Flow boiling visualization and heat transfer in metal-foam-filled mini tubes – Part I: Flow pattern map and experimental data

International Journal of Heat and Mass Transfer xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

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

Flow boiling visualization and heat transfer in metal-foam-filled mini tubes – Part I: Flow pattern map and experimental data Gholamreza Bamorovat Abadi, Chanhee Moon, Kyung Chun Kim ⇑ School of Mechanical Engineering, Pusan National University, San 30, Jangjeon-dong, Geumjeong-gu, Busan 609-735, Republic of Korea

a r t i c l e

i n f o

Article history: Received 2 January 2016 Received in revised form 9 March 2016 Accepted 11 March 2016 Available online xxxx Keywords: Metal foam Flow boiling Heat transfer Visualization

a b s t r a c t High-porosity open-cell metal foams are well known to enhance the heat transfer mechanism in rectangular or circular channels. Their high surface area to volume ratio makes them a great candidate for manufacturing high-performance small-scale heat exchangers. This two-part experimental study investigated the two-phase flow boiling inside a circular copper mini tube. In Part I, the flow pattern was visualized by high-speed imaging in glass tubes. Flow pattern maps, the heat transfer coefficient, and pressure drop are presented for mean vapor quality of 0.1–0.7, heat flux of 20–40 kW/m2, and mass flux of 400–700 kg/m2 s. The experiments were also performed without the metal foam in the mini tube for comparison to the original data. In this range of experimental conditions, the metal foam increased the heat transfer coefficient up to 3.2 times. Also, as expected, the metal foams adversely affected the pressure drop inside the tubes. Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction Enhancing heat transfer has always been a topic of interest for engineers. New technologies are enabling the manufacture of designs that were previously only conceptual, with which comes the need to modify the traditional components of different cycles. The importance of mini-scale heat exchangers has been realized recently after the introduction of new concepts such as mini organic Rankine cycles, mini refrigeration cycles, and electronic cooling devices. One promising method for enhancing the heat transfer mechanism in channels is inserting high-porosity metal foams, which have a high surface area to volume ratio that leads to higher heat transfer area in a very small volume. This allows for compact designs and increases the heat transfer coefficient [1]. The last decade has seen many single-phase and flow boiling experiments with channels that are fully or partially filled with metal foam. Calmidi and Mahajan [2] investigated single-phase forced convection in aluminum metal foams. They used air as the fluid and foams with high porosities of up to 0.97. Their model showed good agreement with experimental results. Mancin et al. [3] performed experiments on convective heat transfer in metal foams with different pore densities ranging from 5 to 40 pores per inch (PPI). They studied the pressure drop and heat transfer coefficient for heat fluxes between 25 and 40 kW/m2 and ⇑ Corresponding author. E-mail address: [email protected] (K.C. Kim).

compared their data with prediction methods, which showed good agreement. They later performed similar studies for copper metal foams [4]. Lu et al. [5] and Zhao et al. [6] thermally analyzed heat exchangers with metal-foam-filled channels. They used the Brinkmanextended Darcy momentum model and two-equation heat transfer and obtained the velocity and temperature distributions in metalfoam-filled pipes. They concluded that a metal-foam heat exchanger has better thermal performance than a finned tube heat exchanger. Kim et al. [7] presented one of the first experiments for convective heat transfer in metal-foam channels with air flow. Advances in computational fluid dynamics methods and software later enabled new types of simulations. Ranut et al. [8,9] used an accurate microtomography-based CFD method to simulate the heat transfer mechanism in metal foams with air flow. Their X-ray CT method is a powerful tool for capturing the shape of the metal foam and creating an acceptable mesh. A more interesting aspect of the heat transfer mechanism is the phase changes inside the metal-foam-filled channels. There is competition between nucleate boiling and convective boiling inside the channels on the wall or struts, and different flow patterns occur with the introduction of metal foam. The different pressure field completely changes the heat transfer coefficient compared to empty tubes. Recently, the number of articles regarding the flow boiling in metal foam channels has increased. Diani et al. [10] and Mancin et al. [11] investigated the phase change phenomena of R134a, R1234yf, and R1234ze(E) in a

http://dx.doi.org/10.1016/j.ijheatmasstransfer.2016.03.043 0017-9310/Ó 2016 Elsevier Ltd. All rights reserved.

Please cite this article in press as: G. Bamorovat Abadi et al., Flow boiling visualization and heat transfer in metal-foam-filled mini tubes – Part I: Flow pattern map and experimental data, Int. J. Heat Mass Transfer (2016), http://dx.doi.org/10.1016/j.ijheatmasstransfer.2016.03.043

2

G. Bamorovat Abadi et al. / International Journal of Heat and Mass Transfer xxx (2016) xxx–xxx

Nomenclature A Atotal Atube cp Dh G g h hfg k L M _ m P Pr Pr q Q Re T x

area, m2 total surface area of the metal-foam-filled tube surface area of the copper tube specific heat, J/kg K hydraulic diameter, m mass flux, kg/m2 s gravitational acceleration, m/s2 enthalpy, kJ/kg latent heat of evaporation, J/kg thermal conductivity, W/mK distance, m molecular mass, kg/kmol mass flow, kg/s pressure, bar Prandtl number reduced pressure heat flux, kW/m2 heat, W reynolds number temperature, K quality

channel filled with 5-PPI metal foam. The heat transfer was enhanced by up to 4.8 times in their experiments at low mass flux, low heat flux, and high vapor quality. In all cases, the pressure drop increased with the vapor quality or mass flux. Zhu et al. [12] performed nucleate pool boiling experiments on mixtures of R113 refrigerant and VG68 oil. They used copper foam with 10 and 20 PPI and porosity as high as 0.98. The metal foam cover increases the heat transfer coefficient up to 160% compared to a flat plate, and the addition of oil decreases the heat transfer coefficient by up to 15%. Hu et al. [13] investigated the effect of tube diameter and compared the pressure drop data for 7.9, 13.8, and 26.0-mm tubes. When the tube diameters decreased from 13.8 mm to 7.9 mm with the same PPI, the pressure drop decreased due to the incomplete cells in the metal foam. Zhu et al. [14,15] visualized the flow boiling of R410A refrigerant inside 7.9-mm glass tubes and presented a flow map based on their data. The enhancement of the heat transfer coefficient by the foam was 50% greater at low mass fluxes than at higher mass fluxes. Slug flow, plug flow, and annular flow were observed in the experiments, and the metal foam promoted the formation of annular flow. This effect was stronger with higher PPI. The study of flow boiling inside metal-foam-filled tubes is just beginning, and the phenomenon is not fully understood yet. The experimental conditions and thus the validity of the correlations have been limited. Thus, more experimental data are needed to understand the heat transfer mechanism inside such tubes. Therefore, the focus of this study is the flow boiling of refrigerants in small tubes at medium mass fluxes. Part I of this experimental study looks at the flow boiling of R245fa refrigerant inside a 4-mm copper tube filled with 20-PPI and 30-PPI copper metal foam. The flow patterns were visualized inside identical glass tubes by high-speed imaging. The mass flux ranges from 400 to 700 kg/m2 s, and the maximum heat flux is 40 kW/m2. The mean vapor quality ranges from 0.1 to 0.7, and the experiments were performed at the saturation temperature of about 62.75 °C. Experiments were performed under the same experimental conditions using tubes without metal foam for comparison and to understand the effect of the foams. In Part II [16], the experimental data are compared to previous correlations, and new predictive

Greek symbols heat transfer coefficient, W/m2 K b pressure drop factor d liquid film thickness, m e void fraction e0 porosity l viscosity, Pa s q density, kg/m3 r surface tension, N/m

a

Subscript Cb F H l MF nb pre sat tp v w

convective boiling friction heater liquid metal foam nucleate boiling preheater saturation two phase vapor wall

correlations are proposed for the heat transfer coefficient and pressure drop. 2. Test rig 2.1. Cycle A closed loop was prepared for the experiment, as shown in Fig. 1. The main part of the loop is a refrigerant cycle where a 170-W gear pump is used to deliver refrigerant to the test section. A positive-displacement flow meter is installed after the pump. The volume of flow entering the test section is controlled by adjusting the pump frequency. The pump speed is adjustable from 0 to 3600 RPM corresponding to 0–4 LPM. Before the test section is a preheating section. The inlet quality of the refrigerant entering the test section is controlled at the preheater, which is a long electrically heated copper tube (through the Joule effect). The heater and preheater are electrically isolated from

Fig. 1. Diagram of experimental loop.

Please cite this article in press as: G. Bamorovat Abadi et al., Flow boiling visualization and heat transfer in metal-foam-filled mini tubes – Part I: Flow pattern map and experimental data, Int. J. Heat Mass Transfer (2016), http://dx.doi.org/10.1016/j.ijheatmasstransfer.2016.03.043

3

G. Bamorovat Abadi et al. / International Journal of Heat and Mass Transfer xxx (2016) xxx–xxx

the rest of the setup. The preheater power is controlled by adjusting the applied voltage. The energy balance of the preheater determines the flow conditions at the test section inlet. After the test section, the flow passes through a plate heat exchanger that is connected to an air-cooled chiller and works as a condenser. The temperature of the flow leaving the condenser and entering the reservoir tank is adjustable by thermocouples on the chiller. This temperature is used to adjust the preheater inlet temperature. There is a K-type thermocouple and a pressure transducer before and after each component to measure the flow conditions at each point. A pressure difference transducer is installed on the test section to measure the pressure drop of the flow passing through the test section. 2.2. Test section The test section is a horizontal copper tube with an inner diameter of 4 mm and outer diameter of 6 mm. This is followed by a quartz glass tube with the same diameter, as shown in Fig. 2. The copper tube length is 500 mm, which is long enough to make sure that the entrance effects are negligible and that the flow is fully developed. Four thermocouples are installed on the outer surface of the copper tube at every 50 mm, with a total of 40 thermocouples. A tape heater is tightly installed on the copper tube, and the whole system is thermally insulated by wool glass insulation around the tube. The heat flux is controlled by the voltage applied to the tape heater. The glass tube is shorter to reduce heat loss. A Photron high-speed camera is used to visualize the flow through the quartz glass, and images were taken with a shutter speed of 15,000 fps. Halogen lamps on the side opposite to the high-speed camera were used to illuminate the test section. Copper metal foams were prepared with 90% porosity and 20 and 30 PPI and then cut into cylinders by electrical discharge machining. Fig. 3 shows the 20 and 30-PPI metal foam sheets. Table 1 lists the geometrical properties of the metal foams. The heat transfer area is increased by 6.18 and 8.76 for 20 and 30 PPI metal foam respectively, compared to the empty tube. The metal foam cylinders are dipped in a solder pool so as only the tips of the ligaments are covered with solder material. The cylinders are then inserted in the copper tube and the tube is heated to connect the metal foam to the copper tube inner surface. The solder material is 60% tin and 40% lead with melting point of around 188 °C.

Fig. 3. Metal foam sheets.

Table 1 Geometrical properties of the metal foams. Pore density (PPI)

Porosity

Surface-area-to-volume ratio (m2/m3)

Permeability (m2)

Atot/ Atube

20 30

0.9 0.9

2310 3520

1.99e
8 9.41e
9

6.18 8.76

R245fa refrigerant was used as the working fluid. We have previously used this refrigerant for flow boiling experiments and organic Rankine cycle applications [17–19]. 2.3. Uncertainty analysis The overall uncertainty in measuring the length or diameter of tubes was estimated to be in the range of ±0.1 mm. The K-type

Fig. 2. Test section.

Please cite this article in press as: G. Bamorovat Abadi et al., Flow boiling visualization and heat transfer in metal-foam-filled mini tubes – Part I: Flow pattern map and experimental data, Int. J. Heat Mass Transfer (2016), http://dx.doi.org/10.1016/j.ijheatmasstransfer.2016.03.043

4

G. Bamorovat Abadi et al. / International Journal of Heat and Mass Transfer xxx (2016) xxx–xxx

thermocouples have an accuracy of ±0.4 °C. The uncertainty in measuring the temperature is ±1.5 °C. The flowmeter has an accuracy of ±1% full scale (0.2–10 liter per minute), and the pressure measurement was achieved using pressure transducers with accuracy of ±0.01 bar. For the pressure drop, a Yokogawa pressure difference transmitter was used, which is accurate to ±0.04% full scale (1–100 kPa). The maximum uncertainty of the pressure drop is estimated to be ±2%. The heater and preheater both use voltage controls that are accurate to ±50 mV. Table 2 summarizes the uncertainties in this experiment. A method suggested by Moffat [20] was used for calculating the uncertainty of the heat transfer coefficient. The REFPROP 9.1 library was used to determine the properties of R245fa.

DPf ¼ DPtotal
DPacceleration

ð3Þ

where DPacceleration is calculated by [15]:

"

DPacceleration

ð1
xÞ2 ¼G þ eqv ð1
eÞql 2

x2

#

"

ð1
xÞ2
G þ eqv ð1
eÞql 2

out

x2

# ð4Þ in

The void fraction is calculated by the Steiner version of the Rouhani–Axelsson drift flux model for cross-sectional void fraction [21]:

e¼

#
1   x 1
x 1:18ð1
xÞ½g rðql
qv Þ0:25 þ ð1þ0:12ð1
xÞÞ þ qv qv ql Gq0:5 l x

"

ð5Þ

2.4. Data reduction 2.5. Single-phase validation The experimental conditions are listed in Table 3. The local heat transfer coefficient is calculated using Eq. (1) [18]:

a¼

qH T w
T sat

ð1Þ

where a is the local heat transfer coefficient, qH is the heat flux from the heater, Tw is the mean inside wall temperature, and Tsat is the saturation temperature calculated using the mean pressure inside the test section [15]. The inner area of the copper tube is used to calculate qH. The inside wall temperature is calculated using the outside wall temperatures and a one-direction heat conduction equation. The quality of the flow entering the test section is calculated using an energy balance over the preheater and by Eq. (2) [18]:


 1 Q pre x¼
ðhsat
hin Þ _ hfg m

ð2Þ

The latent heat is calculated based on the temperature and pressure at the outlet of the preheater, and its variation along the preheater is neglected. hin is the enthalpy of the fluid at the inlet of the preheater, and hsat is that at the outlet. The pressure drop is measured between the inlet and the outlet of the copper tube. The total pressure drop is corrected by subtracting the acceleration pressure drop

Since the heat flux is calculated based on the given electric energy to the heater, single-phase heat transfer coefficient calculation is necessary to determine the effect of heat loss on the experiment. An energy balance on the heater is used to compare the energy given to the heater to the energy received by the fluid through enthalpy change using Eqs. (6) and (7). The inlet and outlet enthalpy values are calculated based on the measured temperature and pressure values at inlet and outlet of the test section.

_ out
hin Þ Q ¼ mðh

ð6Þ

Q ¼ VI

ð7Þ

The energy balance shows that heat loss at the heater is around 5%, meaning that using Eq. (7) as the input heat is justified. To validate the test section, the single-phase overall heat transfer coefficient is also calculated based on single-phase experiments. The experimental Nusselt number is calculated by:

Nu ¼

Parameter

ð8Þ

k

where a is the average single-phase heat transfer coefficient calculated by Eq. (1). The average value of the thermal conductivity k is used in Eq. (8). The experimental Nusselt number is then compared to the well-known Gnielinski equation [22]:

Nu ¼ Table 2 Experimental uncertainties.

aD

ðf =8ÞðRe
1000ÞPr 0:5

1 þ 12:7ðf =8Þ ðPr0:67
1Þ

ð9Þ

Uncertainty 2

Mass flux (kg/m s) Temperature (K) Pressure (bar) Pressure drop (kPa) Length, diameter (m) Vapor quality Heater power (W) Preheater power (W) Heat transfer coefficient (W/m2 K)

±5% ±1.5 ±2% ±2% ±0.2 ±1% ±3% ±3% ±8.5–16.12%

Table 3 Experimental conditions. Parameter

Range

Mass flux (kg/m2 s) Heat flux (kW/m2) Inlet quality Saturation temperature (°C) Tube inner diameter (m) Metal foam pore per inch value (PPI) Metal foam porosity

400–700 20–40 0.1–0.7 62.75 4 mm 20, 30 0.9 Fig. 4. Comparison of experimental Nusselt number to Gnielinski equation.

Please cite this article in press as: G. Bamorovat Abadi et al., Flow boiling visualization and heat transfer in metal-foam-filled mini tubes – Part I: Flow pattern map and experimental data, Int. J. Heat Mass Transfer (2016), http://dx.doi.org/10.1016/j.ijheatmasstransfer.2016.03.043

G. Bamorovat Abadi et al. / International Journal of Heat and Mass Transfer xxx (2016) xxx–xxx

f ¼

1 ð0:79LnðReÞ
1:64Þ2

ð10Þ

A comparison between the experimental Nusselt number and the predicted values is shown in Fig. 4. 100% of the experimental data is captured within ±10% of the values predicted by the Gnielinski equation. Single-phase experiments were also performed to validate the test section and the methodology for measuring the pressure drop. The distance between the measuring points is shown in Fig. 2, which also shows the absolute pressure transducer installed at the inlet and outlet of the test section. The experimental singlephase pressure drop was compared to the predicted single-phase pressure drop using the Blasius equation:

5

The vapor quality change along the test section ranges from 4% to 15% depending on the heat and mass flux. Fig. 7 presents the flow pattern for pure R245fa in an empty tube at a mass flux of 400 kg/m2 s and heat flux of 30 kW/m2.

2

DP ¼

f ¼

2fLG ql D

0:079 Re0:25

ð11Þ

ð12Þ

Fig. 5 compares the experimental single-phase pressure drop and the predicted pressure drop values. 100% of the experimental data falls within ±10% of the predicted values.

3. Results and discussion

Fig. 6. Vapor quality variation along the test section for different mass and heat fluxes.

3.1. Visualization and flow pattern map Flow patterns have been visualized in a quartz glass mini tube filled with metal foam. Before the visualization section, the glass tube and the preceding copper tube are filled with metal foam, and the length of the tube assures that the flow is fully developed. Flow patterns were visualized for an empty glass tube under the same operating conditions. It is somewhat difficult to determine the flow patterns in metal-foam-filled tubes due to the metal foam blocking the view and changing the flow behavior. However, the focus is on determining the intermittent/annular transition line, so the pre-annular flow is considered intermittent. The same flow pattern terminology as in empty tubes is used for metal-foamfilled tubes and then the differences between them are determined through visualization. The reported quality in the following figures is the mean vapor quality. The vapor quality variation along the test section is reported in Fig. 6 for different mass and heat fluxes.

Fig. 5. Comparison of experimental single-phase pressure drop to Blasius equation.

Fig. 7. Flow patterns in flow boiling of pure R245fa in an empty tube at mass flux of 400 kg/m2 s and heat flux of 30 kW/m2, (a) quality of 0.1, intermittent (b) quality of 0.3, intermittent (c) quality of 0.4, annular.

Fig. 8. Flow patterns in flow boiling of pure R245fa in an empty tube at mass flux of 700 kg/m2 s and heat flux of 30 kW/m2, (a) quality of 0.1, intermittent (b) quality of 0.2, annular.

Please cite this article in press as: G. Bamorovat Abadi et al., Flow boiling visualization and heat transfer in metal-foam-filled mini tubes – Part I: Flow pattern map and experimental data, Int. J. Heat Mass Transfer (2016), http://dx.doi.org/10.1016/j.ijheatmasstransfer.2016.03.043

6

G. Bamorovat Abadi et al. / International Journal of Heat and Mass Transfer xxx (2016) xxx–xxx

Fig. 7 (a) shows the intermittent flow for when the quality is 0.1. When increasing the quality to 0.3, the flow is still intermittent, while the flow becomes annular at a quality of 0.4. Fig. 8 shows the flow pattern at a mass flux of 700 kW/m2 and heat flux of 30 kW/m2. Fig. 8 (a) shows that the flow is intermittent at a quality of 0.1 but it becomes annular when increasing the quality to 0.2, as shown in Fig. 8(b). A flow pattern map is presented based on the result of visualization in an empty tube is presented in Fig. 9 for pure R245fa at a heat flux of 30 kW/m2. The transition line from intermittent flow to annular flow is presented. At low mass flux, the flow is intermittent at quality of up to 0.35, while at higher mass fluxes, the annular flow starts at lower qualities. At a mass flow of 700 kg/m2 s, the annular flow starts at a quality of around 0.15. The heat flux was found to have small effect on the transition line, therefor the images and flow maps are presented for one value of the heat flux. The flow patterns of boiling of R245fa in a metal-foam-filled tube were investigated under the same experimental conditions as those for the empty tube. 20 and 30-PPI metal foam were chosen for this experiment. Fig. 10 shows the intermittent flow with the 20-PPI foam at a mass flux of 400 kg/m2 s and quality of 0.1. The flow direction is from right to left. A slug of vapor in the flow moves along the flow direction, as indicated by the red curve in Fig. 10(a) and (b). The slug flow in the metal-foam-filled tube is different from that in the empty tube. Fig. 7(a) shows a big slug of vapor in slug flow in an empty tube. In contrast, the vapor slug in Fig. 10 is broken into small slugs as it passes through the pores of the metal foam. Any such flow that is not annular is considered to be intermittent. Fig. 11 shows flow patterns for the same 20-PPI metal-foamfilled tube at a mass flux of 400 kg/m2 s and heat flux of 30

Fig. 9. Flow pattern map for pure R245fa refrigerant in empty mini tube at a heat flux of 30 kW/m2 showing intermittent/annular transition line.

Fig. 11. Flow patterns in flow boiling of pure R245fa in 20-PPI metal-foam-filled tube at a mass flux of 400 kg/m2 s and heat flux of 30 kW/m2, (a) quality of 0.3, annular (b) quality of 0.5, annular.

kW/m2. In Fig. 11(a) the quality is 0.3. When increasing the quality from 0.1 to 0.3 (Figs. 7 and 11(a)), the flow pattern changes from intermittent to annular. The annular flow is distinguishable around the metal foam, but the core vapor is still disturbed and there are sometimes bubbles in the core vapor. In Fig. 11(b), the quality is increased to 0.5. The flow is still annular, and at high quality, it looks more like the annular flow in the empty tube with a distinguishable vapor core and liquid annular flow. The flow is mostly annular at higher mass velocities, as shown in Fig. 12. Fig. 12(a) shows the flow pattern of R245fa in 20-PPI metal-foam-filled tube at a mass flux of 700 kg/m2 s, heat flux of 30 kW/m2, and quality of 0.1. Under these conditions, the flow is annular. Fig. 12(b) shows the flow pattern for the same conditions except for a quality of 0.3. A flow map for the metal-foam-filled tubes is presented in Fig. 13 based on the flow boiling visualization data for pure R245fa in metal-foam-filled mini tube at a heat flux of 30 kW/m2. The transition line from intermittent flow to annular flow is presented. The flow map is very similar to that for an empty tube (Fig. 9), but the flow is intermittent at the low mass flux of 400 kg/m2 s up to a quality of 0.25. At higher mass fluxes, the annular flow starts at much lower qualities. At a mass flow of 700 kg/m2 s, the annular flow starts at quality of around 0.05. Comparing Figs. 13 and 9 suggests that the annular flow starts at lower qualities compared to the flow inside an empty tube at the same mass velocity. To explore the effect of PPI, experiments were repeated using 30-PPI copper metal foam with the same porosity as the 20-PPI metal foam. Fig. 14 shows the flow pattern for a mass flux of

Fig. 10. Flow patterns in flow boiling of pure R245fa in 20 PPI metal-foam-filled tube at a mass flux of 400 kg/m2 s and heat flux of 30 kW/m2, quality 0.1, intermittent, (a and b) movement of slug of flow along the flow direction (from right to left).

Please cite this article in press as: G. Bamorovat Abadi et al., Flow boiling visualization and heat transfer in metal-foam-filled mini tubes – Part I: Flow pattern map and experimental data, Int. J. Heat Mass Transfer (2016), http://dx.doi.org/10.1016/j.ijheatmasstransfer.2016.03.043

G. Bamorovat Abadi et al. / International Journal of Heat and Mass Transfer xxx (2016) xxx–xxx

7

No study has examined this range of mass flux, particularly for mini tubes as small as those considered. Hence, it is not possible to compare the flow pattern maps with other maps, which mostly deal with low mass fluxes and much bigger tubes. Instead, Fig. 17 compares the flow pattern map of R245fa in a 20-PPI metal-foam-filled tube with 30-PPI metal foam and an empty tube. With the metal foam, annular flow occurs at lower mass flux compared to the empty tube at the same flow vapor quality. 3.2. Heat transfer coefficient

Fig. 12. Flow patterns in flow boiling of pure R245fa in 20-PPI metal-foam-filled tube at a mass flux of 700 kg/m2 s and heat flux of 30 kW/m2, (a) quality of 0.1, annular (b) quality of 0.3, annular.

Fig. 13. Flow pattern map for pure R245fa refrigerant in 20 PPI metal-foam-filled mini tube at a heat flux of 30 kW/m2 showing intermittent/annular transition line.

400 kg/m2 s and heat flux of 30 kW/m2 s. Fig. 14(a) and (b) shows that the flow is intermittent at a quality of 0.1. Very similar to the case of 20-PPI metal foam in Fig. 10, movement of a vapor slug in the liquid is shown from right to left in Fig. 14(a) and (b). When increasing the quality, the flow pattern becomes annular, as shown in Fig. 14(c). Fig. 15 shows that the mass flux is increased from 400 kg/m2 s in Fig. 14 to 700 kg/m2 s. Fig. 15(a) shows the annular flow for a quality of 0.1, and Fig. 15(b) shows that for when the quality is 0.3. It is obvious at this point that the PPI value has little effect on the flow pattern. Comparing Figs. 10–12 and 14 and 15 makes it clearer that under the same experimental conditions, the flow pattern is almost the same in mini tubes with either 20-PPI or 30-PPI metal foam. The flow pattern map for 30-PPI metal foam is presented in Fig. 16. Compared to Fig. 13, the annular flow occurs at lower qualities, so the transition line for the 30-PPI metal foam in Fig. 16 is pushed to the left.

Heat transfer coefficient data are presented for the foam-filled and empty tubes. Fig. 18 compares the two-phase heat transfer coefficients of the filled tubes and the empty ones versus flow vapor quality at a heat flux of around 20 kW/m2. The figure also shows the effect of mass flux on the heat transfer coefficient by comparing the heat transfer coefficient at mass fluxes of 400– 700 kg/m2 s. In this range of mass and heat flux, the empty tube’s heat transfer coefficient has an increasing trend with increasing quality and mass flux, and the flow is mainly annular. Therefore, for this experimental range, convective flow boiling is dominant over nucleate boiling. The same trend is observed for the metalfoam-filled tubes, but the heat transfer coefficient is improved compared to the empty tube. To investigate the effect of heat flux on the heat transfer coefficient data, experiments were repeated at a higher heat flux of around 40 kW/m2. It was implied in the last section that the heat flux does not have a strong effect on the flow pattern, which is also the case for the heat transfer coefficient. The heat flux does not have a strong effect on the heat transfer coefficient, while the mass flux affects the heat transfer coefficient strongly. This is shown in Fig. 19, which presents the heat transfer coefficient data for all the tubes at a heat flux of 40 kW/m2 for flow boiling with the same diameter. To see the effect of increasing heat flux, Fig. 20 shows the heat transfer coefficient data for all the cases at a mass flux of 400 kg/m2 s. The figure compares the heat transfer coefficient data for two heat fluxes of around 20 and 40 kW/m2. In Fig. 21, the same comparison is done for a mass flux of 700 kg/m2 s. Comparing Figs. 18–21 indicates that increasing the PPI of the metal foam increases the heat transfer coefficient in the same conditions, but it does not change the trends or the flow pattern. The heat transfer coefficient is always higher for 30 PPI than for 20 PPI in the same conditions. Comparing Figs. 20 and 21, the heat flux does not have a strong effect on the heat transfer coefficient which is the case for low heat flux and high mass flux conditions where the convective flow boiling is the dominant contributor and the nucleate boiling is rather suppressed. The heat transfer coefficient is still increased by increasing the heat flux but this increase is negligible compared to the increase of the heat transfer coefficient with mass flux. An improvement factor is defined to show the effect of the metal foam [15]:

IF ¼

aMF aEmpty

ð13Þ

aMF is the heat transfer coefficient for a metal-foam-filled tube and aEmpty is that for a smooth empty tube. Along with the pressure drop factor b, this improvement factor is an important design factor that could determine whether using metal foam would be economical. Fig. 22 presents the improvement factor for the 20-PPI metalfoam-filled tube. As seen before, the heat flux does not have a strong effect, but the mass flux still plays an important role. In this case, the effect of mass flux is inversed in that increasing the mass flux decreases the improvement factor. Increasing the quality also decreases the improvement factor. Both trends can be explained by comparing Fig. 22 to the previous figures showing the heat transfer

Please cite this article in press as: G. Bamorovat Abadi et al., Flow boiling visualization and heat transfer in metal-foam-filled mini tubes – Part I: Flow pattern map and experimental data, Int. J. Heat Mass Transfer (2016), http://dx.doi.org/10.1016/j.ijheatmasstransfer.2016.03.043

8

G. Bamorovat Abadi et al. / International Journal of Heat and Mass Transfer xxx (2016) xxx–xxx

Fig. 14. Flow patterns in flow boiling of pure R245fa in 30-PPI metal-foam-filled tube at a mass flux of 400 kg/m2 s and heat flux of 30 kW/m2, (a and b) movement of slug of flow along the flow direction (from right to left), quality 0.1, intermittent, (c) quality 0.3, annular.

Fig. 16. Flow pattern map for pure R245fa refrigerant in 30-PPI metal-foam-filled mini tube at a heat flux of 30 kW/m2 showing intermittent/annular transition line.

improvement factor is always greater than that of the 20-PPI metal-foam-filled tube. Fig. 15. Flow patterns in flow boiling of pure R245fa in 30-PPI metal-foam-filled tube at a mass flux of 700 kg/m2 s and heat flux of 30 kW/m2, (a) quality of 0.1, annular (b) quality of 0.3, annular.

coefficient. The slope of the heat transfer coefficient curve versus quality for the empty tube is always greater than that for the metal-foam-filled tubes. Hence, with increasing quality, the heat transfer coefficient for the empty tube increases more than that of the metal-foam-filled tube, and the improvement factor decreases when increasing quality. Increasing the mass flux also increases the heat transfer coefficient of the empty tube, more so than for the metal-foam-filled tubes. Therefore, the improvement factor decreases when increasing mass flux. Fig. 23 shows similar trends and a comparison for 30-PPI metal-foam-filled channels. Compared to Fig. 22, the

3.3. Pressure drop The pressure drop was measured by directly measuring the pressure difference between the inlet and outlet of the copper tube, which was simultaneously used for heat transfer coefficient measurement. The pressure difference transducer used in this study is accurate to 0.04 kPa and is installed in a higher location than the horizontal tube test section. Its digital output is monitored by the LabView program used in the heat transfer measurement. The valve after the test section was used to control the absolute pressure at the inlet of the test section. Changing the mass flow rate tends to change the pressure inside the test, so controlling this valve ensures that the pressure difference measurement is always performed at the same absolute pressure at the inlet of the test section.

Please cite this article in press as: G. Bamorovat Abadi et al., Flow boiling visualization and heat transfer in metal-foam-filled mini tubes – Part I: Flow pattern map and experimental data, Int. J. Heat Mass Transfer (2016), http://dx.doi.org/10.1016/j.ijheatmasstransfer.2016.03.043

G. Bamorovat Abadi et al. / International Journal of Heat and Mass Transfer xxx (2016) xxx–xxx

9

Fig. 17. Intermittent/annular transition line for 20-PPI and 30-PPI metal-foamfilled tube compared to empty tube.

Fig. 20. Heat transfer coefficient data for metal-foam-filled and empty tubes at a mass flux of 400 kg/m2 s.

Fig. 18. Heat transfer coefficient data for metal-foam-filled and empty tubes at a heat flux of 20 kW m2.

Fig. 21. Heat transfer coefficient data for metal-foam-filled and empty tubes at a mass flux of 700 kg/m2 s.

Fig. 19. Heat transfer coefficient data for metal-foam-filled and empty tubes at a heat flux of 40 kW m2.

Fig. 22. Improvement factor (IF) for 20-PPI metal-foam-filled tube.

The measured pressure difference consists of the two-phase frictional pressure drop and the momentum pressure drop caused by the flow acceleration (expansion). Eqs. (3) and (4) are used to

determine the frictional pressure drop. The inlet and outlet quality values were separately calculated and used in Eq. (4). The frictional two-phase pressure drop in mini tubes is expected to increase

Please cite this article in press as: G. Bamorovat Abadi et al., Flow boiling visualization and heat transfer in metal-foam-filled mini tubes – Part I: Flow pattern map and experimental data, Int. J. Heat Mass Transfer (2016), http://dx.doi.org/10.1016/j.ijheatmasstransfer.2016.03.043

10

G. Bamorovat Abadi et al. / International Journal of Heat and Mass Transfer xxx (2016) xxx–xxx

Fig. 25. Experimental frictional two-phase pressure drop in metal-foam-filled tube – 20 and 30 PPI.

Fig. 23. Improvement factor (IF) for 30-PPI metal-foam-filled tube.

when increasing the quality or increasing the mass flux. These expectations were validated experimentally. Fig. 24 presents the experimental pressure drop versus vapor quality for the empty tube at two different mass fluxes and saturation temperature of 62.75 °C. The pressure drop strongly depends on the mass velocity and the vapor quality. It would also strongly depend on the tube diameter and the saturation temperature, which were not studied here. The maximum saturation temperature drop due to the pressure drop occurs at the maximum mass flux and quality and equals 1.35 °C for the empty tube. The pressure drop in the metal-foam-filled tubes is expected to increase hugely because of the metal foam blocking the flow passage. The increase in the pressure drop in metal-foam-filled tubes compared to empty tubes depends on the geometry of the metal foam, porosity, PPI, and tube diameter. The pressure drop ratio b is:

b¼

DPMF DPEmpty

ð14Þ

In small tubes, the small diameter means that the metal foam inside the tube is not perfect: it might have deficiencies formed during machining, broken ligaments, imperfect pores, and so on. This suggests that the pressure drop ratio would decrease with the diameter [13]. Fig. 25 presents the two-phase pressure drop

Fig. 24. Experimental frictional two-phase pressure drop in empty tube.

Fig. 26. Pressure drop ratio b for 20 and 30-PPI metal-foam-filled tube.

in the 20-PPI metal-foam-filled tube at two mass fluxes of 400 and 700 kg/m2 s. Compared to the empty tube, the pressure drop value is much bigger, but the trend is still the same. As expected, the pressure drop increases with the vapor quality and mass velocity. The pressure drop ratio b for 20-PPI metal foam is given in Fig. 26. The pressure drop ratio increases with the mass flux and is also a function of the quality and metal foam PPI. The pressure drop is expected to increase with the PPI. Fig. 25 also presents the pressure drop data for 30-PPI metal foam. Even though both metal foams have the same porosity, increasing the PPI increases the pressure drop due to the smaller pores. The pressure drop ratio for the 30-PPI metal foam is also given in Fig. 26. The pressure ratio increases with the mass flux, more so than the case with 20-PPI metal foam. Comparing Figs. 25 and 26, it is concluded that the metal foam in the tubes greatly increases the pressure drop. Moreover, the metal foam with higher PPI has a higher pressure drop, and the pressure drop ratio is a function of the quality, mass flux, and metal foam PPI. The maximum saturation temperature drop due to the pressure drop occurs at the maximum mass flux and quality and equals 36.997 °C for the 20 PPI and 45.188 °C for 30 PPI metal foam.

Please cite this article in press as: G. Bamorovat Abadi et al., Flow boiling visualization and heat transfer in metal-foam-filled mini tubes – Part I: Flow pattern map and experimental data, Int. J. Heat Mass Transfer (2016), http://dx.doi.org/10.1016/j.ijheatmasstransfer.2016.03.043

G. Bamorovat Abadi et al. / International Journal of Heat and Mass Transfer xxx (2016) xxx–xxx

4. Conclusion Flow boiling experiments have been performed using mini tubes. Part I of this two-part study presented the flow pattern maps based on flow boiling visualization experiments and highspeed images. Heat transfer coefficient data were calculated, and the pressure drop in the mini tube was measured. Data were compared for metal-foam-filled tubes and empty tubes. The flow pattern maps show either annular flow at high vapor qualities or intermittent slug flow at lower vapor qualities. The metal foam does not change the flow pattern, although it might change the conventional definition of each flow pattern. In each case, the differences in heat transfer characteristics between the metal-foam-filled tube and the empty tube were pointed out. Two parameters were defined for comparing the data. The improvement factor is the ratio of the heat transfer in the metalfoam-filled tube to that in the empty tube. The analysis shows that the improvement factor is as high as 3.2 in the conditions studied. The increase in the heat transfer area plays a definitive role in increasing the heat transfer coefficient but the role of the shift of the flow pattern from intermittent to annular flow, at the same mass flux and vapor quality, due to insertion of the metal foam is not negligible either. Another parameter is the pressure drop ratio, which is the ratio of the pressure drop in the metal-foamfilled tube to that in the empty tube. A pressure drop ratio of up to 22 was observed in the foam-filled tubes. In Part II, predictive correlations are proposed. Acknowledgments This study was supported by the Energy Efficiency & Resources Core Technology Program of the Korea Institute of Energy Technology Evaluation and Planning (KETEP) granted financial resources from the Ministry of Trade, Industry & Energy, Republic of Korea (Nos. 20142010102800, 20132020000390). This work was also supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIP) through GCRCSOP (No. 2011-0030013). References [1] L. Tadrist, M. Miscevic, O. Rahli, F. Topin, About the use of fibrous materials in compact heat exchangers, Exp. Therm. Fluid Sci. 28 (2004) 193–199. [2] V.V. Calmidi, R.L. Mahajan, Forced convection in high porosity metal foams, ASME J. Heat Transfer 122 (3) (2000) 557–565, http://dx.doi.org/10.1115/ 1.1287793.

11

[3] S. Mancin, C. Zilio, L. Rossetto, A. Cavallini, Heat transfer performance of aluminum foams, J. ASME J. Heat Transfer 133 (6) (2011), http://dx.doi.org/ 10.1115/1.4003451. [4] S. Mancin, C. Zilio, A. Diani, L. Rossetto, Experimental air heat transfer and pressure drop through copper foams, Exp. Therm. Fluid Sci. 36 (2012) 224– 232. [5] W. Lu, C.Y. Zhao, S.A. Tassou, Thermal analysis on metal-foam filled heat exchangers. Part I: metal-foam filled pipes, Int. J. Heat Mass Transfer 49 (2006) 2751–2761. [6] C.Y. Zhao, W. Lu, S.A. Tassou, Thermal analysis on metal-foam filled heat exchangers. Part II: tube heat exchangers, Int. J. Heat Mass Transfer 49 (2006) 2762–2770. [7] S.Y. Kim, J.W. Paek, B.H. Kang, Flow and heat transfer correlations for porous fin in a plate-fin heat exchanger, J. Heat Transfer 122 (3) (2000) 572–578, http:// dx.doi.org/10.1115/1.1287170. [8] P. Ranut, E. Nobile, L. Mancini, High resolution microtomography-based CFD simulation of flow and heat transfer in aluminum metal foams, Appl. Therm. Eng. 69 (2014) 230–240. [9] P. Ranut, E. Nobile, L. Mancini, High resolution X-ray microtomography-based CFD simulation for the characterization of flow permeability and effective thermal conductivity of aluminum metal foams, Exp. Therm. Fluid Sci. 67 (2015) 30–36. [10] A. Diani, S. Mancin, L. Doretti, L. Rossetto, Low-GWP refrigerants flow boiling heat transfer in a 5 PPI copper foam, Int. J. Multiphase Flow 76 (2015) 111– 121. [11] S. Mancin, A. Diani, L. Doretti, L. Rossetto, R134a and R1234ze(E) liquid and flow boiling heat transfer in a high porosity copper foam, Int. J. Heat Mass Transfer 74 (2014) 77–87. [12] Y. Zhu, H. Hu, G. Ding, H. Peng, X. Huang, D. Zhuang, Jun Yu, Influence of oil on nucleate pool boiling heat transfer of refrigerant on metal foam covers, Int. J. Refrig. 34 (2011) 509–517. [13] H. Hu, Y. Zhu, H. Peng, G. Ding, S. Sun, Effect of tube diameter on pressure drop characteristics of refrigerant–oil mixture flow boiling inside metal-foam filled tubes, Appl. Therm. Eng. 62 (2014) 433–443. [14] Y. Zhu, H. Hu, S. Sun, G. Ding, Flow boiling of refrigerant in horizontal metalfoam filled tubes: part 1 – two-phase flow pattern visualization, Int. J. Heat Mass Transfer 91 (2015) 446–453. [15] Y. Zhu, H. Hu, S. Sun, G. Ding, Flow boiling of refrigerant in horizontal metalfoam filled tubes: part 2 – a flow-pattern based prediction method for heat transfer, Int. J. Heat Mass Transfer 91 (2015) 502–511. [16] G.B. Abadi, C. Moon, K.C. Kim, Flow boiling visualization and heat transfer in metal-foam-filled mini tubes – Part II: Developing predictive methods for heat transfer coefficient and pressure drop, http://dx.doi.org/10.1016/j. ijheatmasstransfer.2016.03.042. [17] G.B. Abadi, E. Yun, K.C. Kim, Experimental study of a 1 kW organic Rankine cycle with a zeotropic mixture of R245fa/R134a, Energy 93 (2015) 2363–2373. [18] G.B. Abadi, E. Yun, K.C. Kim, Flow boiling characteristics of R134a and R245fa mixtures in a vertical circular tube, Exp. Therm. Fluid Sci. 72 (2016) 112–124. [19] G.B. Abadi, D.Y. Kim, S.Y. Yoon, K.C. Kim, Thermal performance of a 10-kW phase-change plate heat exchanger with metal foam filled channels, Appl. Therm. Eng. (2016), http://dx.doi.org/10.1016/j.applthermaleng.2016.01.156. [20] R.J. Moffat, Using uncertainty analysis in the planning of an experiment, ASME J. Fluids Eng. 107 (2) (1985) 173–178. [21] L. Wojtan, T. Ursenbacher, J.R. Thome, Investigation of flow boiling in horizontal tubes: part I — a new diabatic two-phase flow pattern map, Int. J. Heat Mass Transfer 48 (2005) 2955–2969. [22] V. Gnielinski, New equations for heat and mass transfer in turbulent pipe and channel flow, Int. Chem. Eng. 16 (2) (1976) 359–368.

Please cite this article in press as: G. Bamorovat Abadi et al., Flow boiling visualization and heat transfer in metal-foam-filled mini tubes – Part I: Flow pattern map and experimental data, Int. J. Heat Mass Transfer (2016), http://dx.doi.org/10.1016/j.ijheatmasstransfer.2016.03.043