctuations as seen in Figure 4(c, d). The fluctuations amplitude for the inlet, outlet and wall temperatures and pressure drop signal were 5 K, 4.5 K, 0.1 K and 3.9 kPa respectively. As the heat flux increased, the periodic nature of the signal disappeared and non-periodic fluctuations were observed for all heat flux values. Additionally, after boiling incipience, the fluctuations in fluid inlet temperature was higher than that in the outlet temperature. The wall temperature does not show significant fluctuations where the amplitude was about 0.3 K.

(c) Fluctuations in Tinlet, Toutlet and ∆P at q”= 64.9 kW/m2

Research Article - JTEN - 2015 - 69 expands rapidly in the upstream and downstream sides of the channel.

(d) Fluctuations in Twall at q”= 64.9 kW/m2 Figure 4: Fluctuations in fluid inlet, outlet and wall temperature and pressure drop at boiling incipience heat flux at G = 200 kg/m2s

As the mass flux increases, the fluctuations in the above mentioned parameters were only observed at boiling incipience and became insignificant as heat flux increases. For example, Fig. 5 depicts the fluctuations recorded at boiling incipience for G = 400 kg/m2s. When boiling started at q” = 99 kW/m2, significant fluctuations in the temperatures and pressure drop were recorded. The fluctuations amplitude at this condition was 2 K, 5 K, 0.1 K and 5 kPa respectively for the inlet, outlet and wall temperature and pressure drop. Analyzing the pictures captured using the high speed camera at boiling incipience indicated flow reversal, as seen in Fig. 6 for G = 400 kg/m2s as an example. The flow reversal can also be detected from the pressure drop signal in Figs. 4 and 5 where the signal shows negative values. The flow visualization was conducted at three locations along the channel, near the channel inlet, at the middle and near the channel outlet. Figure 6 displays the periodic evolution of flow patterns at boiling incipience for G = 400 kg/m2s at the middle of the channel. The freehand sketch (red lines) indicates the boundaries of the bubbles to clarify the visualization. The figure demonstrates that small bubbles nucleate at the channel corners at time designated as t = 0 ms. After 5 ms, one bubble grew and spanned the channel width with smaller bubbles still observed at the channel edges. After 6 ms, the big bubble merged with the small bubbles and an elongated bubble was formed. The tail of this elongated bubble seems deformed (no spherical shape) while the nose of this bubble is not shown in the picture, see Fig. 6 (t = 7 ms). This picture shows also another bubble with a spherical head coming from the inlet region. These pictures demonstrate that bubbles nucleate, grow and merge rapidly before departure from the nucleation sites. This results in the formation of an elongated bubble which

(b) Fluctuations in Twall at q”= 99 kW/m2 Figure 5: Fluctuations in fluid inlet, outlet and wall temperature and pressure drop at boiling incipience heat flux at g = 400 kg/m2 s

The nose of the bubble moves faster than the tail. This could be attributed to the partial blockage of the channel resulting from the rapid growth of the bubble without departure from the nucleation site. Thus, the quantity of liquid in the downstream side of the channel becomes smaller compared to the upstream side. Given that the heat flux is constant, this can result in higher evaporation rate downstream of the bubble (thin liquid film have less thermal resistance). The pictures indicate also that the tail of the bubble moves slowly towards the channel inlet (back flow) until another bubble comes from the entry region of the channel. After 9 ms, the picture shows “annular-like” flow patterns but it is not annular flow. This pattern resulted from the coalescence of elongated bubbles. Additionally, this “annular-like” flow pattern was prevailing when the camera was moved to a location near the channel inlet and outlet. It is interesting to note that, after 32 ms, fresh liquid entered the channel and small 768

Research Article - JTEN - 2015 - 69 bubbles were observed to nucleate near the channel corners. This means that there is a periodic evolution of these flow patterns and the period is 32 ms for this mass flux. This behaviour was also observed for all mass fluxes but the period decreased with increasing mass flux, i.e. 68 ms for G = 200 kg/m2s, 32 ms for G = 400 kg/m2s, 21 ms for G = 600 kg/m2s and 14 ms for G = 800 kg/m 2s.

values, the effect of heat flux was very small but the heat transfer coefficient decreases with increasing heat flux. This behaviour could be attributed to the rapid bubble growth and elongation along the channel at high heat flux, which might suppress the nucleation process. Additionally, the evaporation rate of the liquid film around the elongated bubble may be so high, such that dry patches might form underneath the bubble.

t = 32 ms t = 5 ms Figure 6: Sequence of flow images at the middle of the channel at boiling incipience for tin = 89 °c, g = 400 kg/m2s and q” = 99 kw/m2

Heat Transfer Results Figure 7 shows the effect of heat flux on the local heat transfer coefficient plotted versus local vapour quality (hx) and axial distance (h-z) for G = 200 kg/m 2s at low heat flux values. As seen in the figure, boiling started near the middle of the channel (see the sudden increase in the heat transfer coefficient) with a clear small heat flux effect in the region near the channel exit. This could be interpreted as nucleate boiling dominant mechanism. The small heat flux effect observed at the lowest heat flux could be attributed to the absence of local dryout during the nucleation cycle. As discussed above in Fig. 6, there is bubble nucleation, growth and expansion is periodic. During this process, the elongated bubble seems stagnant for a while before the fresh liquid pushes it to the downstream side. Thus, at low heat flux values, it is probable that the elongated bubble leaves the channel before the liquid film dries out. Note for q”= 135.6 kW/m 2 the last thermocouple shows a large decrease in the heat transfer coefficient indicating partial dryout at this location. For moderate and high heat flux values, the effect of heat flux became very complex and partial dryout is observed at the last thermocouple for all heat flux values. Figure 8 depicts the behaviour of the heat transfer coefficient at high heat flux values. It is obvious that the axial distance and the local quality plots show insignificant heat flux effect at very low quality values, x < 0.1. For higher quality

(b) Figure 7: Effect of heat flux on the local heat transfer coefficient at low heat flux for g = 200 kg/m2s, (a) h-x plot, (b) h-z plot

Research Article - JTEN - 2015 - 69 values, (see Fig. 10), there is a clear mass flux effect in the very low quality region. However, the axial distance plot indicates that there is a clear mass flux effect over the second half of the channel (from the middle to the exit) where the heat transfer coefficient increased with increasing mass flux.

(b) Figure 8: Effect of heat flux on the local heat transfer coefficient at high heat flux for g = 200 kg/m2s, (a) h-x plot, (b) h-z plot

(b) Figure 10: Effect of mass flux on the local heat transfer coefficient at low heat flux 697 kw/m2, (a) h-x plot, (b) h-z plot

The heat transfer results presented in this section indicate very complex behaviour. This complexity makes it difficult to infer the dominant heat transfer mechanism(s) in this study. Some researchers [15, 28, 35-37] conducted flow boiling studies using water in microchannels and agreed on the trend of the local heat transfer coefficient versus local vapour quality, which was different compared to the present study. They agreed on that the heat transfer coefficient decreases rapidly from a peak value at x = 0 then decreases moderately with increasing vapour quality. However, there were contradictions in the mass and heat flux effects. Sobierska et al. [15] reported clear heat flux effect for all mass flux range (100 – 700 kg/m 2s) while Kandlikar and Steinke [28] reported heat flux effect only

(b) Figure 9: Effect of mass flux on the local heat transfer coefficient at low heat flux 293 kw/m2, (a) h-x plot, (b) h-z plot

Figure 9 depicts the effect of mass flux on the local heat transfer coefficient plotted on the h-x and h-z plane at q = 293 kW/m 2. The figure demonstrates that the effect of mass flux on the local heat transfer coefficient at low heat flux is not clear, see Fig. 9 (a, b). At high heat flux 770

Research Article - JTEN - 2015 - 69 for the lowest mass flux in their study (157 kg/m 2s). For higher mass flux values, Kandlikar and Steinke [28] did not see any heat flux effect and attributed this to the rapid bubble growth and flow reversal. Sobierska et al. [15] reported heat transfer coefficient that decreases with increasing mass flux. On the contrary, Galvis and Culham [35] reported that the heat transfer coefficient decreases with heat flux up to a certain heat flux value after which it shows small dependence on both heat and mass flux. However, they reported nucleate boiling as a dominant heat transfer mechanism. Lim et al. [37] did not find any mass flux effect over the quality range up to x = 0.2 with the heat transfer coefficient decreases with quality for x <

Conclusions And Future Works

In this study, flow boiling experiments were performed in a single copper microchannel over a wide range of heat flux (56-865 kW/m2) and mass flux (200-800 kg/m 2s). The main conclusions drawn from this study are: -

z Greek ƞ ρ ΔP Subscripts app c ch cu e f FD fin i l lg local loss meas p sat sp sub tc tp outlet w 90

Flow visualization demonstrated that bubbles nucleate and rapidly grow without detaching from the nucleation cavity within the current experimental range. The rapid growth of the bubble before leaving the channel wall resulted in the formation of an elongated bubble, which blocks the channel temporarily (periodically). This periodic blockage results in the onset of flow boiling instability with large fluctuations in temperature and pressure. At boiling incipience, nucleation process including bubble growth into elongated bubbles occurred periodically. The period was detected from flow visualization through the sequence of pictures. The period was found to decrease with increasing mass flux, i.e. the elongated bubble can stay longer before leaving the channel at low mass flux. Thus there is a possibility of periodic partial dryout as mass flux decreases. There is a small heat flux effect only at low heat and mass flux values. At high heat and mass flux values, there was no heat flux effect. The heat transfer dominant mechanism is not clear in this study. A longer test section is needed in order to cover a wide range of vapour quality to help clarify the dominant heat transfer mechanism(s).

Nomenclature

Liquid specific heat Hydraulic diameter Single-phase fanning friction factor Mass flux Heat transfer coefficient Height of the microchannel Specific enthalpy Pressure loss coefficient Thermal conductivity Length Fin parameter Mass flow rate Nusselt number Power Heat energy Heat flux Average surface roughness Reynolds number Temperature Thickness, time Velocity of the liquid Width of the microchannel Thermodynamic mass vapour quality Axial distance Fin efficiency Density Pressure drop Apparent Sudden contraction channel Copper Sudden enlargement Fluid Fully developed Fin Inlet liquid Vapourization Local Loss measured Plenum Saturation Single-phase Subcooled Thermocouple Two-phase Outlet wall Normal to the flow direction

[J/kg°C] [m] [-] [kg/m 2s] [W/m 2K] [m] kJ/kg [-] [W/mK] [m] [-] [kg/s] [-] [W] [W] [W/m2] [µm] [°C] [m], [ms] [m/s] [m] [-] [m] [-] [kg/m3] [Pa]

Research Article - JTEN - 2015 - 69 microchannel, Experimental Thermal and Fluid Sciences, 31 (2006) 111-119. [16] A. Megahed, I. Hassan, Two-phase pressure drop and flow visualization of FC-72 in a silicon microchannel heat sink, International Journal of Heat and Fluid Flow, 30 (2009) 1171-1182. [17] T. Chen, S.V. Garimella, Measurements and highspeed visualizations of flow boiling of a dielectric fluid in a silicon microchannel heat sink, International Journal of Multiphase Flow, 32 (2006) 957-971. [18] K. Balasubramanian, M. Jagirdar, P.S. Lee, C.J. Teo, S.K. Chou, Experimental investigation of flow boiling heat transfer and instabilities in straight microchannels, International Journal of Heat and Mass Transfer, 66 (2013) 655-671. [19] W. Qu, I. Mudawar, Measurement and prediction of pressure drop in two phase micro-channel heat sinks, International Journal of Heat and Mass Transfer, 46 (2003) 2737-2753. [20] K.H. Chang, C. Pan, Two-phase flow instability for boiling in a microchannel heat sink, International Journal of Heat and Mass Transfer, 50 (2007) 2078-2088. [21] D. Bogojevic, K. Sefiane, A.J. Walton, H. Lin, G. Cummins, Two-phase flow instabilities in a silicon microchannels heat sink, International Journal of Heat and Fluid Flow, 30 (2009) 854-867. [22] G. Wang, P. Cheng, H. Wu, Unstable and stable flow boiling in parallel microchannels and in a single microchannel, International Journal of Heat and Mass Transfer, 50 (2007) 4297-4310. [23] S.G. Kandlikar, Nucleation characteristics and stability considerations during flow boiling in microchannels, Experimental Thermal and Fluid Sciences, 30 (2006) 441447. [24] H. Tuo, P. Hrnjak, Periodical reverse flow and boiling fluctuations in a microchannel evaporator of an airconditioning system, International Journal of Refrigeration, 36 (2013) 1263-1275. [25] S.L. Qi, P. Zhang, R.Z. Wang, L.X. Xu, Flow boiling of liquid nitrogen in microtubes: part I - the onset of nucleate boiling, two-phase flow instability and two-phase flow pressure drop, International Journal of Heat Mass Transfer, 50 (2007) 4999-5016. [26] J.G. Collier, J.R. Thome, Convective Boiling and Condensation, third ed., Oxford University Press, Oxford, 1994. [27] X.F. Peng, B.X. Wang, Forced convection and flow boiling heat transfer for liquid flowing through microchannels, International Journal of Heat and Transfer, 36 (1993) 3421-3427. [28] M.E. Steinke, S.G. Kandlikar, An experimental investigation of flow boiling characteristics of water in parallel microchannels, Journal of Heat Transfer, 126 (2004) 518-526. [29] J. Lee, I. Mudawar, Two-phase flow in high-heat-flux micro-channel heat sink for refrigeration cooling

Acknowledgments

The first author acknowledges the support provided by the Turkish Higher Research Council.

Share and Cite