Concurrent upward liquid slug dynamics on both surfaces of annular channel acquired with liquid film sensor

Accepted Manuscript Concurrent Upward Liquid Slug Dynamics on Both Surfaces of Annular Channel Acquired with Liquid Film Sensor Takahiro Arai, Masahiro Furuya, Taizo Kanai, Kenetsu Shirakaw PII: DOI: Reference:

S0894-1777(14)00139-3 http://dx.doi.org/10.1016/j.expthermflusci.2014.05.018 ETF 8234

To appear in:

Experimental Thermal and Fluid Science

Received Date: Revised Date: Accepted Date:

25 March 2014 27 May 2014 29 May 2014

Please cite this article as: T. Arai, M. Furuya, T. Kanai, K. Shirakaw, Concurrent Upward Liquid Slug Dynamics on Both Surfaces of Annular Channel Acquired with Liquid Film Sensor, Experimental Thermal and Fluid Science (2014), doi: http://dx.doi.org/10.1016/j.expthermflusci.2014.05.018

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.

Concurrent Upward Liquid Slug Dynamics on Both Surfaces of Annular Channel Acquired with Liquid Film Sensor Takahiro ARAI* Central Research Institute of Electric Power Industry, 2-11-1 Iwado-Kita, Komae City, Tokyo 201-8511 Tel: +81-3-3480-2111

Masahiro FURUYA Central Research Institute of Electric Power Industry, 2-11-1 Iwado-Kita, Komae City, Tokyo 201-8511 Tel: +81-3-3480-2111

[email protected]

[email protected]

Taizo KANAI Central Research Institute of Electric Power Industry, 2-11-1 Iwado-Kita, Komae City, Tokyo 201-8511 Tel: +81-3-3480-2111

Kenetsu SHIRAKAWA Central Research Institute of Electric Power Industry, 2-11-1 Iwado-Kita, Komae City, Tokyo 201-8511 Tel: +81-3-3480-2111

[email protected]

[email protected]

*Corresponding author: Central Research Institute of Electric Power Industry, 2-11-1 Iwado-Kita, Komae City, Tokyo 201-8511 [email protected] Tel: +81-3-3480-2111 Fax: +81-3-3480-2844

Keywords: Two-phase flow, Annular channel, Liquid film sensor, Disturbance wave, Ripple wave.

1

Abstract The interfacial behavior of upward liquid film flow is an important phenomenon to evaluate interfacial transfer accompanying the entrainment and deposition of droplets. This research focuses on a vertical annular channel, and an air-water liquid film flow experiment was conducted under atmospheric pressure conditions. The diameters of inner and outer pipes in the annular channel were 12 and 18 mm respectively. The experiment featured multi-point electrode sensors installed in both the inner and outer pipe surface at the same height, and the ability to measure the liquid film distribution on both surfaces in the annular channel simultaneously. As for the sensor structure, 10 × 32 measuring points were arranged in a lattice pattern on the sensor surface and the spatial resolution was 2 × 2 mm, hence the liquid film thickness distribution could be measured rapidly, at over 1250 slices per second. Since the sensor was manufactured by a flexible multilayer substrate, it was applicable to a cylindrical channel surface. In the experiment, water was supplied from the inner pipe surface and uniformly distributed in the circumferential direction, whereupon liquid film distributions were measured 300 mm downstream from the water supply position. The time series data of the liquid film distribution demonstrated circumferential distributions of liquid film thickness and interfacial wave velocity. When the superficial gas velocity was smaller than 20 m/s, a liquid film formed on both inner and outside pipe surfaces, regardless of the superficial 2

liquid velocity. With increasing superficial gas velocity, the film thickness of the outer pipe surface became thinner than that of the inner pipe surface. Measurement of the liquid film thickness on both surfaces of the annular channel also showed that a liquid slug with wavelength of several millimeters passed concurrently through both surfaces in the annular channel.

1.

Introduction

An annular flow is an important phenomenon in industrial plants, such as a steam boiler and a heat exchanger. In particular, since the liquid film flow formed in the channel wall is closely concerned with the heat transfer characteristic of these industrial equipment, it is necessary to understand the characteristic of an interfacial wave. Figure 1 shows a schematic of the upward annular flow in a vertical pipe, in which the liquid film flows along with the channel wall, and the interfacial wave changes with the flow conditions. Many researches have been conducted about the interfacial wave characteristic of the upward annular flow. Based on existing knowledge, the interfacial wave is classified into two types, a ripple wave and a disturbance wave (Hall-Yaylor, N., Hewitt, G. F., and Lacey, P. M. C. 1963, Hewitt. G. F., and Govan, A. H. 1990). The ripple wave has small amplitude and high frequency, and affects shear stress on the

3

gas-liquid interface in the annular flow. It is formed with low superficial liquid velocity and high superficial gas velocity, and distributed relatively uniformly. Conversely, the disturbance wave has a complicated interface where the amplitude of the disturbance wave is relatively large, and is formed when the superficial liquid velocity is large. The disturbance wave can generate droplets from the wave crest while moving more quickly than the ripple wave, making it an important factor affecting droplet entrainment. The measurement of liquid film thickness in annular flow has mainly targeted the vertical pipe since it was easy to understand the basic characteristics of wave behavior and various techniques was applicable to measure the liquid film thickness. Especially, the disturbance wave behavior has been measured by various technique suxch as multi-points probes (K. Sekoguchi and M. Takeishi, 1989) and conductance probe (B. J. Azzopardi, 1986, P. Sawant et al., 2008). A lot of actual industrial equipment includes multi-channel structures, with many heat exchanger tubes installed in proximity, and a liquid film formed on their outer surface. Therefore, it is important to understand the characteristic of the upward liquid film flow in the annular channel. Moreover, the actual interfacial structure of disturbance wave is so complicated that it is neccesary to measure the film thickness fluctuation as a two-dimensional distribution. However, there is little knowledge of the upward liquid film flow in the annular channel. In this research, an annular channel consisting of two 4

cylinder pipes was used as a test section. The objective of this research was to measure the two-dimensional liquid film distribution on both surfaces of the annular channel simultaneously at a high-sampling rate, and evaluate the characteristics of the interfacial wave based on the liquid film thickness and the interfacial wave velocity.

2. 2.1

Experiment Experimental apparatus

Figure 2 shows a schematic of the experimental apparatus. The apparatus consisted of a test

section, an upper plenum, a lower plenum, a water tank, a water-circulating pump and a

compressor. Air was supplied to the lower part of the test section through the lower plenum

from the compressor and water was injected into the test section through the sintered metal

pipe located at the lower part of the inner pipe. When the air-water two-phase flow went

upward through the annular channel, liquid film distributions on both surfaces of inner and

outer pipes were measured by a pair of liquid film sensor, whereupon the two-phase flow

was separated at the upper plenum, and the air released into the atmosphere. The water

was returned to the water tank and re-circulated through the test loop. During the

experiment, pure water was used, to which a small amount of NaOH solution had been

added to adjust its conductivity. Very small amount of NaOH meant there was almost no

5

influence on the physical-properties of water. The test section was 800 mm long.

Figure 3 shows a schematic of the test section. The diameters of the inner and outer pipes in

the annular channel were 12 and 18 mm, respectively and the hydraulic diameter of the test section was 6 mm. The air injector consisted of a sintered metal pipe of 120 μm porosity and the outer diameter of the sintered metal pipe was similar to that of the inner pipe in the test

section. The liquid film was measured after the developing length of 300mm. The

measurement area of the sensor located on the outer pipe surface was ±90 degrees in the

azimuthal direction, and 60 mm in the axial direction, while the measurement area of the

sensor located on the inner pipe surface was ±130 degrees in the azimuthal direction, and

also 60 mm in the axial direction.

2.2 Liquid film measurement Figure 4 shows a schematic of the sensor circuitry. The sensor consists of excited electrodes,

working electrodes, insulating regions, and grounded electrodes. The excited electrodes were

connected in a vertical sequence to form a layer of excited electrodes, while the working

electrodes were connected in a horizontal sequence to form a working electrode layer. The

excited electrodes layer crossed the working electrodes layer in a reticular pattern. At each

intersection of excited and working electrodes, the electrodes were exposed to the sensor

surface as electrode pairs. These electrode pairs were arranged in a reticular pattern on the

6

sensor surface and their liquid film thickness was measured to determine conductivity in

each case. When each electrode is connected to an independent signal wire, two signal wires

are required to constitute an electrode pair. For example, as shown in the figure, when the

sensor had 4×4 electrode pairs, at least 32 signal wires were needed based on existing

methods. However, when 4×4 electrode pairs were arranged by the sensor, both excited and

working electrode wires were reduced to only four, respectively.

To measure the signals acquired from all electrode pairs at high-time resolution, a

signal-processing system with a high sampling rate was required. Accordingly, we applied a

system of wire mesh sensor measurement, which was developed and applied to measure

two-phase flow by Prasser et al. (1998). This system was applied to measure the void

fraction distribution in a vertical pipe. In recent years, the measurement system has also

been applied to liquid film measurement (Damson, M. and Prasser, H.-M., 2009, 2010). We

also developed a liquid film sensor based on the same measurement principle, and showed

that the cylindrical liquid film sensor could be manufactured (Arai, T. et al., 2010).

Figure 5 is a photograph of the cylindrical liquid film sensor, which is twisted around a

stainless-steel pipe of 12-mm diameter, and fixed to the pipe surface. The elliptical deep

brown area is the insulating region, and the two circular parts inner the insulating region

indicate an electrode pair. The other area, which covered the entire sensor surface, is the

grounded electrode. Although the sensor surface was almost smooth; the insulating region

7

was about 20

m lower than the remaining area due to the multilayer printed circuit board

manufacturing process. Ten electrode pairs were distributed in a circumferential direction,

and 32 electrode pairs were arranged in an axial direction in a reticular pattern. The axial

length of the measurement area was 60 mm. The upper and lower ends of the sensor were

bonded to the surface of the stainless-steel pipe as smoothly as possible to avoid forming a

discontinuous surface. The signal wires were soldered to the back of the sensor and

connected to an external device through the inner of the pipe.

Figure 6 shows the sensor calibration method. A liquid film sensor was installed on the

half-pipe surface with the same curvature as the experiment, and the liquid film thickness

formed on the sensor surface was adjusted with a gap adjuster. A calibration curve was

obtained based on the relationship between the known liquid film thickness and the electric

potential at the time. Ten kinds of gap adjuster were prepared and the liquid film thickness

was adjusted between 0.07 and 1.8 mm.

Figure 7 shows the calibration curve of the liquid film sensor used in this experiment. With

liquid film thickness increasing up to 0.6mm, the measured voltage increased almost in

parallel. Since the measured voltage was saturated when the liquid film thickness exceeded

1.2mm, the maximum measurement thickness of the liquid film sensor used in this research

was 1.2mm.

8

2.3 Experimental conditions Figure 8 shows the flow conditions in the experiment, plotted on the flow regime map of the

upward two-phase flow in a vertical pipe proposed by Mishima and Ishii (1984). It can be

observed that the entire experiment corresponds to the annular flow condition in a vertical

pipe. The liquid film distribution on both surfaces of the inner and outer pipes was

concurrently measured at 1250 slices per second as time series data for 25 seconds under

each of the flow conditions.

3.

Experimental data analysis

Based on the time series data of the liquid film thickness distribution obtained by the liquid

film sensor, the interfacial wave velocity could be calculated. When an interfacial wave

passed between two measuring points separated in the axial direction, there was a time lag.

The one-dimensional interfacial velocity in the axial direction was calculated as:

uw =

s

τ

(1)

where s was the distance between the two measuring points, namely 10 mm in this experiment, and τ the time lag. The time lag was calculated by cross-correlation analysis. The time lag in the cross-correlation function represents the average time required for the

disturbance wave to travel from the first measuring point to the second one. The cross-correlation coefficient Rfg was calculated using the following equation:

9

∫ {f (t ) − f }{g (t + ϕ ) − g}dt b

R fg (ϕ ) =

a

∫ {f (t ) − f } dt ∫ {g (t + ϕ ) − g} dt b

2

a

b

(2)

2

a

where f(t), g(t) represented the time series of liquid film thicknesses at the first and second

measurement points, respectively.

Cross-correlation analysis could also be used to investigate the relationship between the

disturbance waves on both surfaces of the annular channel. A liquid film flow formed in each

surface of the annular channel adjacent in close proximity such as 3 mm. To clarify the

characteristics of the interfacial wave in the annular channel, it was important to evaluate

the interaction of the liquid film flow formed on both surfaces of the annular channel. In this

research, the cross-correlation analysis was performed by Eq. (2) using the experimental

data of the liquid film fluctuation on both surfaces of the inner and outer pipes in the same

position. For example, it was assumed that the interfacial waves on both surfaces of the

inner and outer pipes passed concurrently at the same cycle. Cross-correlation analysis of

the liquid film fluctuation on both surfaces of the annular channel showed a remarkable

peak in the correlation coefficient at time lag 0. In addition, power spectrum analysis was

also conducted with time series data on liquid film fluctuation to evaluate the wave

frequency characteristic.

All the experimental data analysis was performed with commercially available software

MATLAB. In the following section, the characteristics of the liquid film behavior in the

10

annular channel is analyzed as film thickness, interfacial wave velocity, and wave frequency

in detail.

4.

Results and discussion

4.1 Upward liquid film behavior in the annular channel Figure 9 shows the experimental result with superficial liquid and gas velocities of 0.17 and 20 m/s, respectively. Figure 9(a) shows snapshots of the liquid film distribution on both surfaces of the inner and outer pipes, respectively. The disturbance wave over 1mm of liquid film thickness passed at intervals of about 100 ms. The disturbance wave length in the axial direction was relatively long, peaking at around 30 mm. Since the disturbance wave formed in the same position on both surfaces of the inner and outer pipes, it emerged that the disturbance waves passed concurrently. Figures 9(b), (c), (d) show the liquid film thickness, the cross-correlation of liquid film on both surfaces of the annular channel, and the wave frequency at 50 mm from the bottom of the liquid film sensor and an azimuthal angle of -14 degrees, respectively. Based on power spectrum analysis, the first peaks of the interfacial wave frequency on both surfaces of liquid film were about 10 Hz, which indicates that the disturbance wave dominated. Moreover, the peak of the spectrum was observed up to 100 Hz in the liquid film on the inner pipe surface and the ripple wave was also remarkable. Figure 9(e) shows the 11

azimuthal distribution of the time-averaged liquid film thickness and the interfacial velocity. It emerged that the azimuthal distribution was almost uniform, with little difference in interfacial velocity on both surfaces of the inner and outer pipes. The liquid film thickness on the outer pipe surface exceeded that on the inner pipe surface. According to the cross-correlation of liquid film fluctuation on both surfaces of the annular channel, the correlation coefficient was relatively high. It is suggested that the disturbance wave on both surfaces synchronizes. Figure 10 shows the experimental result with superficial liquid and gas velocities of 0.17 and 26 m/s, respectively. Since the superficial gas velocity increased beyond the experimental result shown in Fig. 9, a shortened axial length of the disturbance wave was observed. However, the dominant frequency of the interfacial wave was about 10 Hz, namely the same as that of the experimental result shown in Fig. 9. According to the interfacial wave on the outer pipe surface, the peak of the interfacial wave spectrum over 100 Hz was observed. This result shows that a ripple wave of high frequency was concurrently observed. Since the amplitude of the disturbance wave decreased and the ripple wave relatively expanded with the increase in superficial gas velocity, the correlation of the liquid film flow on both surfaces of the annular channel fell in relative terms. Figure 11 shows the experimental result with superficial liquid and gas velocities of 12

0.17 and 52 m/s, respectively. With a further increase in superficial gas velocity, the disturbance wave was observed only on the inner pipe surface. In the liquid film flow on the inner pipe surface, the disturbance wave over 0.5 mm of liquid film thickness passed at 30-70 Hz of wave frequency. In the liquid film flow on the outer pipe surface, the ripple wave below 0.1 mm of wave amplitude dominated. Consequently, no correlation of the liquid film fluctuation on both surfaces was observed. Compared with the difference between the interfacial wave velocities on both surfaces of the annular channel, the interfacial wave velocity on the inner pipe surface rose with increasing superficial gas velocity. Conversely, the interfacial wave velocity on the inner pipe surface decreased rather than that of the experimental result as shown in Fig. 10, which is the main reason why the dominant liquid film flow on the outer pipe surface transited to the ripple wave from the disturbance wave. Therefore, the liquid film thickness on the outer pipe surface was thinner than that on the inner pipe surface. In this experiment, discussion about the development distance of the annular flow was required. Kataoka and Ishii (1982) proposed a developing length z ∞ necessary to establish an equilibrium annular flow as follows:

z∞ ≅

440.0 DWe 0.25 Re 0f.5

13

(3)

Re f =

ρ L jL D μL

2

ρ G jG D ⎛ ρ L − ρ G ⎜⎜ We = σ ⎝ ρG

Reynolds number

⎞ ⎟⎟ ⎠

(4)

13

Weber number

(5)

where D, ρG, ρL,, , μL, and σwere the test section diameter, gas phase density, liquid phase density, area-averaged superficial gas velocity, area-averaged liquid velocity, liquid phase viscosity and surface tension, respectively. In this experiment, z∞ /D was calculated as about 90 when the hydraulic diameter of the test section was used as D. Wolf et al. (2001) obtained the following results for the developing characteristics of upward liquid flow in a vertical pipe of inner diameter 31.8 mm and 10.8 m long. Both the liquid film thickness and flow rate become almost constant by 120-330 D. Based on the spectrum analysis of the liquid film fluctuation, an interfacial wave was fully developed by 50-100 D, while both the frequency and velocity of the disturbance wave become almost constant by 100-300 D. These results indicate that each parameter of the upward liquid film flow, whereby water is supplied through a porous tube, became constant by 100 D. Since z/D in this experiment was about 50, the upward liquid film flow may be under development. However, the liquid film distribution on both surfaces of the annular channel could be measured concurrently by a pair of sensors, and it is shown that the effect of the inlet condition of the test section on the developing characteristics of upward liquid film flow was also evaluated. 14

As mentioned above, a pair of liquid film sensors could acquire a time series of the upward liquid film distributions on both surfaces of the annular channel concurrently. The results show that a disturbance wave over 1mm of liquid film thickness formed on both surfaces of the inner and outer pipes at near 20 m/s, and mutually synchronized. With the increase in superficial gas velocity, the amplitude of the disturbance wave declined, and the wave frequency shifted to a high frequency region. Therefore, the correlation of the liquid film fluctuation on both surfaces of the annular channel relatively decreased. 4.2 Interfacial wave characteristics in the annular channel Figure 12 shows the time-averaged liquid film thickness on the inner and outer pipes,

respectively. With a Weber number of under 1000, the liquid film thickness on the inner

pipe was almost constant. Except for the disturbance wave, the liquid film seldom changed,

although characteristics of the disturbance wave, such as the amplitude, axial length and

wave frequency, all varied. However, further increasing the Weber number accelerated the

moving velocity, not only of the disturbance wave but also the ripple wave, and decreased

the total liquid film thickness. Conversely, the liquid film thickness on the outer pipe

surface monotonically decreased with increasing Weber number. In this experiment, since

the water was supplied from the inner pipe surface, it is considered that the upward liquid

15

film flow had a tendency to flow along the inner pipe surface with increasing superficial gas

velocity.

Figure 13 shows the effect of the Weber number on interfacial wave velocity. The

experimental results were plotted for every liquid film Reynolds number, and the existing

result in a vertical pipe 9.4 mm in diameter, as obtained by Sawant et al. (2008), was also

plotted. Compared with these results, it emerged that the interfacial wave velocity on the

inner pipe surface showed virtually the same tendency. In other words, the interfacial wave

velocity on the inner pipe surface was the same as that of the fully developed liquid film

flow. Conversely, the interfacial wave velocity on the outer pipe surface showed a different

tendency with increasing superficial gas velocity. As mentioned in Fig. 11, the dominant

liquid film flow on the outer pipe surface transited to the ripple wave from the disturbance

wave with increasing superficial gas velocity. When the Weber number was under 1000, the

interfacial wave velocity showed almost the same tendency since the disturbance wave

dominated on the outer pipe surface. However, when the Weber number exceeded 1000, the

disturbance wave was hardly formed but the ripple wave became a dominant flow. Therefore, the interfacial wave velocity once decreased around We = 1000. As the Weber

number increased further, the ripple wave velocity, i.e. the interfacial wave velocity

increased.

Figure 14 shows the maximum of cross-correlation coefficient obtained by cross-correlation

16

analysis of the interface wave for each flow condition. The cross-correlation was relatively

high under low superficial gas velocity amd high superficial liquid velocity conditions. The

time delay at the maximum of cross-correlation coefficient was within a range of 1 to 5 ms.

The cross-correlation coefficient decreased with increasing superficial gas velocity. Since the

liquid film fluctuation formed on the outer pipe surface became very small, particularly

when the superficial liquid velocity was less than 0.035 m/s, the cross-correlation coefficient

was about 0. Based on the experimental results of the liquid film distribution and the

maximum correlation coefficient under each flow condition, it can be judged that the

correlation of the liquid film fluctuation between both surfaces in the annular channel was

relatively high and the interfacial waves synchronized with each other when the maximum

correlation coefficient exceeded 0.2. In other words, when the superficial liquid velocity was

less than 0.035 m/s, the correlation of the liquid film fluctuation between both surfaces in

the annular channel was low since it was hard for the disturbance wave to form on the outer

pipe surface. As the superficial liquid velocity exceeded 0.035 m/s, the dominant interfacial

wave changed from the ripple wave to the disturbance wave. Therefore, the correlation of

the liquid film fluctuation became high under the low-superficial gas flow condition such as

24 m/s or less. Since the dominant interfacial wave on the outer pipe surface changed from a

disturbance wave to a ripple wave with increasing superficial gas velocity in this

experiment, the correlation of the liquid film fluctuation also decreased.

17

5.

Conclusion

The upward liquid film distribution on both surfaces of the annular channel was acquired by

liquid film sensors at 1250 slices per second. Based on the experimental result, the

interfacial wave characteristic in the annular channel was clarified as follows:

A large disturbance wave formed on both surfaces of the inner and outer pipes at a

superficial gas velocity of around 20 m/s. Cross-correlation analysis of the upward liquid

film fluctuation on both surfaces of the annular channel showed that the disturbance wave

moved concurrently along the channel surface. The liquid film thickness decreased with

increasing superficial gas velocity. In particular the liquid film flow on the outer pipe

surface transited to the ripple wave with increasing superficial gas velocity, and the liquid

film thickness on the same decreased remarkably beyond that on the surface of the inner

pipe. Consequently, cross-correlation of the liquid film fluctuation in both of pipe surfaces

declined with increasing superficial gas velocity.

The interfacial wave velocity on the inner pipe surface rises with increasing superficial gas

and liquid velocity. This result was almost in agreement with the existing experimental

result, which measured the interfacial wave velocity of fully developed annular flow in a

vertical pipe quantitatively. In the case of the interfacial wave on the outer pipe surface, the

disturbance wave decreased rapidly with increasing superficial gas velocity. Since the

dominant interfacial wave changed to a ripple wave from a disturbance wave, the interfacial

18

wave velocity slowed down. A further increase in the interfacial gas velocity triggered a

further rise in the interfacial wave velocity again.

Acknowledgements The authors would like to thank Professor H.-M. Prasser of ETH for his many helpful

suggestions. They also wish to thank Mr. Takeo Yoshioka of CERES Inc. and Mr. Yoshiyuki

Shiratori of Electric Power Engineering Systems Co., Ltd., for helping with these

experiments.

6.

References Arai, T., Furuya, M., Taizo, K., 2010. Development of high-speed and

multidimensional measurement technique for liquid film behavior, Proc. ISFV14. Azzopardi, B. J., 1986. Disturbance wave frequencies, velocities and spacing in vertical annular two-phase flow, Nucl. Eng. Des., 92, 121-133. Damson, M., Prasser, H.-M., 2009. High-speed liquid film sensor for two-phase flows with high spatial resolution based on electrical conductance, Flow Measurement and Instrumentation, 20, 1-14. Damson, M., Prasser, H.-M., 2010. Experimental studies of the effect of functional spacers to annular flow on subchannels of a BWR fuel element, Nucl. Eng.

19

Des., 240, 3126-3144. Hall-Yaylor, N., Hewitt, G. F., and Lacey, P. M. C., 1963. The motion and frequency of large disturbance wave in annular two-phase flow of air-water mixtures, Chem. Eng. Sci., 18, 532-552 . Hewitt. G. F., and Govan, A. H., 1990. Phenomena and prediction in annular two-phase flow, ASME FED-Vol. 99, Advances in Gas-Liquid Flows, Kim, J. H., et al., eds., 41-56. Kataoka, I., Ishii, M. 1982. Entrainment rate in annular two-phase flow, Argonne National Laboratory Report. ANL-82, Argonne, Illinois. Mishima, K., Ishii, M., 1984. Flow regime transition criteria for upward two-phase flow in vertical tubes, Int. J. Heat Mass Transfer, 27, 5, 723-737. Prasser, H.-M. et al., 1998. A new electrode-mesh tomograph for gas-liquid flows, Flow Measurement and Instrumentation, 9, 111-119. Sawant, P., Ishii, M., Hazuku, T., Takamasa, T., Mori, M., 2008. Properties of disturbance wave in vertical annular two-phase flow, Nucl. Eng. Des., 238, 3528-3541. Sekoguchi, K. and Takeishi, M., 1989. Interfacial struxtures in upward huge wave flow and annular flow regimes, Int. J. Multiphase Flow, 15, 3, 295-305. Wolf, A., Jayanti, S., Hewitt, G. F., 2001. Flow development in vertical annular flow,” Chem. Eng. Sci., 56, 3221-3235. 20

Figure list Fig. 1. Schematic of the upward annular flow Fig. 2. Schematic of experimental apparatus Fig. 3. Schematic of the test section Fig. 4. Measurement principle of liquid film Fig. 5. Photograph of the cylindrical liquid film sensor Fig. 6. Calibration of the liquid film thickness Fig. 7. Calibration curve of liquid film sensor Fig. 8. Experimental flow condition Fig. 9. Upward liquid film behavior (=0.17m/s, =20m/s) Fig. 10. Upward liquid film behavior (=0.17m/s, =26m/s) Fig. 11. Upward liquid film behavior (=0.17m/s, =52m/s) Fig. 12. Effect of flow conditions on the mean liquid film thickness Fig. 13. Effect of flow conditions on the interfacial wave velocity Fig. 14. Cross-correlation of interfacial wave on both surfaces of the annular channel

21

Liquid film

Vertical pipe Disturbance wave

Ripple wave Droplet

Fig. 1. Schematic of the upward annular flow

22

Air

P Pressure T Temperature F Flow rate

Signal processor

P

Upper plenum Liquid film sensors m m 0 0 8

Water tank

Water injection area

T

Flow meter F

P

Water pump

F

P

Lower plenum

Flow meter

Fig. 2. Schematic of experimental apparatus

23

Air compressor

Outer pipe m m m m 8 2 1 1 D D I O

Inner pipe Liquid film sensor

0 0

m m 0 6

Liquid film sensor

z

m m 0 0 3

r e t a W

Water injection from sintered metal ri A

Fig. 3. Schematic of the test section

24

Excited Electrode Working Electrode

Acquired Signals S4

S3

S2

S1 Power supply

Switching Excitation Pulses

Fig. 4. Measurement principle of liquid film

25

Fig. 5. Photograph of the cylindrical liquid film sensor

26

Gap adjuster Liquid film

& Sensor

Fig. 6. Calibration of the liquid film thickness

27

Normarized voltage (-)

1.2 1.0 0.8 0.6 0.4 0.2 0.0 0.0

0.5 1.0 1.5 Liquid film thickness, δ (mm)

Fig. 7. Calibration curve of liquid film sensor

28

2.0

Superficial liquid velocity, (m/s)

1 Annular Slug 0.1

0.01 1

Churn

10 100 Superficial gas velocity, (m/s) Fig. 8 Experimental flow condition

29

Inner pipe

0

0

Outer pipe

0 ms

40 ms

80 ms

120 ms

160 ms

200 ms

240 ms

280 ms

360 ms

320 ms

1.0

Normalized power spectrum

1.2

Inner pipe Outer pipe

0.8 0.6 0.4 0.2 0.0 0

100

200 300 Time, t (ms)

400

500

(b) Time series of the liquid film thickness Cross-correlation coefficient, Rfg (-)

Liquid film thickness, δ (mm)

(a) Snapshots of the liquid film thickness distribution 1.0

Inner pipe Outer pipe

0.8 0.6 0.4 0.2 0.0 0

100

200

(c) Interfacial wave frequency

0.6

0.4

0.2

0.0

-0.2 -200

-100

0 100 Time lag, Δ t (ms)

30

300

Frequency, f (Hz)

200

400

Inner pipe Outer pipe

1.0

4

0.8 Interfacial wave velocity

3

0.6 0.4

2

Liquid film thickness

1

0.2 0.0

-100

0

Azimuthal angle, θ (deg)

100

Interfacial wave velocity, uw (m/s)

Time-averaged liquid film thickness, δ (mm)

5

1.2

0

(d) Cross-correlation function (e) Azimuthal liquid film distribution Fig. 9. Upward liquid film behavior ( = 0.17 m/s, = 20 m/s)

31

Inner pipe

1

0

0

Outer pipe

0 ms

40 ms

80 ms

120 ms

160 ms

200 ms

240 ms

280 ms

360 ms

320 ms

1.0

Normalized power spectrum

1.2 Inner pipe Outer pipe

0.8 0.6 0.4 0.2 0.0 0

100

200 300 Time, t (ms)

400

500

(b) Time series of the liquid film thickness Cross-correlation coefficient, Rfg (-)

Liquid film thickness, δ (mm)

(a) Snapshots of the liquid film thickness distribution 1.0

Inner pipe Outer pipe

0.8 0.6 0.4 0.2 0.0 0

100

200

0.4

0.2

0.0

-100

0 100 Time lag, Δ t (ms)

32

400

(c) Interfacial wave frequency

0.6

-0.2 -200

300

Frequency, f (Hz)

200

1.0

6

Inner pipe Outer pipe

5 Interfacial wave velocity

0.8

4

0.6

3

0.4

2

Liquid film thickness

1

0.2 0.0

-100

0

Azimuthal angle, θ (deg)

100

Interfacial wave velocity, uw (m/s)

Time-averaged liquid film thickness, δ (mm)

1.2

0

(d) Cross-correlation function (e) Azimuthal liquid film distribution Fig. 10. Upward liquid film behavior ( = 0.17 m/s, = 26 m/s)

33

Inner pipe

1

0

0

Outer pipe

40 ms

80 ms

120 ms

160 ms

200 ms

240 ms

280 ms

320 ms

360 ms

(a) Snapshots of the liquid film thickness distribution Liquid film thickness, δ (mm)

1.2 1.0

Inner pipe Outer pipe

0.8 0.6 0.4 0.2 0.0 0

Normalized power spectrum

0 ms

100

200 300 Time, t (ms)

400

1.0

500

Inner pipe Outer pipe

0.8 0.6 0.4 0.2 0.0 0

100

200

300

Frequency, f (Hz)

(b) Time series of the liquid film thickness

34

400

(c) Interfacial wave frequency

0.4

0.2

0.0

-0.2 -200

1.2

-100

0 100 Time lag, Δ t (ms)

6

Inner pipe Outer pipe

5

1.0 0.8

4

Interfacial wave velocity

3

0.6 0.4

2

Liquid film thickness

1

0.2 0.0

200

-100

0

Azimuthal angle, θ (deg)

100

Interfacial wave velocity, uw (m/s)

Cross-correlation coefficient, Rfg (-) Time-averaged liquid film thickness, δ (mm)

0.6

0

(d) Cross-correlation function (e) Azimuthal liquid film distribution Fig. 11. Upward liquid film behavior ( = 0.17 m/s, = 52 m/s)

35

Mean liquid film thickness, δ (mm)

Mean liquid film thickness, δ (mm)

0.6 Inner pipe

Ref=240 Ref=400

0.5

Ref=560 Ref=800

0.4

Ref=1100

0.3 0.2 0.1 0.0 1 10

2

3

4

10 10 10 Weber number, We (-)

5

10

0.6 Outer pipe

Ref=240

0.5

Ref=400 Ref=560 Ref=800

0.4

Ref=1100

0.3 0.2 0.1 0.0 1 10

2

3

4

10 10 10 Weber number, We (-)

5

10

Fig. 12. Effect of flow conditions on the mean liquid film thickness

36

Interfacial wave velocity, uw (m/s)

Interfacial wave velocity, uw (m/s)

12

12

Re =240 Outer pipe Re =400 10

Inner pipe 10 8 6 4 2 0 1 10

2

3

4

10 10 10 Weber number, We (-)

10

f

Ref=240

f

Ref=400

Ref=560

Ref=560

Ref=800

Ref=800

Re 8f=1100

Ref=1100

Sawant et al., Ref=500

Sawant et

Sawant et al., Ref=950

Sawant et

Sawant et al., Ref=1480

Sawant et

Sawant et al., Ref=3100

Sawant et

Sawant et al., Ref=5700

Sawant et

6 4 2

0 1 10

5

2

3

4

10 10 10 Weber number, We (-)

Fig. 13. Effect of flow conditions on the interfacial wave velocity

37

10

5

0.5

0.4

0.3 0.2 0.1 0

51

39

26

)( t n ie c if f e o c n o ti la re r o c f o m u m ix a M

19

64 77

Fig. 14. Cross-correlation of interfacial wave on both surfaces of the annular channel

38