Accepted Manuscript Pool boiling performance and bubble dynamics on microgrooved surfaces with reentrant cavities Yalong Sun, Gong Chen, Shiwei Zhang, Yong Tang, Jian Zeng, Wei Yuan PII: DOI: Reference:
S1359-4311(17)31809-4 http://dx.doi.org/10.1016/j.applthermaleng.2017.07.044 ATE 10711
To appear in:
Applied Thermal Engineering
Received Date: Revised Date: Accepted Date:
17 March 2017 21 June 2017 4 July 2017
Please cite this article as: Y. Sun, G. Chen, S. Zhang, Y. Tang, J. Zeng, W. Yuan, Pool boiling performance and bubble dynamics on microgrooved surfaces with reentrant cavities, Applied Thermal Engineering (2017), doi: http:// dx.doi.org/10.1016/j.applthermaleng.2017.07.044
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Pool boiling performance and bubble dynamics on microgrooved surfaces with reentrant cavities Yalong Sun, Gong Chen, Shiwei Zhang*, Yong Tang, Jian Zeng, Wei Yuan Key Laboratory of Surface Functional Structure Manufacturing of GuangDong Higher Education Institute, South China University of Technology, GuangZhou, 510640, China Corresponding author: Telephone/fax: +86 20 87114634; Email:
[email protected].
Abstract Microgrooved surfaces with reentrant cavities (MSRCs) were fabricated on pure copper substrates by employing the orthogonal Ploughing/Extrusion (P/E) method. The pool boiling heat transfer characteristics of this enhanced structure at different liquid subcooling degrees were investigated with deionized water as the working fluid at atmospheric pressure. The results indicated that the MSRCs had a higher heat transfer coefficient (HTC) than the smooth copper sheet (SCS) at any liquid subcooling. In addition, the wall superheat at the onset of nucleate boiling (ONB) for the MSRC was lower. Studies on the visualization of bubble growth process were also conducted with a high-speed digital camera, which showed that there were two kinds of bubbles on the MSRC surface depending on the different locations of nucleation sites. i.e., in the reentrant cavities and in micro-grooves. The departure diameter of bubbles that nucleated in the grooves increased with increasing heat flux for qa < 465.58 kW/m2 and then started decreasing sharply. However, heat flux had little influence on the departure diameter of bubbles growing in the reentrant cavities. Increasing the liquid subcooling temperature reduced the bubble departure diameter while increased the HTC.
Key words: pool boiling; heat transfer; bubble growth; micro groove; reentrant cavity 1
Nomenclature CHF
Critical heat flux, kW/m2
HTC (h)
Heat transfer coefficient, kW·m-2·K-1
IMN
Interconnected microchannel nets
MSRC
Microgrooved surface with reentrant cavities
ONB
Onset of nucleate boiling
P/E
Ploughing/Extrusion
SCS
Smooth copper sheet
WEDM
Wire electrical discharge machining
Dd
Bubble departure diameter, mm
Dpe, Ppe
P/E depth/pitch, mm
L1
Distance between thermocouple T2 and the lower surface of the solder layer, mm
L2
Solder layer thickness, mm
qa
Real heat flux supplied to test section, kW/m2
Tavg
Average temperature of T1, T2 and T3, °C
Ti
Thermocouple reading (i=1-6), °C
Tsat
Liquid saturation temperature, °C
Tw
Wall temperature, °C
Tsat
Wall superheat, Tw - Tsat, °C
Tsub
Degree of subcooling, °C
Subscript
i
Thermal conductivity, W/(m·K)
avg
Average
d
Departure
sub
Subcooled
sat
Saturated
w
Wall
2
1. Introduction The fast development of electronic devices towards miniaturization and lightweight increases the heat flux while reduces the heat dissipation space. Therefore, effective heat dissipation manners have become the key to ensure the safe operation of electronic equipment. Pool boiling utilizes the phase change process to transfer a large quantity of heat, and the superiority of heat dissipation performance of pool boiling has been demonstrated by many other researchers [1-3]. Therefore, it is widely recognized as an effective heat dissipation method, for a wide range of application areas, such as LEDs [4], CPUs [5], 3D chips [6] and hard disk drives [7], etc. In order to further enhance the heat transfer performance during pool boiling, a number of researches have been conducted in the past decades with concentration on producing enlarged heat dissipation surfaces. Since the works on the effect of surface finish on boiling heat transfer characteristics in 1931 by Jackob et al. [8], various microstructures have been developed as enhanced cooling surfaces, such as fins [9-11], deposited nanoparticles on the cooling surfaces [12-15], micro channels [16, 17] and porous-layer coatings [18]. Further investigations revealed that the pool boiling heat transfer could be further augmented by integrating the modification methods at the meantime, e.g., micro-grooves with porous coating on the surfaces [19, 20]. Tang et al. [21] fabricated the interconnected microchannel nets (IMN) on porous matrix by employing the wire electrical discharge machining (WEDM) method, and it showed a lower wall superheat and a higher HTC compared to the pure-copper IMN. Gheitaghy et al. [22] compared the pool boiling heat transfer characteristics of plain surface, mesochannel surface, electrodeposited microporous surface and mesochannel surface with porous-layer coating. The results showed that the mesochannel surface with porous-layer coating yielded a 2.1 and 3.9 times enhancement on critical heat flux (CHF) and HTC than the plain surface, respectively. In general, the research direction of the studies mentioned above and other similar ones mainly focused on the combination of the microchannels and the porous 3
materials or depositing nanoparticles on structured surface. Although these types of modification were proven to be efficient in pool boiling heat transfer enhancement, the complex manufacturing process and high production cost involved may prohibit them from industrial application on large scale [23, 24]. Therefore, in order to meet the requirements in industrial practices, a new type of high-performance microchannels should be developed which allows for facile and cost-effective techniques. Microchannels with high-aspect ratio, which are favorable for effective boiling heat transfer [25], while still maintaining a low cost and technical feasibility were developed by Tang et al. [26, 27] using the P/E method. During the ploughing process, the metal was forced upward, rather than peeled off from the base material, and thus high-aspect ratio grooves were produced with relatively low cutting depth. Besides, the rough surface for the P/E grooves and the favorable capillary performance of the P/E grooves has been justified in Ref [28]. Due to the superior heat and mass transfer characteristics of P/E grooves, it has also been successfully applied to the field of heat pipes [29], heat sinks [30] and heat columns [31]. More importantly, the P/E process could be conducted on general machine tools, such as planing tools, which is facile [32]. Inspired by the idea of enhanced bubble nucleation on the channel walls, the performance of P/E microchannels were further enhanced by introducing the cross-connected finned microgrooves in copper strips [33]. It was found that the P/E depth has the greatest influence on the formation of micro pores. Further studies conducted by Kuo et al. [34] revealed that the microchannels with reentrant cavities on the side walls yielded a higher flow boiling HTC when compared to plain-wall microchannels. The research conducted by Zhou et al. [35] indicated that the boiling mechanisms are strongly dependent on the wall surface conditions. The flow boiling heat transfer performance of the enhanced and the plain-wall microchannels were compared by simulations, which demonstrated that the reentrant cavities in the side walls of microchannels have the ability to facilitate nucleating and delay CHF. Most recently, Zeng et al. [36] fabricated the MSRC structure on aluminum plates using the orthogonal P/E method. The MSRCs featured high-aspect ratio grooves and reentrant cavities on the side walls. The capillary performance of 4
the MSRCs were characterized in the capillary rise and permeability tests, which justified that the reentrant cavities on the side walls can enhance the capillary force. In addition, the high-aspect ratio grooves for MSRCs are believed to provide high capillary pressure, which could facilitate the liquid replenishment at high heat flux, avoiding the formation of Leidenforst film [37] during pool boiling. Besides, the reentrant cavities could also increase the nucleation sites density, which could possibly enhance the bubble nucleation on the initial phase of boiling. Therefore, in order to further prove the promising application of the MSRCs as enhanced wicks for two-phase heat transfer devices, the pool boiling heat transfer performance of the MSRCs are investigated in this paper, which is the main objective in the present study. In this paper, MSRCs were produced on the pure copper substrates by employing the orthogonal P/E method. Pool boiling tests were conducted and compared to the SCS using deionized water as the working fluid in different subcooling degrees at atmospheric pressure. The ONB, HTC and the visualization of bubble growth characteristics were evaluated to explore the feasibility of the heat transfer enhancement and application in high heat flux region.
2. Manufacturing of MSRC The MSRCs in this study were fabricated by orthogonal P/E method on pure copper substrates. One may refer to Ref. [36] for the detailed manufacturing process. Fig. 1 shows the characterizations of the MSRC-1 sample, the overall dimensions of the tested samples are 20 mm × 20 mm × 2 mm, as shown in Fig. 1.a. The manufacturing operation of MSRC consists of two ploughing processes. As shown in Fig. 1.c, after the first ploughing process, with the parameters of ploughing depth Dpe = 0.3 mm, ploughing pitch Ppe = 0.75 mm, micro-grooves with fins were obtained on the copper substrate, which has a rough surface due to the friction between the extrusion surface and the forming surface. Unlike traditional cutting which cuts the metal out to become chips, the residual metal formed during ploughing process still maintained on the substrate and translated into the fins on the top of 5
micro-grooves, which increased the aspect ratio. Then the copper sheet was placed at 90° to the original position, and the second ploughing process was conducted with
Pitch (Ppe) 1.2 mm
20 mm
Depth (Dpe) 0.5 mm
two ploughing depths
Sectional view of the reentrant cavities 20 mm (a)
(b)
First P/E direction
Second P/E direction
Fins on the top of the groove
Reentrant cavities
(c)
(d)
Fig. 1 Characterization of MSRC-1.
of 0.3 mm and 0.5 mm while the same ploughing pitch of 1.2 mm. The second ploughing process broke down the first groove array periodically with the pitch of 1.2 mm. The residual metal was pushed to both sides of the cutting tool, which partially covered the segmented grooves. Therefore, the reentrant cavities were formed on the side walls of the second micro grooves. Fig. 1.b and Fig. 1.d show the sectional view and the enlarged view of the reentrant cavities, respectively. Apparently, the second array of MSRC served as the main channels for the fluid flow. After the two ploughing processes, the WEDM was used to obtain the accurate dimensions of the samples. In this study, two samples with the second ploughing depths of 0.5 and 0.3 mm 6
are referred as MSRC-1 and MSRC-2, respectively. For better comparison, a SCS was also made with the same dimensions, which was polished with 1200 mesh sandpaper before the test. Details of the structure parameters of the tested samples are shown in Table 1.
Table 1. Structural parameters for the samples tested in this study. Sample
Dimensions/mm
Matrix
P/E parameters depth (Dpe)×pitch (Ppe) mm
Material
First array
Second array
MSRC-1
20×20×2
Pure copper
0.3×0.75
0.5×1.2
MSRC-2
20×20×2
Pure copper
0.3×0.75
0.3×1.2
SCS
20×20×2
Pure copper
-
-
3. Experimental setup and data reduction 3.1 Experimental setup Fig. 2 and Fig. 3 show the picture and schematic of the test platform, respectively. The test platform consists of a heating module, data acquisition part, visualization system, temperature control system and a boiling chamber. The saturated and subcooled deionized water were used as working fluid in the pool boiling experiments, and the degree of subcooling was 0, 10 and 20 °C .
7
Fig. 2. Photograph of the test platform. Inlet Outlet
HIOKI
HIOKI
Sample
LR8501
L2 0.2mm
Data acquisition part
T1
Solder
T2 7.0mm
T3 L1 2.5mm 7.0mm
6mm
T4 6mm
8888
T5
Temperature controller
T6
6mm
Visualization system
A 8888
B
Copper block
(b)
Φ6.0mm
Powermeter Variac (a)
Catridge heaters
(c)
Fig. 3. Schematic of the test platform (a), details of the arrangement of thermocouples (b) and cartridge heaters (c).
The boiling chamber was made of Pyrex glass for high-speed visualization study. Teflon plates were assembled at both ends of the Pyrex glass tube, and rubber O-rings were placed at the interface of glass tube (with the outer diameter of 70 mm and length of 150 mm) and Teflon plates to avoid liquid leakage. It should be noted there was an open hole on the upper Teflon plate which ensured that the tests were performed under atmospheric pressure. Before each experimental test, the deionized water was fully degassed by vigorous boiling for about half an hour. The temperature of subcooled water was maintained by the temperature control system, which consists 8
of a thermocouple, an auxiliary heater of 150 W and a copper-coil condenser. The tip of thermocouple was placed at 10 mm above the tested surface, acting as a monitor of the temperature of water. When the measured temperature was lower than the setting one, the auxiliary heater began to work; otherwise, the auxiliary heater stopped heating while the condenser took away the excess heat to lower the water temperature. During the whole testing process, the liquid temperature was maintained within ±1
°C of the desired temperature. As shown in Fig. 3.b, the heating module was made by a pure copper block, which was embedded in a Teflon block. The Teflon block has a low thermal conductivity to reduce the heat loss applied to the sample. Eight cartridge heaters (6 mm in diameter and 40 mm in length) of 120 W capacity each were symmetrically distributed in the copper block, as shown in Fig. 3.c. The cartridge heaters were powered by a variac and monitored by a power meter. The samples were welded on the upper surface (20×20 mm) of the copper block, and the high thermal conductivity solder (Pb-Sn-Ag-Sb) was used, the melting point of the solder paste is 150 °C ,so the experiments were all conducted below 150 °C to avoid the solder from melting. It should be noted that when assembled, the upper surface of the copper block was designed to be 0.2 mm below the boiling chamber. Therefore, the amount of solder used was ensured to be the same for different test samples as the gap was exactly filled during the welding process. Thus, the thermal resistance brought by the solder-layer was also the same for every test. The data acquisition part consisted of thermocouples and HIOKI data acquisition system (LR8410-30 wireless logging station and LR8510 wireless voltage/temp unit). Three K-type thermocouples (T1, T2, T3) were uniformly placed under 2.5 mm of the testing surface, which would be used to calculate the wall-temperature. In the vertical direction, four K-type thermocouples (T2, T4, T5, T6) were placed with the interval of 6 mm to calculate the actual heat flux according to Fourier Law. The details of the arrangement of thermocouples can be seen in Fig. 3.b. The visual observations were carried out by the high-speed camera (Fastec TS5, American Fastec Corporation, with 9
960 × 540 pixels @ 922 fps), with a high power lamp serving as the lighting source.
3.2 Data reduction During each test, the power applied to the copper block was increased at an interval of 15-30 W by adjusting the variac. When the temperature fluctuations of the thermocouples were less than 0.5 °C in 2 min, measurements were taken at 1 s interval for 2 min. The final results were the average of the collected data. In this study, the input heat flux (qa) to the test samples were calculated by the one-dimensional Fourier Law. qa c
dT dz
(1)
Where c is the thermal conductivity of the pure copper and is 397 W/(m·K), dT dz
is the vertical temperature gradient in the copper block determined by the four
vertical thermocouples readings, z is the coordinate perpendicular to the base surface, T is the reading of thermocouples. The wall temperature (Tw) of the test surface was given as follows, T1 T 2 T 3 3 L1 L 2 Tw Tavg qa ( ) Tavg
c
(2) (3)
t
Where Tavg is the average temperature of T1, T2 and T3, t is the thermal conductivity of the solder and is 50 W/(m·K); L1 is the distance between the thermocouple T2 and the lower surface of the solder layer; and L2 is the thickness of the solder layer, and qa is the input heat flux from Eq. (1). The HTC (h) was determined as follows h
qa Tsat
(4)
Where Tsat is the wall superheat calculated by
Tsat Tw Tsat
(5)
Where Tsat is the liquid saturation temperature; Tw is the wall temperature from Eq. (3).
10
3.3 Uncertainty analysis The calculations of uncertainties were based on a standard error analysis method [38]. The temperature measurement uncertainty of the precise K-type thermocouple was ±1 °C . The uncertainty of the subcooling degree was ±1 °C , the limitation of WEDM accuracy and the solder layer thickness was estimated as ±0.01 mm, and the location uncertainties of thermocouples were estimated to be ±0.05 mm. Thus the calculated uncertainties for the applied heat flux and HTC are within 2.34% and 6.33%, respectively.
4 Results and discussion 4.1 Boiling curves 1400
1400
1300 1100 -2
1000 900
CHF
1100
800 700 600 500 400 ONB
300
1000 900 700 600 500 400
100
100 0
5
10
15
20
25
30
35
40
45
Wall-superheat ΔTsat/ºC
ONB
300 200
-5
CHF
800
200 0 -10
MSRC-2-20 ºC MSRC-2-10 ºC MSRC-2-0 ºC SCS-20 ºC SCS-10 ºC SCS-0 ºC
1200
Heat Flux/kW ·m-2
1200
Heat Flux/kW·m
1300
MSRC-1-20 ºC MSRC-1-10 ºC MSRC-1-0 ºC SCS-20 ºC SCS-10 ºC SCS-0 ºC
0 -10
-5
0
5
10
15
20
25
30
35
40
45
Wall-superheat ΔTsat/ºC
(a)
(b)
Fig. 4. Boiling curves for the MSRCs and SCS at different liquid subcooling degrees: (a) MSRC-1 and SCS; (b) MSRC-2 and SCS.
Fig. 4 shows the boiling curves of MSRCs and SCS at different liquid subcooling temperatures. All of the tests have not reached the CHF except the SCS at the subcooling degree of 0 °C , as shown in Fig. 4. The ONB is defined as the turning point of the boiling curves. As shown in Fig.4, the MSRCs induce the ONB at a lower wall superheat. For example, when the subcooling degree is 20 °C , the wall superheat at ONB are 1.45 °C and 4.70 °C for MSRC-1 and SCS, respectively. It should be noted that there are no obvious shift for MSRCs under the subcooling of 0
°C and 10 °C , due to the fact that the tested minimum heat flux has already triggered the boiling incipience. As shown in Fig. 6.a, discrete bubbles are clearly observed on the surface of MSRC-1, which means the existence of the nucleate boiling. After the 11
ONB, despite of a small nonlinear increase region, the boiling curves are almost linear, which suggests that nucleate boiling regime governs in the whole testing process. The MSRCs are able to induce the ONB at lower wall superheat mainly due to the confinement of bubble nucleus and high local heat flux in the reentrant cavities. As is justified in various studies [19, 34], the reentrant cavities, thanks to the semi-closed structure, serve as vapor traps, which confine the small nucleus inside it at the single phase. The nucleus acts as the nucleation site. On the contrary, for the SCS sample, since bubbles are completely surrounded by the liquid, they have been condensed quickly before they are able to grow up, indicating that the SCS needs a higher heat flux to maintain the growth of bubbles. Furthermore, bubbles nucleating inside the reentrant cavities absorb heat from all sides of the cavity, while those on the planar heated surface only get heat from vertical direction. As a result, bubbles inside the cavities have a higher local heat flux. Therefore, the MSRCs accelerate the bubble nucleation at a much lower wall superheat as compared to the SCS sample. As is shown in Fig. 4.a and b, both the MSRC-1 and MSRC-2 yield higher heat flux than the SCS at any given liquid subcooling and wall superheat. For example, at
Tsat = 20 °C and Tsub = 0 °C , the MSRC-1 and MSRC-2 attain the heat flux of 552.68 and 541.75 kW/m2 , respectively, which is about 18.99% higher than that for the SCS (qa = 459.87 kW/m2). In addition, we find that the MSRCs are competitive compared with other heat sinks in literatures. For example, Zhang et al. [20] developed a porous interconnected microchannel net with copper power sintering method, which attained the heat flux of 543.1 kW/m2 under the same condition mentioned above. Similar heat flux is attained by a micro-structured surfaces with regular patterns developed by Moita et al. [39], which is 510 kW/m2 at Tsat = 20
°C and Tsub = 0 °C . It should be noted that either sintering or other MEMS procedures have to be employed in the aforementioned literatures [20, 39], which are expensive and inefficient. Therefore, the MSRC provides a direction of development for high-performance heat sinks with a cost-effective and stable method. The enhancement of boiling heat transfer for the MSRC is possibly due to the enhanced nucleation in the reentrant cavities and vapor path in the micro-grooves. As mentioned 12
above, the reentrant cavities trap the nucleus and provide higher local heat flux, facilitating the nucleate boiling. As a result, the MSRCs are able to dissipate a greater amount of heat in unit time. In addition, the micro-grooves serve as vapor path for easier bubble escape. More importantly, as the bubbles depart from the cavities, micro-convection is induced in the grooves, similar to that observed in microchannels with porous layer on the top of the fins [11]. Besides, it should be noted that the superiority of the MSRCs to the SCS still maintained obvious at the high heat flux. For example, at Tsat = 35 °C and Tsub = 0 °C , the MSRC -1 and MSRC -2 attain the heat flux of 1228.98 and 1167.05 kW/m2, respectively, which is about 16.40% and 13.38% higher than that for the SCS (qa= 1029.3 kW/m2), respectively. The copper foam with crossing V-shaped grooves fabricated by Qu et al. [40] yields the heat flux of 1289.6 kW/m2, which is similar to that of the MSRCs under the same conditions. Similar result was also observed by Deng et al. [41] on the porous structures with reentrant cavities. Therefore, it could be seen that the MSRC remains competitive even in the high heat flux range. It is hypothesized that the superior heat transfer performance for the MSRCs at high heat flux results from the capillary force provided by the micro-grooves and reentrant cavities, which has been justified in our previous work [36]. It should be noted that at high heat flux, bubbles coalesce into vapor slugs on the top of the heated surface, which avoids liquid replenishment and leads to local dryout. However, the MSRCs are able to keep themselves wetted even at high heat flux thanks to the capillary liquid suction. Therefore, the MSRCs yield better heat transfer performance than the SCS at high heat flux. Higher heat flux can be attained by increasing the liquid subcooling for both the MSRC-1 and MSRC-2. For example, at Tsub = 20 °C , the MSRC-1 is able to dissipate 683.17 kW/m2 at Tsat = 20 °C , which is about 23.61% higher than that at Tsub = 0 K (qa= 552.67 kW/m2) on the same condition. This finding complies with what is observed in various literatures [21, 42-43]. This is possibly results from the reduced departure diameter at higher liquid subcooling, preventing the formation of large vapor slugs on the heated surface, which would be discussed in the later section. 13
4.2 Nucleate boiling heat transfer coefficient Fig. 5 shows the relationship between HTC and heat flux at different liquid subcooling. A general increasing trend of HTC is observed for both MSRCs and SCS, including the rapid growth region at low to medium heat fluxes (qa < 350 kW/m2) followed by the stable growth region afterwards. Within the tested heat flux region, the CHF for all the three samples are not reached except the SCS at the subcooling degree of 0 °C , the maximum HTC yielded by the MSRC-1 is 38.84, 36.59, 35.22 kW·m2·K-1 at the subcooling degree of 20 °C , 10 °C and 0 °C , respectively, which are about 15.80%, 12.93% and 14.50% higher than those of the SCS, respectively, and that by the MSRC-2 is 40.25, 36.74, 34.48 kW·m2·K-1 under the same conditio n.
40
45
35
40 35 -1
HTC/kW·m ·K
25
30
-2
-2
HTC/kW·m ·K
-1
30
20
MSRC-1-20 ºC MSRC-1-10 ºC MSRC-1-0 ºC SCS-20 ºC SCS-10 ºC SCS-0 ºC
15 10
25 20
MSRC-2-20 ºC MSRC-2-10 ºC MSRC-2-0 ºC SCS-20 ºC SCS-10 ºC SCS-0 ºC
15 10
5
5 0
100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400
0
100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 -2
Heat Flux/kW·m
-2
Heat Flux/kW·m
(a)
(b)
Fig. 5. HTC and heat flux relationship at different degrees of liquid subcooling: (a) MSRC-1 and SCS; (b) MSRC-2 and SCS.
The MSRCs yield higher HTC than the SCS at any given liquid subcooling and heat flux. For example, at subcooling degree of Tsub = 10 °C and heat flux qa = 745.76 kW/m2, the HTC for the MSRC-1 is 32.02 kW·m2·K-1, which is 12.31% higher than that for the SCS. Such an enhancement is mainly attributed to the micro-grooves and reentrant cavities. As has been discussed above, the combination of micro-grooves and reentrant cavities increases the capillary force, ensuring the sufficient liquid replenishment even at the high heat flux. Besides, the reentrant cavities also increase the bubble nucleation density and surface area, enlarging the 14
vapor-liquid heat transfer interface area. As a result, the total amount of heat dissipated by the MSRCs is much higher than that for the SCS. Besides, the heat conduction and phase change at the gas-liquid interface, the micro-convection induced in the grooves as the bubble escape from the cavities further enhance the HTC, and more importantly, the turbulence interrupts the bubble coalescence, avoiding the formation of large vapor slugs. In the low to moderate heat flux region, HTC can be significantly enhanced by increasing the liquid subcooling. For example, the HTC for MSRC-1 at qa = 109.94 kW/m2 and Tsub = 10 °C is 19.46 kW·m2·K-1, which is 42.93% higher than that at
Tsub = 0 °C . The same trend is also found for the MSRC-2. Such an enhancement has also been observed by many other researchers [16, 21, 42]. The large enhancement of HTC can be mainly attributed to the fact that the increased subcooling degree restrains the bubble growth, which ensures a sufficient amount of liquid flow into the heating surface. However, at the moderate to very high heat flux region (qa = 350-1400 kW/m2), the liquid subcooling degree only has a minor effect on the HTC of SCS. This phenomenon may be explained by the fact that the SCS has a much worse intrinsic capabilities of bubble nucleation and liquid-replenishment than the MSRCs, which limits the effects of liquid subcooling on the HTC enhancement. On the contrary, the HTC of MSRC samples are still able to increase at high heat flux region when the liquid subcooling degree increases from 10 °C to 20 °C ; and the HTC curves also keep an increasing tendency with the increase of heat flux, which indicates that the MSRC do not reach the boiling limit in the tested range of heat flux. The HTC of MSRC-1 is slightly enhanced when the liquid subcooling degree changes from 0 °C to 10 °C . Besides, the HTC curve maintains a rising tendency even at the ultrahigh heat flux region (i.e., qa > 1000 kW/m2 ) when the liquid subcooling is 10
°C and 0 °C . In conclusion, the MSRCs yield higher HTC than the SCS at any liquid subcooling. The HTC for the MSRCs can be further enhanced by increasing the liquid subcooling, which is especially obvious at the low to medium heat flux region. The MSRCs are able to induce the ONB at lower wall-superheat than the SCS, mainly due 15
to the increased nucleation and higher local heat flux provided by the reentrant cavities. The superior heat transfer performance of the MSRCs is jointly attributed to the enlarged nucleate density, sufficient liquid supply as well as micro-convection induced by departing bubbles.
4.3 Visualization research 4.3.1 Bubble growth characteristics In this study, bubble growth process was observed with high speed camera with 922 frames/s for both MSRC-1 and SCS samples. The high-speed images were taken from the side direction, so the bubble growth process can be clearly recorded. Fig. 6 and Fig. 7 show the differences of bubble growth process with the increase of heat flux at the subcooling of 0 °C .
(a) 35.51 kW/m2
(b) 72.98 kW/m2
(c) 146.42 kW/m2
(d) 270.15 kW/m2
(e) 465.58 kW/m2
(f) 669.32 kW/m2
(g) 891.95 kW/m2
(h) 1195.95 kW/m2
Fig. 6. Bubble visualizations of MSRC-1 ( Tsub = 0 °C ).
(a) 36.64 kW/m2
(b) 73.21 kW/m2
(c) 146.18 kW/m2
(d) 270.70 kW/m2
(e) 469.57 kW/m2
(f) 670.10 kW/m2
(g) 891.93 kW/m2
(h) 1196.09 kW/m2
Fig. 7. Bubble visualizations of SCS ( Tsub = 0 °C ).
With low heat flux applied, discrete bubbles are observed both on the MSRC-1 and SCS sample, as illustrated in Fig. 6.a and Fig. 7.a. However, it is clear that the bubble density is much larger for the MSRC-1 than that for the SCS, mainly due to the fact that the bubbles nucleating in the reentrant cavities are isolated from the bulk 16
liquid, which prevents the subcooled liquid to condensate the vapor embryos. Besides, larger local heat flux can be obtained inside the cavity because the bubble is heated by the cavity walls surrounding it, further facilitating the bubble nucleation. On the contrary, a much higher wall-superheat is required to induce the bubble growth in SCS surface due to the quick condensation of bubble nucleus by the cool bulk liquid. As the heat flux increases, more nucleation sites are activated for the MSRC-1, as shown in Fig. 6.b. The stable and successive bubble departure dissipates a great amount of heat, which leads to the rapid growth of HTC as shown in Fig. 5. However, only a small number of bubble nucleation sites are activated for the SCS in Fig. 7.b, and the bubbles are intermittently formed at activated sites, indicating the weaker heat transfer performance. The more active bubble nucleation analyzed above possibly explains the enhancement of heat transfer by the MSRC-1 at the low heat flux. As heat flux increases to qa = 146.42 to 270.15 kW/m2 as shown in Fig. 6.c-d, the bubbles of MSRC-1 mainly depart from the cavities on the top despite the coalescence into vapor slugs after departure, with fewer bubbles nucleating in the microgrooves. Therefore, sufficient liquid replenishment is ensured for MSRC-1 due to the capillary suction by the microgrooves. Nevertheless, for the SCS, with more bubble nucleation sites are activated, the bubbles tend to merge before detaching from the surface, and vertical coalescences are observed, leading to the formation of vapor slugs, as shown in Fig. 7c-d. It should be noted that the vapor slugs isolate the heated surface from the subcooled water. As a result, the liquid replenishment of SCS is undermined, which reduces the HTC for the SCS, as shown in Fig. 5. With the further increase of heat flux from 465.58 to 669.32 kW/m2 as shown in Fig. 6.e-f, almost all bubbles are coalesced with their predecessors and vapor columns are formed on the top of reentrant cavities. However, these vapor columns are very unstable when they move upside, the break of vapor columns induced a disturbance in the pool, which increased the microconvection heat transfer [44]. For the SCS, as shown in Fig. 7.e-f, bubbles coalesce into large vapor slugs before they depart. It can be seen that the vapor slug is relatively stable, characterized by the regular shape even after departure, which isolates the bulk liquid and the heating surface, leading to the 17
increase of wall superheat. At high heat fluxes, as shown in Fig. 6.g-h, bubbles almost cover the heating surface and large mushroom clouds are formed, which further isolate the heating surface from the subcooled water and lead to a surge of wall temperature. On the contrary, the MSRC structures made by P/E method have a rough surface along with micro fins on the finished surface and reentrant cavities in the sidewalls, resulting in an obvious increase in the capillary force [28]. This speeds up the liquid flow rate towards the heating surface, ensuring that there is enough liquid to keep the heating surface wet. Furthermore, micro-grooves on the MSRCs can ensure the liquid supply in the horizontal direction, and thus the cycles of bubble nucleation, growth and expulsions can be maintained. Without structures on the heating surface, the continual bubble merger prevents the liquid flow towards the heating surface, as shown in Fig. 7.g-h.
4.3.2 Bubble departure diameter Further study on the bubble growth mechanism for the MSRC-1 reveals that there exists two types of bubbles in terms of the site it nucleates, i.e., groove bubbles and cavity bubbles. Fig. 8 and 9 show the variation of bubble departure diameters (Dd) with increase of heat fluxes at the subcooling degree of 0 °C , with the upper to lower row representing the groove bubbles and cavity bubbles for the MSRC-1, and bubbles on the SCS. Besides, the mean departure diameter in Fig. 9 is calculated by considering at least 10 bubbles to ensure the accuracy.
18
A: Bubbles nucleated in the micro-grooves 1mm
1mm
1mm
60-1
90-1
(a) 146.42 kW/m2 (b) 219.58 kW/m2 (c) 270.15 kW/m2 Dd =1.79 mm Dd =2.03 mm Dd =1.92 mm
1mm
1mm
1mm
110-1
(d) 319.88 kW/m2 Dd =2.26 mm
190-1
(e) 465.58 kW/m2 Dd =2.56 mm
240-1
(f) 594.35 kW/m2 Dd =1.89 mm
B: Bubbles nucleated in the reentrant cavities 1mm
1mm
60-2
(a) 146.42 kW/m2 Dd =1.08 mm
90-2
1mm
1mm
1mm
110-2
2 (b) 219.58 kW/m2 (c) 270.15 kW/m Dd =1.14 mm Dd =1.16 mm
(d) 319.88 kW/m2 Dd =1.23 mm
190-2
(e) 465.58 kW/m2 Dd =1.16 mm
1mm
(f) 594.35 kW/m2 Dd =1.11 mm
C: Bubbles nucleated on the smooth copper sheet
1mm 60
2
(a) 146.18 kW/m Dd =1.29 mm
1mm
1mm 90
(b) 220.51 kW/m2 Dd =1.37 mm
110
2
(c) 270.7 kW/m Dd =1.75 mm
1mm
1mm 2
(d) 320.57 kW/m Dd =1.97 mm
130
2
(e) 371.8 kW/m Dd =2.01 mm
1mm 240
(f) 595.22 kW/m2 Dd =2.75 mm
Fig. 8. High-speed images of the departing bubble at micro-grooves of MSRC-1 (upper row) and reentrant cavities of MSRC-1 (middle row) and the SCS (bottom row) at different heat fluxes.
By comparing Fig. 8.A and B, one can clearly identify that the bubbles departing from the cavities are much smaller than those from the groove, which justifies our assumption of bubble nucleation enhancement described in previous section. For example, at qa = 465.58 kW/m2, Dd for the cavity bubble is 1.16 mm, which is 54.67% smaller than that from the groove. The reduced departure diameter for bubbles from the cavity is attributed to the reentrant characteristic, which traps the small nucleus and isolates the small bubbles from the cool bulk liquid. Moreover, the bubble nucleation is further enhanced by the higher local heat flux as a result of the heating by all the walls surrounding the bubble. In addition, due to the strong capillary suction of the reentrant cavities, the interface of the liquid and bubble is unstable, which induces a more chaotic flow in the micro-layer under the bubble. Therefore, the bubble detaches from the cavity more easily, leading to a smaller departure diameter of the MSRC-1.
19
Bubble departure diameter Dd /mm
3.2 2.8
MSRC-1-0 ºC-groove bubble MSRC-1-0 ºC-cavity bubble SCS-0 ºC
2.4 2.0 1.6 1.2 0.8
150 200 250 300 350 400 450 500 550 600 -2
Heat Flux/kW•m
Fig. 9. Average bubble departure diameter as a function of heat flux.
There is a distinct difference of departure diameter as a function of heat flux between the groove and cavity bubble. As shown in Fig. 9, the diameter for the groove bubble increases from 1.79 to 2.56 mm as the heat flux rises from 146.42 to 465.58 kW/m2 but drops sharply to 1.89 mm at qa = 594.35 kW/m2. This trend complies well with the model proposed by Kolev [45], which is mainly caused by the predominance of the dynamic forces (e.g., the inertia force, the evaporation force) at the high heat flux region. However, the diameter of cavity bubble remains constant with the increase of heat flux except for the minor fluctuation. A similar phenomenon has been also observed by Chien and Webb [46]. This can be mainly attributed to the restriction of diameter of the reentrant cavities. It is hypothesized that when the diameter of bubble grows to the size of the cavity, the bubble protrudes to the surface driven by the vapor pressure and then departs driven by the buoyance force. Since the bubble remains in contact with the reentrant cavity before it departs as shown in Fig. 8, the departure diameter of bubbles is almost same as the cavity diameter. Therefore, there is no obvious change of bubble diameter for the cavity bubbles as the heat flux changes. Differently, Dd for the SCS keeps growing along with the increases of heat flux and is larger than that of groove bubbles when qa = 595.22 kW/m2, as shown in Fig. 8.C, a similar phenomenon has been also observed by Mchale and Garimella [47]. 20
Furthermore, the Dd of SCS is larger than that of bubbles nucleated in the reentrant cavities at each heat flux. This phenomenon indicates that the mechanism which controls the bubble departure diameter of SCS is different from that of cavity bubbles. Bubbles on the SCS firstly grow inside the microscale cavities, when the diameter of bubble equals to the diameter of cavity, the buoyancy cannot carry the bubble away, thus the bubble will extend out of the cavity and form a liquid microlayer [48]. And the interface of the microlayer moves along the SCS to regain fresh liquid as the bubble grows along the surface, thus the Dd of SCS is larger than that of cavity bubbles. The above analysis shows that the values of Dd for the cavity bubbles are smaller than those for the groove bubbles and keep almost unchanged with the increase of heat flux. And Dd for the groove bubble increases when qa < 465.58 kW/m2 but decrease sharply when qa > 465.58 kW/m2, which is similar with the other studies.
4.3.3 Effects of subcooling As discussed above, the liquid subcooling degree has a significant influence on the heat transfer performance. For example, the HTC for MSRC-1 at qa = 271.08 kW/m2 and Tsub = 20 °C is 28.1 kW·m2·K-1, which is 14.06% higher than that at
Tsub = 10 °C . And such an enhancement is closely linked with the bubble growth characteristics. Therefore, the effects of liquid subcooling on the bubble departure diameter and formation of vapor slugs are investigated in this section. Fig.10 shows the bubble images at a similar heat flux with variation in the liquid subcoolings, and the departure diameters are plotted as a function of liquid subcooling in Fig.11. It can be seen that the boiling process is much more vigorous at lower Tsub . For example, at Tsub = 0 °C and qa = 669.32 kW/m2, large vapor slugs cover the MSRC-1 surface, while at Tsub = 20 °C and qa = 670.71 kW/m2, discrete bubbles can still be distinguished. The large vapor slugs covering the heated surface results in the deteriorated heat transfer enhancement on low liquid subcooling condition as is discussed for Fig. 5. It can be formed more easily due to the coalescences of larger bubbles that grow at lower subcooling degree, which is clearly justified in Fig. 11. 21
The departure diameters of cavity bubbles almost remain the same value, but the groove bubbles increase with decrease of liquid subcooling, e.g., the Dd is 1.07 mm at
Tsub = 20 °C but increases to 1.45 mm at Tsub = 0 °C . 1mm
1mm
(a) Δ Tsub= 0 ºC
(b) ΔTsub=10 ºC
1mm
(c) ΔTsub= 20 ºC
Fig. 10. Effects of subcooling on the bubble behaviors of MSRC-1 sample at high heat flux: (a) qa = 669.32 kW/m2; (b) qa = 670.73 kW/m2; (c) qa = 670.71 kW/m2.
Bubble departure diameter Dd /mm
1.8 groove bubble cavity bubble
1.5 1.2
0.9 0.6 0.3 0.0 0
10
20
Degree of subcooling/ºC Fig. 11. Average bubble departure diameter with variation of liquid subcooling degrees.
5. Conclusion This study investigated the boiling heat transfer characteristics of MSRCs as a new type of high-performance heat transfer device which is facile to manufacture to meet the large scale demand in industrial practices. The comparison study is performed under different subcooling degrees with deionized water as the working fluid at atmospheric pressure. The results show that the MSRCs exhibit a lower wall 22
superheat at the ONB and a higher HTC than that of the SCS. The heat transfer performance of the MSRCs are shown to be significantly affected by the liquid subcooling. The visualization study is then conducted with a high-speed camera to characterize the boiling phenomena of the MSRC-1 and the SCS. The key conclusions drawn from this work include: (1). The MSRCs suppress the temperature excursion by lowing the wall superheat at ONB and does not reach CHF during the entire test range. Besides, the MSRCs yield a maximum HTC of 38.84 kW·m2·K-1 at Tsub = 20 °C , which is about 15.80% higher than that of the SCS. (2). The liquid subcooling is shown to have a significant influence on the HTC, which is more obvious at low to medium heat flux region. Lowering the liquid subcooling decreases the HTC. (3). Bubbles on the surface of MSRCs can be classified into two types in terms of the nucleation sites, i.e., cavity bubble and groove bubble. Cavity bubbles are smaller than those on SCS and keep almost unchanged with the increase of heat flux due to the restriction of cavity diameter. Groove bubbles increase from 1.79 to 2.56 mm as the heat flux increases when qa < 465.58 kW/m2 but decreases sharply when qa > 465.58 kW/m2. (4). The bubble departure diameter of the groove bubble decreases with the increases of liquid subcooling. However, liquid subcooling is shown to have little influence on that of the cavity bubble.
Acknowledgements Supported by National Nature Science Foundation of China (No. 51475172), Team Program (No. 2014A030312017) supported by the Natural Science Foundation of Guangdong, Science and Technology Planning Project for Industry-University-Research Cooperation in Guangdong Province (No.2014B090901065 and 2016B090918096) and key program (No. U1401249) of NSFC-Guangdong Joint Funds of China. One of the authors, Shiwei Zhang, would like to acknowledge financial support from the Chinese Scholarship Council (CSC).
23
References [1] Y. Tian, K. Zhang, N. Wang, Z. Cui, L. Cheng, Numerical study of pool boiling heat transfer in a large-scale confined space, Appl. Therm. Eng. 118 (2017) 188-198. [2] H. Gong, A. R. Khan, N. Erkan, L. Wang, K. Okamoto, Critical heat flux enhancement in downward-facing pool boiling with radiation induced surface activation effect, Int. J. Heat Mass Transfer 109 (2017) 93-102. [3] A. Sathyabhama, A. Dinesh, Augmentation of heat transfer coefficient in pool boiling using compound enhancement techniques, Appl. Therm. Eng. 119 (2017) 176-188. [4] G. Setoh, F.L. Tan, S.C. Fok, Experimental studies on the use of a phase change material for cooling mobile phones, Int. Commun. Heat Mass Transfer 37 (9) (2010) 1403–1410. [6] C. Dang, L. Jia, Q. Lu, Investigation on thermal design of a rack with the pulsating heat pipe for cooling CPUs, Appl. Therm. Eng. 110 (2017) 390–398. [7] K.-l. Zhang, Z.-H. Liu, B.-c. Zheng, A new 3D chip cooling technology using micro-channels thermosyphon with super-moist fluids and nanofluid, Energy Convers. Manage. 128 (2016) 44–56. [8] M. Jakob, W. Fritz, Versuche über den verdampfungsvorgang, Forsch. Ingenieurwes. 2 (12) (1931) 435-447. [9] J. Fernández-Seara, Á.Á. Pardiñas, R. Diz, Heat transfer enhancement of ammonia pool boiling with an integral-fin tube, Int. J. Refrig 69 (2016) 175–185. [10] Y. Zhang, J. Wei, Y. Xue, X. Kong, J. Zhao, Bubble dynamics in nucleate pool boiling on micro-pin-finned surfaces in microgravity, Appl. Therm. Eng. 70 (1) (2014) 172-182. [11] A. Jaikumar, S.G. Kandlikar, Enhanced pool boiling for electronics cooling using porous fintops on open microchannels with FC-87, Appl. Therm. Eng. 91 (2015) 426-433. [12] S.Mt Aznam, S. Mori, F. Sakakibara, K. Okuyama, Effects of heater orientation on critical heat flux for nanoparticle-deposited surface with honeycomb porous plate attachment in saturated pool boiling of water, Int. J. Heat Mass Transfer 102 (2016) 1345–1355. [13] S. Das, D.S. Kumar, S. Bhaumik, Experimental study of nucleate pool boiling heat transfer of water on silicon oxide nanoparticle coated copper heating surface, Appl. Therm. Eng. 96 (2016) 555–567. [14] S. Das, B. Saha, S. Bhaumik, Experimental study of nucleate pool boiling heat transfer of 24
water by surface functionalization with SiO2 nanostructure, Exp.Therm. Fluid Sci. 81 (2017) 454-465. [15] S. Das, B. Saha, S. Bhaumik, Experimental study of nucleate pool boiling heat transfer of water by surface functionalization with crystalline TiO2 nanostructure, Appl. Therm. Eng. 113 (2017) 1345-1357. [16] J.C. Passos, R.F. Reinaldo, Analysis of pool boiling within smooth and grooved tubes, Exp. Therm Fluid Sci. 22 (2000) 35-44. [17] A.K. Das, P.K. Das, P. Saha, Performance of different structured surfaces in nucleate pool boiling, Appl. Therm. Eng. 29 (17-18) (2009) 3643–3653. [18] S. Sarangi, J.A. Weibel, S.V. Garimella, Effect of particle size on surface-coating enhancement of pool boiling heat transfer, Int. J. Heat Mass Transfer 81 (2015) 103–113. [19] D. Deng, W. Wan, J. Feng, Q. Huang, Y. Qin, Y. Xie, Comparative experimental study on pool boiling performance of porous coating and solid structures with reentrant channels, Appl. Therm. Eng. 107 (2016) 420–430. [20] S. Zhang, Y. Tang, J. Zeng, W. Yuan, J. Chen, C. Chen, Pool boiling heat transfer enhancement by porous interconnected microchannel nets at different liquid subcooling, Appl. Therm. Eng. 93 (2016) 1135–1144. [21] Y. Tang, J. Zeng, S. Zhang, C. Chen, J. Chen, Effect of structural parameters on pool boiling heat transfer for porous interconnected microchannel nets, Int. J. Heat Mass Transfer 93 (2016) 906–917. [22] A.M. Gheitaghy, H. Saffari, M. Mohebbi, Investigation pool boiling heat transfer in U-shaped mesochannel with electrodeposited porous coating, Exp. Therm Fluid Sci. 76 (2016) 87–97. [23] D. Ciloglu, A. Bolukbasi, A comprehensive review on pool boiling of nanofluid, Appl. Therm. Eng. 84 (2015) 45-63. [24] K.C. Leong, J.Y. Ho, K.K. Wong, A critical review of pool and flow boiling heat transfer of dielectric fluids on enhanced surfaces, Appl. Therm. Eng. 112 (2017) 999–1019. [25] R. H. Nilson, S.K. Griffiths, S.W. Tchikanda, M.J. Martinez, Axially tapered microchannels of high aspect ratio for evaporative cooling devices, J. Heat Transfer 126 (3) (2004) 453. [26] Y. Tang, P. Chen, X. Wang, Experimental investigation into the performance of heat pipe with micro grooves fabricated by Extrusion–ploughing process, Energy Convers. Manage. 51 (10) 25
(2010) 1849–1854. [27] X. Wang, Y. Tang, P. Chen, Investigation into performance of a heat pipe with micro grooves fabricated by extrusion–ploughing process, Energy Convers. Manage. 50 (5) (2009) 1384–1388. [28] D. Deng, Y. Tang, J. Zeng, S. Yang, H. Shao, Characterization of capillary rise dynamics in parallel micro V-grooves, Int. J. Heat Mass Transfer 77 (2014) 311–320. [29] Y. Tang, J. Xiang, Z. Wan, W. Zhou, L. Wu, A novel miniaturized loop heat pipe, Appl. Therm. Eng. 30 (10) (2010) 1152–1158. [30] J.-h. Xiang, C.-l. Zhang, F. Jiang, X.-c. Liu, Y. tang, Fabrication and testing of phase change heat sink for high power LED, Trans. Nonferrous Met. Soc. China 21 (2011) 2066-2071. [31] S.-l. Tao, Y. Tang, M.-q. Pan, Z.-p. Wan, L.-s. Lu, Forming method of wick structure for heat column, Trans. Nonferrous Met. Soc. China 21 (2011) 1587-1594. [32] P. Chen, Y. Tang, X.-k. Liu, X.-q. Liu, Formation of integral fins function-surface by extrusion-ploughing process, Trans. Nonferrous Met. Soc. China 16 (2006) 1029-1034. [33] Y. Chi, Y. Tang, J.-c. Chen, X.-x. Deng, L. Liu, Z.-p. Wan, X.-q. Liu, Forming process of cross-connected finned micro-grooves in copper strips, Trans. Nonferrous Met. Soc. China 17 (2007) 267-272. [34] C.-J. Kuo, Y. Peles, Local measurement of flow boiling in structured surface microchannels, Int. J. Heat Mass Transfer 50 (23-24) (2007) 4513–4526. [35] S. Zhou, X. Xu, B.G. Sammakia, Modeling of boiling flow in microchannels for nucleation characteristics and performance optimization, Int. J. Heat Mass Transfer 64 (2013) 706–718. [36] J. Zeng, L. Lin, Y. Tang, Y. Sun, W. Yuan, Fabrication and capillary characterization of micro-grooved wicks with reentrant cavity array, Int. J. Heat Mass Transfer 104 (2017) 918–929. [37] A. Shahriari, J. Wurz, V. Bahadur, Heat transfer enhancement accompanying Leidenfrost state suppression at ultrahigh temperatures, Langmuir 30 (40) (2014) 12074-12081. [38] J.R. Taylor, An introduction to Error Analysis, second ed., University Science Books, 1997. [39] A.S. Moita, E. Teodori, A.L.N. Moreira, Influence of surface topography in the boiling mechanisms, Int. J. Heat Fluid Flow 52 (2015) 50-63. [40] Z.G. Qu, Z.G. Xu, C.Y. Zhao, W.Q. Tao, Experimental study of pool boiling heat transfer on horizontal metallic foam surface with crossing and single-directional V-shaped groove in saturated water, Int. J. Multiphase Flow 41 (2012) 44-55. 26
[41] D. Deng, J. Feng, Q. Huang, Y. Tang, Y. Lian, Pool boiling heat transfer of porous structures with reentrant cavities, Int. J. Heat Mass Transfer 99 (2016) 556-568. [42] K.N. Rainey, S.M. You, S. Lee. Effect of pressure, subcooling, and dissolved gas on pool boiling heat transfer from microporous, square pin-finned surfaces in FC-72, Int. J. Heat Mass Transfer 46 (1) (2003) 23-35. [43] J. Mikielewicz, D. Mikielewicz. A simplified energy dissipation based model of heat transfer for subcooled flow boiling, Int. J. Heat Mass Transfer 106 (2017) 280-288. [44] X. Ji, J. Xu, Z. Zhao, W. Yang, Pool boiling heat transfer on uniform and non-uniform porous coating surfaces, Exp. Therm Fluid Sci. 48 (7) (2013) 198-212. [45] N.I. Kolev, The influence of mutual bubble interaction on the bubble departure diameter, Exp. Therm Fluid Sci. 8 (2) (1994) 167-174. [46] L.-H. Chien, R.L. Webb, Measurement of bubble dynamic on an enhanced boiling surface, Exp. Therm Fluid Sci. 16 (1998) 177-186. [47] J.P. McHale, S.V. Garimella, Bubble nucleation characteristics in pool boiling of a wetting liquid on smooth and rough surfaces, Int. J. Multiphase Flow 36 (2010) 249-260. [48] M.G. Cooper, A.J.P. Lloyd, The microlayer in nucleate pool boiling, Int. J. Heat Mass Transfer 12 (1969) 895-913.
Figure captions Fig. 4 Characterization of MSRC-1.
Fig. 2. Photograph of the test platform. Fig. 3. Schematic of the test platform (a), details of the arrangement of thermocouples (b) and cartridge heaters (c). Fig. 4. Boiling curves for the MSRCs and SCS at different liquid subcooling degrees: (a) MSRC-1 and SCS; (b) MSRC-2 and SCS. Fig. 5. HTC and heat flux relationship at different degrees of liquid subcooling: (a) MSRC-1 and SCS; (b) MSRC-2 and SCS. Fig. 6. Bubble visualizations of MSRC-1 ( Tsub = 0 °C ). Fig. 7. Bubble visualizations of SCS ( Tsub = 0 °C ). 27
Fig. 8. High-speed images of the departing bubble at micro-grooves of MSRC-1 (upper row) and reentrant cavities of MSRC-1 (middle row) and the SCS (bottom row) at different heat fluxes. Fig. 9. Average bubble departure diameter as a function of heat flux. Fig. 10. Effects of subcooling on the bubble behaviors of MSRC-1 sample at high heat flux: (a) qa = 669.32 kW/m2; (b) qa = 670.73 kW/m2; (c) qa = 670.71 kW/m2. Fig. 11. Average bubble departure diameter with variation of liquid subcooling degrees. Table 1. Structural parameters for the samples tested in this study.
28
Highlights
MSRCs promote the ONB and yield higher HTC than smooth surface
Cavity bubble and groove bubble are nucleated on the MSRCs
Groove bubble increases for qa < 465.58 kW/m2 but decreases sharply afterwards
Liquid subcooling has little influence on the departure diameter of cavity bubble
29
