An experimental study of flow boiling heat transfer from porous foam structures in a channel

Accepted Manuscript An Experimental Study of Flow Boiling Heat Transfer from Porous Foam Structures in a Channel I. Pranoto, K.C. Leong PII:

S1359-4311(14)00291-9

DOI:

10.1016/j.applthermaleng.2014.04.027

Reference:

ATE 5551

To appear in:

Applied Thermal Engineering

Received Date: 18 June 2013 Revised Date:

28 March 2014

Accepted Date: 12 April 2014

Please cite this article as: I. Pranoto, K.C. Leong, An Experimental Study of Flow Boiling Heat Transfer from Porous Foam Structures in a Channel, Applied Thermal Engineering (2014), doi: 10.1016/ j.applthermaleng.2014.04.027. 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.

ACCEPTED MANUSCRIPT

Research Highlights “Pocofoam” 61% porosity foam enhanced the cooling performance by up to 2.5 times.

•

The coolant mass flux had affected significantly the cooling performance.

•

The evaporator gap had significant effect on the flow boiling heat transfer.

•

Flow boiling heat transfer coefficient of 16.5 kW/m2·K was achieved in this study.

•

“Laminar” and “turbulent” bubble regimes were observed in this study.

AC C

EP

TE D

M AN U

SC

RI PT

•

ACCEPTED MANUSCRIPT An Experimental Study of Flow Boiling Heat Transfer from Porous Foam Structures in a Channel

I. Pranoto and K.C. Leong* School of Mechanical and Aerospace Engineering, Nanyang Technological University

Abstract

RI PT

50 Nanyang Avenue Singapore 639798, Singapore

This paper presents a study of the flow boiling heat transfer and bubble characteristics from

SC

porous graphite foam structures in a channel. The flow boiling performance and phenomena for different coolant mass fluxes, evaporator gaps, and foam properties were investigated

M AN U

with a two-phase cooling experimental facility. Two types of graphite foams with different thermophysical properties viz. “Pocofoam” of 61% porosity and “Kfoam” of 72 % porosity were used as the enhanced evaporator structure and tested with the dielectric liquid FC-72. Evaporator gaps of 6, 4, and 2 mm were tested with coolant mass fluxes of 50, 100, and 150 kg/m2·s. The experimental results show that the evaporator gap, coolant mass flux, and foam properties have effects on the flow boiling characteristics and performance. It was found that

TE D

the use of “Pocofoam” of 61% porosity and “Kfoam” of 72% porosity had enhanced the boiling heat transfer coefficients by up to 2.5 and 1.9 times, respectively as compared to those of a smooth surface. The results also show that the boiling performance increases with the increase of coolant mass flux and the evaporator gap. “Pocofoam” graphite foam of 61%

EP

porosity exhibits better boiling performance than “Kfoam” graphite foam of 72% porosity for all tested cases. From this study, a maximum local flow boiling heat transfer coefficient of

AC C

16.5 kW/m2·K was achieved with “Pocofoam” of 61% porosity at 6 mm gap and the mass flux of 150 kg/m2·s.

Flow boiling images were recorded to study the flow boiling

phenomena and the bubble growth mechanism. “Laminar” and “turbulent” bubble regimes were observed from these images and larger “mushroom cloud” bubbles and a vapor film layer were identified at a heat flux of 83.3 W/cm2.

Keywords: two-phase cooling; graphite foam; evaporator gaps; flow boiling heat transfer; flow boiling in a channel; bubble dynamics. * Corresponding author. Tel: +65-6790-5596; fax: +65-6792-4062. E-mail address: [email protected] (K.C. Leong) 1

ACCEPTED MANUSCRIPT Nomenclature free flow area of the channel (m2)

Ah

base area of heater (m2)

cp

specific heat (kJ/kg·K)

d

evaporator gap (mm)

dp

pore diameter (m)

F

dynamic factor of single-phase convection

fd

bubble departure frequency (Hz)

G

coolant mass flux (kg/m2·s)

hfc

single phase convection heat transfer coefficient (kW/m2·K)

HF

graphite foam height (mm)

hfb

flow boiling heat transfer coefficient (kW/m2·K)

hfb(avg)

average value of flow boiling heat transfer coefficients (kW/m2·K)

hlv

latent heat of vaporization of the coolant (kJ/kg)

hnb

nucleate boiling heat transfer coefficient (kW/m2·K)

I

current (A)

k

thermal conductivity (W/m·K)

M

number of frames during the waiting time

N

number of frames during the departure time

Psat

saturation pressure (Pa)

q

heating power (W)

q"

heat flux (W/cm2)

S

dynamic factor of nucleate boiling

td

bubble departure time (s)

Tl

liquid temperature (°C)

Tw tw V

SC

M AN U

TE D

EP

AC C

Tsat

RI PT

Aff

saturation temperature (°C)

wall temperature (°C) bubble waiting time (s) voltage (V)

∆TW ( d x − d1 )

average wall temperature difference between the porous channels with various evaporator gaps to that at d1 (°C)

∆TW (G1 −Gx )

average wall temperature difference between G1 and various coolant mass fluxes (°C)

2

ACCEPTED MANUSCRIPT h fbGx

ratio of hfb(avg) values of the porous channel with various mass fluxes to that at G1

h fbG1

ratio of hfb(avg) values of the porous channel at d1 to that evaporator gaps of d2 and d3

h fbd1 h fbd x

β

total internal-surface-area-to-volume ratio (m2/m3)

ε

porosity, relative heat loss

ρ

graphite foam density (kg/m3) surface tension (mN/m)

σ

f

liquid phase

GF

graphite foam

v

vapor phase

i

channel inlet

1. Introduction

M AN U

channel exit

TE D

e

SC

Subscripts

RI PT

Greek Symbols

An effective cooling system for electronic devices must be able to remove high heat fluxes from compact surfaces and maintain the devices at the desired low temperature. The

EP

development of cooling technologies for electronic devices had therefore evolved from natural convection and single-phase forced convection to phase change cooling systems using pool and flow boiling. Many research efforts on the enhancement of boiling performance

AC C

have focused on increasing the boiling effective surface area [1-3]. Pool boiling systems based on the thermosyphon are only applicable for stationary systems where the evaporator is always below the condenser. To overcome this limitation, the liquid from the condenser has to be pumped to the evaporator. Hence, flow boiling cooling systems have been developed for the removal of high heat fluxes. Many researchers have studied the use of pumped two-phase cooling for various applications. Rainey et al. [4] studied flow boiling from microporous coated surfaces in subcooled FC-72. Their experimental results showed that microporous surfaces outperformed plain surfaces and enhanced the Critical Heat Flux (CHF). Muwanga and Hassan [5] studied flow boiling heat transfer of FC-72 in a microtube by liquid crystal thermography. They 3

ACCEPTED MANUSCRIPT observed the effect of mass and heat fluxes on the heat transfer coefficient. Their results showed that the heat transfer coefficient was influenced slightly by heat flux at a lower mass flux (770 kg/m2⋅s) and was less dependent on heat flux at a higher mass flux (1040 kg/m2⋅s). They also observed that the wall temperature oscillated for a wide range under two-phase conditions with amplitude of 1 to 7 ºC for low heat flux and frequency of about 10 Hz.

for high power devices with water as the working fluid.

RI PT

Wang et al. [6] experimentally investigated micro capillary pumped loop (CPL) cooling They found that the groove-

enhanced surface design improved the thermal performance of the cooling system. They suggested that the micro CPL system must be used at low pressures and that automatic

SC

circulation had to be maintained for preheating to ensure the equilibrium of transmission without an additional power supply. In addition, the system started up successfully and depriming did not occur until the heat flux had reached 185.2 W/cm2. A MEMS-based

M AN U

integrated capillary pumped loop (CPL) cooling module was developed by Jung et al. [7]. The module consisted of an evaporator and condenser with cone-shaped capillary structure insertions. Their experimental results showed that the module possessed good transient characteristics. They stated that the CPL can handle a heat flux of 6.22 W/cm2 and an allowable evaporator surface temperature of up to 110ºC with FC-72 as the coolant. Further

TE D

investigations on pumped two-phase cooling were carried out by Agostini et al. [8]. They developed a pumped two-phase cooling system with a silicon multi-microchannel heat sink using refrigerant R236fa as the working fluid. They stated that the system had successfully removed heat fluxes up to 255 W/cm2 and maintained the base chip temperature below 52ºC.

EP

They also reported that the flow boiling heat transfer coefficient increased with heat flux, decreased with mass velocity, and decreased slightly with vapor quality.

AC C

Currently, porous media are increasingly being used as enhanced boiling surfaces. Porous foams with large surface-area-to-volume ratio, low density, and high bulk thermal conductivity can be used as boiling evaporators for high heat flux device systems. Many studies have been conducted to apply porous media to enhance the performance of thermal management on the electronic devices [9-14]. Many research investigations on the application of porous material in two-phase flow cooling systems have been conducted over the years. Zhao et al. [15] investigated flow boiling heat transfer in horizontal metal-foam filled tubes with R134a as the working fluid. Their experimental results showed that the heat transfer coefficient was almost doubled by reducing the cell size from 20 to 40 pores per inch (PPI) for a given porosity. They also stated that the boiling heat transfer coefficient kept

4

ACCEPTED MANUSCRIPT increasing steadily, albeit slowly, by increasing the vapor quality for high mass fluxes, while the same trend was not observed for low mass fluxes.

They obtained heat transfer

coefficients of copper foam tubes which were approximately three times higher than that of plain tubes. An experimental study on the flow boiling heat transfer performance of FC-72 on silicon chips was conducted by Ma et al. [16]. They fabricated micro-pin-fins on the chip

RI PT

surface by using a dry etching technique to enhance the boiling heat transfer. Different fluid velocities, liquid subcooling and chip configuration were investigated in their study. They found that all the micro-pin-finned surfaces showed considerable heat transfer enhancement and increased CHF compared to the smooth surface. They also stated that the CHF values for all surfaces increased with fluid velocity and subcooling. In their study, the maximum CHF

SC

value of nearly 150 W/cm2 was reached for chip PF30-120 at the fluid velocity of 2 m/s and liquid subcooling of 35 K. Kim et al. [17] studied the flow boiling characteristics of three

M AN U

porous copper foams viz. 95% porosity and 10 PPI, 95% porosity and 20 PPI, and 92% porosity and 20 PPI, which were soldered to a heated wall of a 10 mm wide × 37 mm long × 7 mm high channel. Their experimental results showed that the high porosity and large pore size foam, i.e., the 95%, 10 PPI copper foam, gave the best result, achieving a heat transfer coefficient of 10 kW/m2·K. They stated that the results for water were in good agreement

TE D

with the available sparse porous matrix correlations, using the recommended dispersion conductivity coefficient of 0.06. Finally, the results of their study indicated that the porous foam had more influence in enhancing the convective heat transfer coefficient for FC-72 than water despite its relatively low dispersion coefficient.

EP

In summary, flow boiling systems provide the advantages of enhanced cooling performance and orientation flexibility compared to pool boiling systems. It is noted from

AC C

the above brief review that porous media enhance the cooling performance and increase the CHF. However, there are still many aspects and parameters pertaining to the mechanism and characteristics of flow boiling on porous media evaporator that need to be explored, especially in highly conductive graphite foams. To the best of the authors’ knowledge, the literature in this area is still rather scarce. In this study, a pumped flow boiling cooling system was developed by using porous graphite foams as the evaporator insert in the channels.

The effects of the coolant mass flux, evaporator gap, and thermophysical

properties of the graphite foam on the bubble characteristics and performance were investigated.

5

ACCEPTED MANUSCRIPT 2. Design of Experiments 2.1 Experimental facility In this study, a pumped two-phase cooling facility was designed to remove the heat from a heat simulator of dimensions 80 mm long × 60 mm wide. The designed facility consists of six main parts: flow boiling channel (test section), heat simulator, air-cooled condenser,

RI PT

micro-pump, and tubing system. The heat simulator consists of eight cartridge heaters which were installed in a copper block to generate up to 5000 W of heat. It was insulated with a Teflon base to reduce the heat loss. The evaporator channel was located and clamped tightly on top of the heater. Conductive thermal grease, “Omegatherm 201”, was applied to reduce The

SC

the thermal contact resistance between the heater and evaporator channel surface.

evaporator channel was made of aluminum and designed with four channels. Three sets of

M AN U

guide vanes connect the channels and reduce the overall flow resistance. Graphite foam inserts were attached on the bottom of channels by conductive thermal adhesive, “Omegabond 101”. The evaporator cover was also made of aluminum and clamped tightly to the evaporator through 18 M4 screws. A square rubber ring was attached to the interface between the evaporator channel and cover to ensure air- and liquid-tightness of the system. The dielectric liquid coolant, FC-72, was pumped to the evaporator channel by a Cole-

TE D

Parmer “Micropump R-73011-18” pump equipped with a high-precision Cole-Parmer gear pump drive which can be varied from 0 ~ 5000 rpm and can provide accurate and pulseless fluid delivery. The chemical resistance of the pump made it compatible with a wide range of liquids including FC-72. The pump head was mounted on the motor and an inline filter was

EP

installed before the inlet of the pump to maximize its life span. The heater boils the phase change coolant inside the evaporator channel. The resulting liquid-vapor mixture flows to the

AC C

air-cooled condenser section. The condensed liquid returns to the reservoir before it is circulated by the pump. The schematic diagram of the facility and the design of the channel are shown in Figs. 1 and 2, respectively. The condenser was designed to ensure that the vapor can be completely condensed to its liquid state before it is circulated to the evaporator by the micro-pump. The condenser tubes were made of aluminum which is compatible with most coolants including FC-72. The condenser tubes were fitted with aluminum fins to enhance heat transfer with the aluminumbrazed joints thus ensuring that the system was leak-free. The condenser of dimensions 200 (Lc) × 50 (Wc) × 230 (Hc) mm was mounted on a metal base. Two fans were used to remove heat from the condenser. The speed of the fans can be controlled to vary the heat removal

6

ACCEPTED MANUSCRIPT rate. “Swagelok” stainless steel tubes and fittings were used in the entire closed-loop system to ensure leak-tightness. The tubing system was designed and tested for pressure operational conditions from -1 to 3.5 bar. Seven “Swagelok” needle valves were installed in the system to conduct pressure and vacuum tests and to charge/discharge the coolant.

RI PT

2.2 Evaporator inserts

To enhance the boiling heat transfer performance of the system, the graphite foam inserts were attached on the bottom of each channel on the evaporator. Two types of graphite foams viz. “Pocofoam” graphite foam of 61% porosity and “Kfoam” graphite foam of 72% porosity

SC

developed by Oak Ridge National Laboratory, USA, were used as evaporator inserts. The foam samples were fabricated into block structures with dimensions of 80 mm × 60 mm × 6 Scanning Electron Microscope (SEM)

M AN U

mm by Electrical Discharge Machining (EDM).

images of internal structure of the foams and the foams dimensions are shown in Fig. 3. Selected properties of the graphite foams are shown in Table 1. In all the experiments, there will always be a gap in the channel. It is important to note that with “Pocofoam” graphite foams of 61% porosity and “Kfoam” of 72% porosity, the coolant flows predominantly

2.3 Phase change coolant

TE D

through the evaporator gaps and any flow through the foams is considered to be negligible.

In two-phase cooling systems, the choice of the phase change coolant plays an important

EP

role in heat transfer performance. In particular, the thermophysical properties of the coolant directly influence the heat transfer coefficient of the system. In cooling systems used for

AC C

electronic devices, the coolant has to be chemically compatible with the components and possess large dielectric strength to prevent electrical connections. FC-72 was selected as the phase change coolant used in the experiments. The liquid is manufactured by 3M Ltd. and is widely used for electronic cooling applications. Selected thermophysical properties of FC-72 at 25ºC and 1 atm are shown in Table 2.

2.4 Data reduction and uncertainty analysis Temperatures, pressures, coolant flow rates, and heating powers were measured in the experiments. The electrical power can be adjusted by a power controller and displayed by a power meter. A total of 15 type K thermocouples were used in the system to measure various 7

ACCEPTED MANUSCRIPT temperatures. Eight flat thermocouples were inserted at the interface between the evaporator and heater surface to measure the average evaporator surface temperature.

Two

thermocouples were inserted in the copper block to measure the temperature of the heater and another two thermocouples were screwed to the evaporator cover to measure the liquid temperature. Three thermocouples were attached in the tubing system to measure the coolant

outlet.

RI PT

temperatures at the reservoir, at the location before the evaporator inlet, and at the evaporator Due to the difficulty of securing the thermocouples within the graphite foam,

temperatures within the foam were not measured. Two “Omega” PX409 pressure transducers were installed on the evaporator cover to measure the pressure inside the evaporator channel.

SC

The system volumetric flow rate was obtained by a “Digmesa” flow meter installed between the micro pump outlet and the evaporator inlet. All the sensors were carefully calibrated before the commencement of the experiments. The digital signals from the sensors were

M AN U

channeled to the “Yokogawa” (MX100) data acquisition system attached to a personal computer.

Pressure tests were performed before the start of the experiments to ensure that the system was leak-proof and system reliability. The steady-state condition was considered to be reached when the difference between two consecutive average heater wall temperatures was

TE D

smaller than 0.2°C. The tests showed that 15 minutes were required to reach quasi steady state from the initial condition, and 5 minutes for the system to reach the next steady state after increasing the power input. The heat flux was increased gradually from 4.2 to 83.3 W/cm2 in increments of 4.2 W/cm2. Systematic uncertainty can be minimized by careful

EP

experimentation. To minimize data reduction uncertainty, the time-averaging method was employed. The time-averaged value is defined as the average value of a parameter measured

AC C

for a prescribed duration (1 minute with a sampling rate of 200 ms) under quasi-steady-state condition.

The uncertainty of the thermocouple measurements is ±0.5°C while the

uncertainties of the pressure transducer, power meter, and flow meter measurements are within ±0.25%, ±0.3%, and ±0.5% of their full-scales, respectively. In this study, the uncertainty of hfb was determined by the method of Taylor [20]. The measured quantities are assumed to be independent and random and to follow normal distributions. In the experiments, q", Tw, and Tsat are functions of hfb and their uncertainties can be calculated from

8

ACCEPTED MANUSCRIPT  δ   δT   δT  δhb =  q "  +  W  +  sat  hb  q "   TW   Tsat  2

2

2

(1)

By using the above method, the average uncertainties of q", Tw, and Tsat were calculated to be 3.3%, 1.5%, and 1.8%, respectively. The average uncertainty of hfb was estimated to be 4%

RI PT

from Eq. (1). The heat loss (qL) of the test section was evaluated before the commencement of the boiling experiments. qL was calculated from the difference between the total power input (qt) to the cartridges heaters and the heat transfer rate to the single-phase liquid of the coolant

SC

through the channel (qC). With qt = VI and qC = GAff c p, f (Tf ,e − Tf ,i ) , the relative heat loss (ε) was calculated by

qt − GAff c p , f (T f ,e − T f ,i )

M AN U

ε=

qt

(2)

By using Eq. (2), the average value of ε was estimated to be about 7.5%. Tf,e and Tf,i in Eq. (2) are the liquid temperatures at the exit and inlet of the channel, respectively.

TE D

2.5 Experimental parameters

In order to investigate the cooling performance of the pumped two-phase cooling system, three parameters, viz. coolant mass flux (G), evaporator gap (d), and foam properties (i.e.

EP

porosity and pore diameter) were considered. In the experiments, different mass fluxes were obtained by adjusting the speed of the micro pump according to the respective free flow area of the channel (Aff). The evaporator gap is defined as the distance from the top of the porous

AC C

insert to the evaporator top surface as illustrated in Fig. 4. Evaporator gaps of 2 mm, 4 mm, and 6 mm were investigated in this study. The evaporator gaps were adjusted by changing the interface material attached on the top cover of the channel. Different thicknesses of the interface material were used to obtain the designated evaporator gaps.

In the study,

experiments on the aluminum smooth surface (i.e. evaporator channel without any graphite foam insert) were performed as a base case to evaluate the heat transfer enhancement of the porous channel (i.e. evaporator channel with the graphite foam inserts). The parameters tested in the experiments are summarized in Table 3.

The wall temperatures Tw were

determined by eight thermocouples placed at different locations on the heater surface and the average wall temperature was calculated as 9

ACCEPTED MANUSCRIPT Tw =

Tw1 + Tw2 + Tw3 + Tw4 + Tw5 + Tw6 + Tw7 + Tw8 8

(3)

The arrangement of the wall thermocouples on the evaporator is shown in Fig. 5.

RI PT

3. Results and Discussion The cooling performances of the pumped two-phase cooling system were analyzed based on the flow boiling heat transfer coefficient.

Many researchers have proposed various

correlations to determine the flow boiling heat transfer coefficient. As proposed by Chen

SC

[21], the local flow boiling heat transfer coefficient can be expressed as the superposition of forced convection and nucleate boiling heat transfer coefficients. A generalized hfb for both

M AN U

vertical and horizontal tubes was also proposed by Kandlikar [22] and Kandlikar and Balasubramanian [23]. In the current experiments, the heating powers were varied from 4.2 to 83.3 W/cm2 at intervals of 4.2 W/cm2. It was found that boiling incipience occurred for a heat flux of 20.8 W/cm2. Therefore, it can be concluded that nucleate boiling heat transfer had dominated the heat transfer process and thus, the flow boiling heat transfer coefficient can be determined from

q" (Tw − Tsat )

(4)

q" =

q Ah

(5)

EP

TE D

h fb =

AC C

In this study, the saturation temperatures Tsat of FC-72 were calculated at their corresponding saturated pressures Psat measured by the pressure transducer and from the equation given by the manufacturer [19] as

Tsat (K) =

1562 9.729 − log10 ( Psat ) 

(6)

The average values of the flow boiling heat transfer coefficients hfb(avg) were calculated to evaluate the boiling performance for different coolant mass fluxes, evaporator gaps, and graphite foam properties.

10

ACCEPTED MANUSCRIPT 3.1 Effects of Coolant Mass Flux Wall temperature readings and boiling curves derived from the experiments on “Pocofoam” graphite foam of 61 % porosity and “Kfoam” graphite foam of 72% porosity for various mass fluxes of 50, 100, and 150 kg/m2·s with evaporator gaps of 6 and 4 mm are presented in Figs. 6 and 7. As defined in Eq. (4), the flow boiling heat transfer coefficient hfb

RI PT

is inversely proportional to the superheat Tw - Tsat. For constant heating flux q", a lower superheat corresponds to a higher hfb. The single-phase convection heat transfer data are not shown in the boiling curves for purpose of clarity. The experimental results show that the coolant mass flux affects the wall temperatures and flow boiling heat transfer coefficients for

SC

a given gap size. The results show that regardless of the evaporator gap and the graphite foam used, the highest coolant mass flux resulted in the lowest wall temperature and the smallest wall superheat at all tested heat fluxes. The higher the mass flux, the better is the

M AN U

flow boiling heat transfer performance. The average wall temperature differences ∆TW (G1 −Gx ) between G1 and various coolant mass fluxes; and the ratio of hfb(avg) values of the porous channels with various mass fluxes to that at G1 defined as hfbGx (avg ) / hfbG1 (avg ) are presented in Table 4. Gx refers to the coolant mass fluxes which were varied from G2 to G3. It is shown that by increasing the mass flux from 50 kg/m2·s to 100 and 150 kg/m2·s the average wall

TE D

temperatures were reduced by up to 4.3 °C and 9.6°C, respectively. The results also show that enhancements of hfb(avg) by up to 12% and 34% were obtained by increasing the coolant mass fluxes from 50 kg/m2·s to 100 and 150 kg/m2·s, respectively. The boiling curves in Figs. 6(c) and 6(d) for “Pocofoam” foam of 61% porosity and Figs.

EP

7(c) and 7(d) for “Kfoam” foam of 72% show that the coolant mass flux affects significantly the flow boiling heat transfer performance. The results show that at a given heat flux, the

AC C

higher coolant mass fluxes had resulted in lower wall superheats for all evaporator gaps and heat fluxes. The values of hfb(avg) for coolant mass fluxes of 50, 100, and 150 kg/m2·s with “Pocofoam” foam of 61% porosity and “Kfoam” of 72% porosity at d = 6 mm and d = 4 mm are presented in Table 4. The results also show that the enhancements of hfb(avg) due to the increase of the mass flux with d = 6 mm and d = 4 mm are not significantly different which suggests that the flow boiling enhancement mechanisms for both tested gaps might be the same. In flow boiling systems, the mechanism of heat transfer is a combination of forced convection and nucleate boiling. In this study, the graphite foam is fully immersed in the coolant with the possibility of the liquid filling all the voids in the pore structure. The heat 11

ACCEPTED MANUSCRIPT from the heater will be conducted through the heater wall and porous inserts resulting in boiling of the liquid inside the pores. Subsequently, the supplied fresh coolant refills the pores and evaporating bubbles will be continuously generated. The coolant flows through the evaporator gaps enhancing the heat removal by forced convection and sweeping away the generated bubbles from the top surface of graphite foams. Therefore, higher coolant mass

RI PT

fluxes will increase the liquid replacement rate to the pores and increase the forced convection capacity resulting in an increase in the bubble departure frequency. The flow boiling mechanism for different coolant mass fluxes will be discussed in greater detail in

SC

Section 3.4.

3.2 Effect of Evaporator Gap

The effects of evaporator gap were investigated for “Pocofoam” graphite foam of 61%

M AN U

porosity and “Kfoam” graphite foam of 72% porosity with mass fluxes of 100 and 150 kg/m2·s. Wall temperature readings and boiling curves for different gaps tested with different graphite foam types and coolant mass fluxes are shown in Figs. 8 and 9. As expected, the experimental results show that the evaporator gap affects the wall temperatures and flow boiling heat transfer coefficient.

The results show that subsequent decreases of the

TE D

evaporator gap from 6 mm to 4 and 2 mm result in lower flow boiling performance. As shown in Figs. 8(a), 8(b), 9(a), and 9(b), the evaporator gaps affect the wall temperatures at a given mass flux. The results show that the wall temperatures had increased gradually when the evaporator gap was decreased from d = 6 mm to 4 and 2 mm. The average wall

EP

temperature differences for various evaporator gaps ∆TW ( d x −d1 ) at G = 100 kg/m2·s and G = 150 kg/m2·s for “Pocofoam” foam of 61% porosity and “Kfoam” of 72% porosity are

AC C

presented in Table 5, respectively where dx refers to the evaporator gaps of d2 and d3. It is shown that by decreasing the gap from 6 mm to 4 and 2 mm, the average wall temperatures were increased by about 3.4 ~ 3.9°C and 5.3 ~ 7.5°C, respectively. To evaluate the boiling performance difference, the hfb(avg) ratio of the evaporator gap of 6 mm to those at 4 and 2 mm hfbd1 / hfbdx are determined, where x refers to the evaporator gap of d2 and d3. The values of

hfbd1 / hfbdx for various evaporator gaps are presented in Table 5. The results show that the flow boiling heat transfer coefficients were reduced by about 5 ~ 6% and 20 ~ 23% by decreasing the evaporator gap from 6 mm to 4 and 2 mm, respectively. The superheats readings for various evaporator gaps on “Pocofoam” graphite foam of 61% porosity and “Kfoam” graphite foam of 72% porosity are shown in Figs. 8(c), 8(d) and 12

ACCEPTED MANUSCRIPT 9(c), 9(d), respectively. It can be seen that subsequent decreases of the evaporator gap from 6 mm to 4 and 2 mm result in higher superheats. By using Eq. (4), hfb for various evaporator gaps are further quantified. The hfb(avg) values of the porous channel with “Pocofoam” of 61% porosity and “Kfoam” of 72% porosity with G = 100 and 150 kg/m2·s can be found in Table 4. It is noted that the highest flow boiling heat transfer coefficient of 16.5 kW/m2·K

RI PT

was achieved by using “Pocofoam” graphite foam of 61% porosity with the evaporator gap of 6 mm and coolant mass flux of 150 kg/m2·s. It can be seen that the effects of decreasing the evaporator gap from 6 to 4 mm on the flow boiling heat transfer coefficients are less significant as compared to the decrease of the evaporator gap from 6 to 2 mm.

SC

Based on the results, it is found that the evaporator gap has an important role in the flow boiling mechanism from the graphite foam structures. In the boiling process, as the heat flux increase, the generated bubbles will coalesce and form larger bubbles on the graphite foam

M AN U

surface and in the channel gap. The sizes of the merged bubbles were found to be much larger than the individual isolated bubbles generated from the foam structure. At a constant mass flux, a smaller evaporator gap will tend to cause more bubble confinement as compared to larger evaporator gaps. The bubble confinement tends to lead to a decrease in the bubble departure frequency, vapor layer formation, and thus a decrease in the boiling performance.

Section 3.4.

TE D

The flow boiling mechanism from different evaporator gaps will be further discussed in

3.3 Effect of Graphite Foam Types

EP

As mentioned in the previous section, two graphite foams of different porosities and thermal conductivities were tested in the experiments. The wall temperatures and boiling curves for “Pocofoam” graphite foam of 61% porosity and “Kfoam” of 72% porosity graphite

AC C

foams with evaporator gaps of 6 and 4 mm at G = 150 kg/m2·s are presented in Fig. 10. The experimental results show that the thermophysical properties of the foams had significantly affected the flow boiling heat transfer. It can be seen that the wall temperatures obtained by “Pocofoam” graphite foam of 61% porosity are lower than those of “Kfoam” graphite foam of 72% porosity for all tested cases. The average difference of wall temperatures obtained by “Kfoam” graphite foam of 72% porosity and “Pocofoam” graphite foam of 61% porosity are found to be 8.0 and 8.4°C for the evaporator gaps of 6 and 4 mm, respectively. The boiling curves show that the use of “Pocofoam” graphite foam of 61% porosity result in much lower superheats compared to “Kfoam” graphite foam of 72% porosity. By using Eq. (4), the

13

ACCEPTED MANUSCRIPT values of hfb(avg) obtained using “Pocofoam” graphite foam of 61% porosity G = 150 kg/m2·s are 16.5 and 15.7 kW/m2·K for the evaporator gaps of 6 and 4 mm, respectively while for “Kfoam” graphite foam of 72% porosity, the values of hfb(avg) are found to be 13.1 and 12.5 kW/m2·K, respectively. It is also found that the hfb(avg) ratio of “Pocofoam” 61% to “Kfoam” 72% porosity graphite foams are 1.26 for both evaporator gaps of 6 and 4 mm.

RI PT

As shown in Fig. 10, the wall temperature readings and boiling curves of the smooth surface are also presented as a basis of comparison. It can be seen that the use of porous graphite foams have enhanced significantly the boiling performance of the smooth surface. The average difference of wall temperatures between the use of “Pocofoam” graphite foam of

SC

61% porosity and the smooth surface are found to be 21.1 and 19.6°C with the evaporator gaps of 6 mm and 4 mm, respectively. On the other hand, the average difference of wall temperatures obtained by “Kfoam” of 72% porosity and the smooth surface are about 15.1

M AN U

and 13.2°C, respectively. “Pocofoam” of 61% porosity had better flow boiling heat transfer performance compared to “Kfoam” of 72% porosity. It is also found that the hfb(avg) ratios of “Pocofoam” graphite foam of 61% porosity to the smooth surface are 2.49 and 2.44 for the evaporator gaps of 6 and 4 mm, respectively, while the hfb(avg) ratios of “Kfoam” graphite foam of 72% porosity to the smooth surface are 1.98 and 1.93, respectively.

TE D

As shown in Table 1, “Pocofoam” graphite foam of 61% porosity possesses higher effective thermal conductivity and lower porosity compared to “Kfoam” graphite foam of 72% porosity which has effects on the flow boiling mechanism. The combined effect of these two properties will determine the bubble departure frequency and the active nucleation

EP

site density of the porous graphite foam. The larger bubble departure frequency and larger nucleation site density of the “Pocofoam” graphite foam of 61% porosity account for its

AC C

superior flow boiling heat transfer performance as compared to “Kfoam” 72% porosity. This will be further discussed in Section 3.4.

3.4 Flow Boiling Mechanism and Discussions The flow boiling heat transfer from porous graphite foams with different coolant mass fluxes and evaporator gaps are discussed in Sections 3.1 ~ 3.3. As presented in Section 3.1, increases in the mass flux from 50 kg/m2·s to 100 and 150 kg/m2·s had increased the flow boiling performance by 12% and 34%, respectively. The flow boiling images for different mass fluxes on “Pocofoam” of 61% porosity and “Kfoam” of 72% porosity at q" = 83.3 W/cm2 for evaporator gaps of 6 and 4 mm are shown in Figs. 11 and 12, respectively. They

14

ACCEPTED MANUSCRIPT show that the increase of the coolant mass flux has reduced the number of the coalesced flowing bubbles on the graphite foam surfaces and evaporator gaps. It can be seen that there was a significant reduction in the number of coalesced bubbles when the mass flux was increased from 50 to 100 kg/m2·s. This observation supports the experimental results of enhancement of flow boiling heat transfer due to an increase in the coolant mass flux. The

RI PT

higher coolant mass flux will sweep away the generated and coalesced bubbles more effectively from the graphite foam surfaces and evaporator gap and increase the liquid replacement rate to the graphite foam pores. Combining these mechanisms with the heat removal enhancement by forced convection from the higher coolant mass flux resulted in the

SC

increase of the bubble departure frequency. It is also shown that there are less significant differences in the bubble phenomena on the graphite foam for mass fluxes of 100 and 150 kg/m2·s as compared to those of 50 and 100 kg/m2·s. Therefore, further increase of the

M AN U

coolant mass flux beyond 150 kg/m2·s might not produce significant difference in the boiling phenomena. Therefore, the selection of the operation coolant mass flux for a flow boiling cooling system should take into consideration the cooling performance, cost efficiency and the system requirements.

As discussed in Section 3.2, the evaporator gaps have affected the flow boiling heat

TE D

transfer coefficient on the graphite foam structures. By decreasing the evaporator gap from 6 to 4 mm, the flow boiling performance was reduced slightly by 5 ~ 6% for all tested graphite foams and coolant mass fluxes. Further decrease of the evaporator gap from 6 to 2 mm resulted in a significant decrease in the boiling performance by 20 ~ 30 %. The flow boiling

EP

images from the graphite foams for various evaporator gaps with the coolant mass fluxes of 100 and 150 kg/m2·s are shown in Figs. 13 and 14, respectively. At the same heat flux and coolant mass flux, much larger coalesced bubbles were formed in the smaller evaporator

AC C

gaps, especially at d = 2 mm and G = 100 kg/m2·s. The larger coalesced bubbles on the graphite foam surfaces and evaporator gaps will result in more bubble confinement and vapor layers covering the top foam surface leading to an increase in the graphite foam local temperatures. Simultaneously, it will also diminish fresh liquid replacement rate to the graphite foam pore structures and decrease the bubble departure frequency. Further decrease of the evaporator gap below 2 mm may cause dry-out on the porous foam which will significantly increase the surface temperatures, decrease the cooling performance and result in the appearance of CHF. Therefore, the selection of the evaporator gap in the flow boiling system becomes important to achieve maximum cooling performance and system reliability.

15

ACCEPTED MANUSCRIPT In this study, the evaporator gap of 6 mm gave better flow boiling performance for all tested cases. Based on the captured images of boiling process, several bubble regimes can be observed. The first regime is termed as “laminar” in which isolated bubbles were generated intermittently from the active cavities. This bubble regime was generally observed from the

RI PT

boiling process at low heat flux for all mass flux levels. By increasing the heat flux, the number of active nucleation sites was also increased. The bubble regime at high heat flux is termed as the “turbulent” regime. In this regime, some bubbles had coalesced with their predecessors and formed larger bubbles. When the heat flux was further increased, more

SC

coalesced bubbles in the channel were formed. This phenomenon is more obvious in the captured images of high heat flux with low mass flux. Similar flow boiling phenomena and regimes were also found in the visualization study performed by Ramaswamy et al. [3]. The

M AN U

continuous bubble coalescence process had led to the formation of larger “mushroom cloud” bubbles and a thin vapor layer on the porous foam surface as shown in Figs. 15 and 16, respectively.

As discussed in Section 3.3, the porous graphite foams have significantly enhanced the flow boiling performance of the smooth aluminum surface. In flow boiling from the smooth

TE D

surface, the nucleation bubbles were generated from the smooth aluminium surface. It was observed from the experiments that relatively fewer bubbles departed from the smooth surface. This is due to the relatively fewer active nucleation sites from the smooth aluminium surface. By attaching the graphite foam inserts, the number of active nucleation sites was The small

EP

increased significantly due to the increased total exposed area to the fluid.

openings between the pores allow the liquid to fill the entire pore network which in turn leads

AC C

to a larger number of nucleation bubbles. The significant role of the internal pore structure of the graphite foam on the enhancement of nucleation sites was discussed in our previous work [24]. It was also mentioned in Section 3.3 that the use of “Pocofoam” graphite foam of 61% porosity resulted in significantly higher boiling heat transfer coefficients as compared to “Kfoam” graphite foam of 72% porosity. The superior boiling heat transfer performance of “Pocofoam” graphite foam of 61% porosity compared to “Kfoam” graphite foam of 72% porosity can be deduced from the properties of the foams. As shown in Table 1, “Pocofoam” graphite foam of 61% porosity possesses higher effective thermal conductivity compared to “Kfoam” foam of 72% porosity.

In the boiling process, the porous channels with

“Pocofoam” graphite foam of 61% porosity inserts will conduct the heat more effectively and

16

ACCEPTED MANUSCRIPT the saturation temperature of the foam will be achieved in a shorter time which results in higher bubble departure frequency. Based on the physical properties of the foams, the total internal surface-area-to-volume ratio (β) of the porous graphite foam was determined using the model proposed by Yu et al. [25].

The calculated values of β for “Pocofoam” graphite foam of 61% porosity and

RI PT

“Kfoam” graphite foam of 72% porosity are about 10400 and 5900 m2/m3, respectively. On the other hand, the unit cell model of the graphite foam proposed by Leong and Li [26] yields values of β for “Pocofoam” of 61% porosity and “Kfoam” of 72% porosity graphite foams that are about 19500 and 12500 m2/m3, respectively. The differences in the β values can be

SC

attributed to the different geometric and selected parameters of the graphite foam unit cell used in the two models which nevertheless show that the “Pocofoam” graphite foam of 61% porosity has a much higher internal surface-area-to-volume ratio compared to the “Kfoam”

M AN U

graphite foam of 72% porosity. Therefore, more active nucleation sites will be generated within the “Pocofoam” graphite foam of 61% porosity resulting in higher flow boiling heat transfer coefficients. Flow boiling images from “Pocofoam” of 61% and “Kfoam” of 72% with different evaporator gaps and heat fluxes are shown in Figs. 17 and 18. It can be seen that the number of generated bubbles from “Pocofoam” graphite foam of 61% porosity is

TE D

much larger than “Kfoam” graphite foam of 72% porosity for all tested cases. This is in line with the results of the effect of the foam types and provides the evidence that “Pocofoam” graphite foam of 61% porosity gives a higher nucleation site density compared to “Kfoam” graphite foam of 72% porosity.

EP

The bubble departure frequency was determined and analyzed to further explore the different boiling performance from the graphite foams. To determine the bubble departure

AC C

frequency, the bubble growth and departure processes were recorded by using a high speed camera at 2005 frames per second (fps). The captured images were analyzed frame by frame to determine the bubble growth and departure phenomena. The bubble departure frequency calculation method is portrayed in Fig. 19 which shows that a period of bubble growth and departure can be divided into waiting and bubble departure times. Once these times are determined, the bubble departure frequency can be calculated by fd =

1 t w + td

17

(7)

ACCEPTED MANUSCRIPT where tw and td are waiting and departure times, respectively. At the frame rate of 2005 fps, the number of frames in the waiting and departure periods are M and N, respectively. Therefore, tw and td can be calculated as (8)

td = (N frames/2005 fps)

(9)

RI PT

tw = (M frames/2005 fps)

By using the above method, the average bubble departure frequency of “Pocofoam” graphite foam of 61% porosity and “Kfoam” graphite foam of 72% porosity at 45.8 W/cm2 were found to be 163 and 145 Hz, respectively. At 58.3 W/cm2, the average values of fd are 169 and 150 Hz for “Pocofoam” graphite foam of 61% and “Kfoam” of 72% porosity,

SC

respectively. Therefore, the combination of larger total internal surface area and higher thermal conductivity of “Pocofoam” graphite foam of 61% porosity gives rise to a higher

M AN U

active nucleation site density and higher bubble departure frequency and thus, better boiling heat transfer performance.

4. Conclusion

In this study, the effects of coolant mass flux, evaporator gap, and graphite foam

TE D

properties were investigated. The flow boiling heat transfer enhancements of two different graphite foams were evaluated and compared with the boiling performance of the smooth surface. A visualization study was performed to understand the flow boiling phenomena and bubble characteristics. The following conclusions can be drawn:-

EP

1) The coolant mass flux has affected significantly the flow boiling cooling performance. The experimental results show that the increase of mass fluxes from 50 kg/m2·s to 100

AC C

and 150 kg/m2·s have increased the boiling heat transfer coefficients up to 12 % and 34%, respectively. Higher coolant mass fluxes are able to sweep away the generated and flowing bubbles effectively and increase the liquid replacement rate to the pore structures. In addition, the larger forced convection heat transfer from the higher coolant mass flux resulted in the increase of the bubble departure frequency. 2) The experimental results show that the evaporator gap had affected the flow boiling performance. It was found that the flow boiling heat transfer coefficients were reduced by about 5 ~ 6% and 20 ~ 23% by decreasing the evaporator gap from 6 mm to 4 and 2 mm, respectively. From the study, a smaller evaporator gap produces larger coalesced bubbles on the graphite foam surfaces and evaporator gaps and cause bubble

18

ACCEPTED MANUSCRIPT confinement and vapor layers which cover the top foam surface resulting in significantly increase of the wall temperatures.

In addition, the fresh liquid

replenishment rate to the graphite foam pore structures will also be diminished which resulted in the decrease of the bubble departure frequency. It was found that the highest flow boiling heat transfer coefficient of 16.5 kW/m2·K was achieved by using

RI PT

“Pocofoam” graphite foam of 61% porosity with the evaporator gap of 6 mm. 3) With the use of “Pocofoam” graphite foam of 61% porosity and “Kfoam” graphite foam of 72% porosity, the flow boiling heat transfer coefficients were increased significantly by up to 2.5 and 1.9 times, respectively as compared to those of the

SC

smooth surface. In addition, the wall temperatures were reduced by 19.6 ~ 21.1°C with the use of the porous graphite foams. The graphite foam thermophysical properties and its interconnected pores structures play an important role in the flow boiling

M AN U

enhancements. It was also found that the use of “Pocofoam” of 61% porosity resulted in a heat transfer coefficient which was 1.3 times higher than that of “Kfoam” 72% porosity. The difference of flow boiling heat transfer performance from the tested graphite foam types was evaluated by determining the surface-area-to-volume ratio and the bubble departure frequency. It was found that the surface-area-to-volume ratio of

TE D

the “Pocofoam” 61% porosity is about 1.2 ~ 1.3 times larger than those of “Kfoam” 72% porosity while the measured bubble departure frequency from “Pocofoam” of 61% porosity is 1.6 ~1.7 times higher than those of “Kfoam” of 72% porosity. 4) The visualization study shows that larger coalesced bubbles and bubble confinements

EP

were observed with the decreases of the coolant mass flux and evaporator gap. The flow boiling images also show that larger bubble density was generated by the use of

AC C

“Pocofoam” 61% porosity as compared to “Kfoam” 72% porosity. “Laminar” and “turbulent” flow boiling regimes were observed in the flow boiling process. Typical “mushroom cloud” bubbles and vapor film layers were found from the boiling process at high heat flux and low coolant mass flux which indicated that the boiling process had approached the onset of CHF.

Appendix A. Supplementary material Supplementary material of the flow boiling video associated with this article can be found in this link: https://www.dropbox.com/sh/hru72qnieo9csuk/t9GL3becNP

19

ACCEPTED MANUSCRIPT Acknowledgement The authors gratefully acknowledge the contribution of Mr. Hui Peng Loy for carrying out some of the experiments and assisting in the data collection.

RI PT

References S.M. You, T.W. Simon, A. Bar-Cohen, A technique for enhancing boiling heat transfer with application to cooling of electronic equipment, IEEE Transactions on Components, Hybrids and Manufacturing Technology 15 (1992) 823-831.

[2]

I. Mudawar, T.M. Anderson, Optimization of enhanced surfaces for high flux chip cooling by pool boiling, ASME Journal of Electronic Packaging 115 (1993) 89-100.

[3]

C. Ramaswamy, Y. Joshi, W. Nakayama, W.B. Johnson, High speed visualization of boiling from an enhanced structure, International Journal of Heat and Mass Transfer 45 (2002) 4761-4771.

[4]

K.N. Rainey, G. Li, S.M. You, Flow boiling heat transfer from plain and microporous coated surfaces in subcooled FC-72, ASME Journal of Heat Transfer 123 (2001) 918925.

[5]

R. Muwanga, I. Hassan, A flow boiling heat transfer investigation of FC-72 in a microtube using liquid crystal thermography, ASME Journal of Heat Transfer 129 (2007) 977- 987.

[6]

C.T. Wang, T.S. Leu, T.M. Lai, Micro capillary pumped loop system for a cooling high power device, Experimental Thermal and Fluid Science 32 (2008) 1090-1095.

[7]

J.Y. Jung, H.S. Oh, D.K. Lee, K.B. Choi, S.K. Dong, H.Y. Kwak, A capillarypumped loop (CPL) with microcone-shaped capillary structure for cooling electronic devices, Journal of Micromechanics and Microengineering 18 (2008) 1-7.

[8]

B. Agostini, J.R. Thome, M. Fabbri, B. Michel, High heat flux two-phase cooling in silicon multimicrochannels, IEEE Transactions on Components and Packaging Technology, 31 (2008) 691-701.

[9]

J.W. Klett, A.D. McMillan, N.C. Gallego, C.A. Walls, The Role of structure on the thermal properties of graphitic foams, Journal of Materials Science 39 (2000) 36593676.

[11]

M AN U

TE D

EP

AC C

[10]

SC

[1]

J.W. Klett, M. Trammel, Parametric investigation of a graphite foam evaporator in a thermosyphon with fluorinert and silicon CMOS chip, IEEE Journal on Device and Materials Reliability 4 (2004) 626-637. M.S. El-Genk, J.L. Parker, Nucleate boiling of FC-72 and HFE-7100 on porous graphite at different orientations and liquid subcooling, Energy Conversion and Management 49 (2008) 733-750.

[12]

L.W. Jin, K.C. Leong, I. Pranoto, H.Y. Li, J.C. Chai, Experimental study of a twophase thermosyphon with porous graphite foam insert, ASME Journal of Thermal Science and Engineering Applications 3 (2011) 024502-1 – 024502-6.

[13]

L.W. Jin, K.C. Leong, I. Pranoto, Saturated pool boiling heat transfer from highly conductive graphite foams, Applied Thermal Engineering 31 (2011) 2685-2693. 20

ACCEPTED MANUSCRIPT H.Y. Li, K.C. Leong, Experimental and numerical study of single and two-phase flow and heat transfer in aluminum foams, International Journal of Heat and Mass Transfer 54 (2011) 4904-4912.

[15]

C.Y. Zhao, W. Lu, S.A. Tassou, Flow boiling heat transfer in horizontal metal-foam tubes, ASME Journal of Heat Transfer 131 (2009) 121002-1 - 121002-8.

[16]

A. Ma, J. Wei, M. Yuan, J. Fang, Enhanced flow boiling heat transfer of FC-72 on micro-pin-finned surfaces, International Journal Heat and Mass Transfer 52 (2009) 2925-2931.

[17]

D.W. Kim, A. Bar-Cohen, B. Han, Forced convection and flow boiling of dielectric liquid in a foam filled channel, 11th Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems, ITHERM, USA, 2008, pp. 8684.

[18]

Poco Graphite Material Specification, 2008.

[19]

3M Fluorinert Electronic Liquid FC-72 Product Information, 2009.

[20]

J.R. Taylor, An Introduction to Error Analysis, 2nd Edition, University Science Books, California, 1997.

[21]

J.C. Chen, A correlation for boiling heat transfer to saturated fluids in convective flow, Industrial and Engineering Chemistry Process and Development 5 (1966) 322329.

[22]

S.G. Kandlikar, A general correlation for two-phase flow boiling heat transfer coefficient inside horizontal and vertical tubes, ASME Journal of Heat Transfer 112 (1990) 219-228.

[23]

S.G. Kandlikar, P. Balasubramanian, An extension of the flow boiling correlation to transition, laminar, and deep laminar flows in minichannels and microchannels, Heat Transfer Engineering 24 (2004) 3-17.

[24]

I. Pranoto, K.C. Leong, L.W. Jin, The role of graphite foam pore structure on saturated pool boiling enhancement, Applied Thermal Engineering 42 (2012) 163172.

[25]

Q. Yu, B.E. Thompson, A.G. Straatman, A unit cube-based model for heat transfer and fluid flow in porous carbon foam, ASME Journal of Heat Transfer, 128 (2006) 352-360.

SC

M AN U

TE D

EP

AC C

[26]

RI PT

[14]

K.C. Leong, H.Y. Li, Theoretical study of the effective thermal conductivity of graphite foam based on a unit cell model, International Journal of Heat and Mass Transfer, 54 (2011) 5491-5496.

21

ACCEPTED MANUSCRIPT Figure Captions Fig. 1 Schematic diagram of the flow boiling experimental facility. Fig. 2 Flow boiling channel (test section).

Fig. 4 Schematic of the test section and the evaporator gap.

RI PT

Fig. 3 SEM images of (a) “Pocofoam” graphite foam of 61% porosity and (b) “Kfoam” graphite foam of 72% porosity. Fig. 5 Schematic of the heat simulator and wall temperature thermocouple arrangements. Fig. 6 Wall temperature readings of different coolant mass fluxes at (a) d = 6 mm and (b) d = 4 mm; and (c) boiling curves of different mass fluxes at d = 6 mm and (d) d = 4 mm with “Pocofoam” graphite foam of 61% porosity.

SC

Fig. 7 Wall temperature readings of different coolant mass fluxes at (a) d = 6 mm and (b) d = 4 mm; and (c) boiling curves of different mass fluxes at d = 6 mm and (d) d = 4 mm with “Kfoam” graphite foam of 72% porosity.

M AN U

Fig. 8 Wall temperature readings and boiling curves for different evaporator gaps tested with “Pocofoam” graphite foam of 61% porosity for (a), (b) G = 100 kg/m2·s; (c), (d) G = 150 kg/m2·s. Fig. 9 Wall temperature readings and boiling curves for different evaporator gaps tested with “Kfoam” graphite foam of 72% porosity and FC-72 coolant for (a), (b) G = 100 kg/m2·s; (c), (d) G = 150 kg/m2·s.

TE D

Fig.10 Wall temperature readings and boiling curves of different graphite foam types with FC-72 coolant at G = 150 kg/m2·s for (a),(b): d = 6 mm; (c),(d): d = 4 mm. Fig. 11 Flow boiling images with various coolant mass fluxes at q" = 83.3 W/cm2 and d = 6 mm for graphite foams (a) “Pocofoam” of 61% porosity and (b) “Kfoam” of 72% porosity. Fig. 12 Flow boiling images with various coolant mass fluxes at q" = 83.3 W/cm2 and d = 4 mm for graphite foams (a) “Pocofoam” of 61% porosity and (b) “Kfoam” of 72% porosity.

EP

Fig. 13 Flow boiling images with various evaporator gaps at q" = 83.3 W/cm2 and G = 100 kg/m2·s graphite foams (a) “Pocofoam” of 61% porosity and (b) “Kfoam” of 72% porosity.

AC C

Fig. 14 Flow boiling images with various evaporator gaps at q" = 83.3 W/cm2 and G = 150 kg/m2·s graphite foams (a) “Pocofoam” of 61% porosity and (b) “Kfoam” of 72% porosity. Fig. 15 Typical “mushroom cloud” of large coalesced bubbles at high heat flux and low mass flux. Fig. 16 Typical thin vapor film formed in channel at low mass flux and high heat flux. Fig. 17 Flow boiling images with different graphite foam types with G = 150 kg/m2·s at q" = 58.3 W/cm2 for (a) d = 6 mm and (b) d = 4 mm. Fig. 18 Flow boiling images with different graphite foam types with G = 150 kg/m2·s at q" = 83.3 W/cm2 for (a) d = 6 mm and (b) d = 4 mm. Fig. 19 Bubble departure frequency determination from the captured images of bubble growth and departure.

22

ACCEPTED MANUSCRIPT Table 1 Selected thermophysical properties of graphite foams [18] Properties

“Pocofoam” 61%

“Kfoam” 72%

0.35

0.65

61

72

245

110

Average pore diameter, dp (mm) Porosity, ε (%)

plane), kGF (W/m⋅K) Density, ρGF (kg/m3)

RI PT

Bulk thermal conductivity (out of

480

SC

900

Table 2 Selected properties of FC-72 [19]

Boiling point [ºC]

Value

M AN U

Properties

56

Vapor pressure [kPa]

30.9

Liquid density, ρl [kg/m³]

1680

Latent heat of vaporization, hlv [kJ/kg] Thermal conductivity, kl [W/m·K)]

88

0.057 10

Dielectric strength [kV]

38

TE D

Surface tension, σ [mN/m]

EP

Table 3 Experimental parameters

AC C

Parameters

Evaporator gap (d)

Coolant mass flux (G)

Graphite foam type

Value d1

6 mm

d2

4 mm

d3

2 mm

G1

50 kg/m2·s

G2

100 kg/m2·s

G3

150 kg/m2·s

“Pocofoam” 61% porosity “Kfoam” 72% porosity

23

ACCEPTED MANUSCRIPT Table 4 Average wall temperature differences and average flow boiling heat transfer coefficient ratio between different coolant mass fluxes

“Pocofoam” 61% d = 4 mm

∆TW (G1 −G2 )

4.3

hfb(G1)

12.5

h fb (G2 ) / h fb (G1 )

1.16

∆TW (G1 −G3 )

9.2

hfb(G2)

14.5

h fb( G3 ) / h fb( G1 )

1.32

hfb(G3)

16.5

“Kfoam” 72% d = 4 mm

h fb (Gx ) / h fb (G1 )

∆TW (G1 −G2 )

4.2

hfb(G1)

11.7

h fb (G2 ) / h fb (G1 )

1.12

∆TW (G1 −G3 )

9.6

hfb(G2)

13.2

h fb( G3 ) / h fb( G1 )

1.34

hfb(G3)

15.7

∆TW (G1 −G2 )

4.1

hfb(G1)

10.8

h fb (G2 ) / h fb (G1 )

1.13

∆TW (G1 −G3 )

8.6

hfb(G2)

12.1

h fb( G3 ) / h fb( G1 )

1.21

hfb(G3)

13.1

M AN U

“Kfoam” 72% d = 6 mm

(°C) hfb(Gx) (kW/m2·K)

RI PT

“Pocofoam” 61% d = 6 mm

∆TW (G1 −Gx )

SC

Graphite foams and Gaps

∆TW (G1 −G2 )

3.6

hfb(G1)

10.2

h fb (G2 ) / h fb (G1 )

1.12

∆TW (G1 −G3 )

8.7

hfb(G2)

11.4

h fb( G3 ) / h fb( G1 )

1.22

hfb(G3)

12.4

TE D

Table 5 Average wall temperature differences and average flow boiling heat transfer coefficient ratio between different evaporator gaps. Graphite foams and Gaps

EP

“Pocofoam” 61% G = 100 kg/m2·s

AC C

“Pocofoam” 61% G = 150 kg/m2·s “Kfoam” 72% G = 100 kg/m2·s “Kfoam” 72% G = 150 kg/m2·s

∆TW ( d x −d1 )

(°C)

∆TW ( d2 −d1 )

3.4

hfbd1 / hfbd 2

1.06

∆TW ( d3 −d1 )

5.3

hfbd1 / hfbd 3

1.20

∆TW ( d2 −d1 )

3.9

hfbd1 / hfbd 2

1.05

∆TW ( d3 −d1 )

6.7

hfbd1 / hfbd 3

1.21

∆TW ( d2 −d1 )

3.8

hfbd1 / hfbd 2

1.06

∆TW ( d3 −d1 )

7.5

hfbd1 / hfbd 3

1.23

∆TW ( d2 −d1 )

4.1

hfbd1 / hfbd 2

1.05

∆TW ( d3 −d1 )

7.3

hfbd1 / hfbd 3

1.20

24

hfbd1 / hfbdx

SC

RI PT

ACCEPTED MANUSCRIPT

EP

TE D

M AN U

Fig. 1 Schematic diagram of the flow boiling experimental facility.

AC C

Fig. 2 Flow boiling channel (test section).

(b)

(a)

Fig. 3 SEM images of (a) “Pocofoam” graphite foam of 61% porosity and (b) “Kfoam” graphite foam of 72% porosity. 25

RI PT

ACCEPTED MANUSCRIPT

M AN U

SC

Fig. 4 Schematic of the test section and the evaporator gap.

AC C

EP

TE D

Fig. 5 Schematic of the heat simulator and wall temperature thermocouple arrangements.

(a)

(b)

26

RI PT

ACCEPTED MANUSCRIPT

(c)

(d)

(b)

AC C

EP

(a)

TE D

M AN U

SC

Fig. 6 Wall temperature readings of different coolant mass fluxes at (a) d = 6 mm and (b) d = 4 mm; and (c) boiling curves of different mass fluxes at d = 6 mm and (b) d = 4 mm with “Pocofoam” graphite foam of 61% porosity.

(c)

(d)

Fig. 7 Wall temperature readings of different coolant mass fluxes at (a) d = 6 mm and (b) d = 4 mm; and (c) boiling curves of different mass fluxes at d = 6 mm and (b) d = 4 mm with “Kfoam” graphite foam of 72% porosity.

27

RI PT

ACCEPTED MANUSCRIPT

(a)

M AN U

SC

(b)

(d)

TE D

(c)

AC C

EP

Fig. 8 Wall temperature readings and boiling curves for different evaporator gaps tested with “Pocofoam” graphite foam of 61% porosity for (a), (b) G = 100 kg/m2·s; (c), (d) G = 150 kg/m2·s.

(a)

(b)

28

RI PT

ACCEPTED MANUSCRIPT

(d)

(c)

(b)

AC C

EP

(a)

TE D

M AN U

SC

Fig. 9 Wall temperature readings and boiling curves for different evaporator gaps tested with “Kfoam” graphite foam of 72% porosity and FC-72 coolant for (a), (b) G = 100 kg/m2·s; (c), (d) G = 150 kg/m2·s.

(d) (c) Fig.10 Wall temperature readings and boiling curves of different graphite foam types with FC-72 coolant at G = 150 kg/m2·s for (a),(b): d = 6 mm; (c),(d): d = 4 mm.

29

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

Fig. 11 Flow boiling images with various coolant mass fluxes at q" = 83.3 W/cm2 and d = 6 mm for graphite foams (a) “Pocofoam” of 61% porosity and (b) “Kfoam” of 72% porosity.

Fig. 12 Flow boiling images with various coolant mass fluxes at q" = 83.3 W/cm2 and d = 4 mm for graphite foams (a) “Pocofoam” of 61% porosity and (b) “Kfoam” of 72% porosity.

30

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

Fig. 13 Flow boiling images with various evaporator gaps at q" = 83.3 W/cm2 and G = 100 kg/m2·s graphite foams (a) “Pocofoam” of 61% porosity and (b) “Kfoam” of 72% porosity.

Fig. 14 Flow boiling images with various evaporator gaps at q" = 83.3 W/cm2 and G = 150 kg/m2·s graphite foams (a) “Pocofoam” of 61% porosity and (b) “Kfoam” of 72% porosity.

31

ACCEPTED MANUSCRIPT

(a)

RI PT

(b)

M AN U

SC

Fig. 15 Typical “mushroom cloud” of large coalesced bubbles at high heat flux and low mass flux.

AC C

EP

TE D

Fig. 16 Typical thin vapor film formed in channel at low mass flux and high heat flux.

Fig. 17 Flow boiling images with different graphite foam types with G = 150 kg/m2·s at q" = 58.3 W/cm2 for (a) d = 6 mm and (b) d = 4 mm.

32

M AN U

SC

(a)

RI PT

ACCEPTED MANUSCRIPT

(b)

AC C

EP

TE D

Fig. 18 Flow boiling images with different graphite foam types with G = 150 kg/m2·s at q" = 83.3 W/cm2 for (a) d = 6 mm and (b) d = 4 mm.

33

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Fig. 19 Bubble departure frequency determination from the captured images of bubble growth and departure.

34

ACCEPTED MANUSCRIPT

Captions for the submitted videos of ATE-5551 by I. Pranoto and K.C. Leong

Vid.1_Flow Boiling_Poco61_6mmGap_G150_q58.3

Caption

Flow boiling from "Pocofoam" of 61% porosity with saturated FC-72 at d = 6 mm, G = 150 kg/m2·s and q" = 58.3 W/cm2.

RI PT

File name

Vid.2_Flow Boiling_Poco61_6mmGap_G150_q83.3

Caption

Flow boiling from "Pocofoam" of 61% porosity with saturated FC-72 at d = 6 mm, G = 150 kg/m2·s and q" = 83.3 W/cm2.

SC

File name

Vid.3_Flow Boiling_Poco61_2mmGap_G50_q83.3_mushroom

Caption

Flow boiling from "Pocofoam" of 61% porosity with saturated FC-72 at d = 2 mm, G =

M AN U

File name

AC C

EP

TE D

50 kg/m2·s and q" = 83.3 W/cm2.