Experimental studies of condensing vapor bubbles in subcooled pool water using visual and acoustic analysis methods

Annals of Nuclear Energy 110 (2017) 171–185

Contents lists available at ScienceDirect

Annals of Nuclear Energy journal homepage: www.elsevier.com/locate/anucene

Experimental studies of condensing vapor bubbles in subcooled pool water using visual and acoustic analysis methods Hongrae Jo, Daeseong Jo ⇑ School of Mechanical Engineering, Kyungpook National University, 80 Daehak-ro, Daegu 41566, Republic of Korea

a r t i c l e

i n f o

Article history: Received 20 December 2016 Received in revised form 9 June 2017 Accepted 18 June 2017

Keywords: Rising bubble Collapsing vapor Pressure waves Hydrophone

a b s t r a c t In this study, smooth bubble condensation that occurs in subcooled pool water was examined to understand condensation phenomena at the beginning of boiling. The behaviors of condensable and noncondensable bubbles were investigated with respect to various temperatures and bubble sizes. The experimental data were analyzed using visual and acoustic methods including Phase Interface Binarization (PIB), Tridimensional Reconstructing Assumption (TRA), and acoustic data conversion. For visual analyses, (1) the PIB method determined bubble departure frequency, condensation time, and rising distance, and (2) the TRA method determined departure bubble size and volume reduction rate. The bubble detached with smaller volume and occurred more frequently with a smaller nozzle and a higher subcooling degree. Because the condensation always occurred during the growth, necking, and detachment of a bubble, the bubble was detached before it grew sufficiently. Lower condensation time and rising distance were associated with higher subcooling degrees and smaller injected bubbles. With respect to the acoustic analysis, sound signals were measured using a hydrophone, and the obtained analog data were converted to sound pressure units. The results revealed that higher volume reduction rate resulted in stronger sound pressure. As a result, the condensation phenomena in the smooth bubble regime were visually observed, and the possibilities of acoustic monitoring for earlier boiling were investigated. Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction In the field of nuclear engineering, it is important to understand boiling phenomena under subcooled pool boiling conditions due to various heat transfer effects such as pressure drop, flow instability, and cooling efficiency. As shown in Fig. 1, different boiling phenomena occur with excess temperature and heat flux. In nucleate boiling, Onset of Significant Void (OSV) exists from the Onset of Nucleate Boling (ONB) to the Departure from Nucleate Boiling (DNB). Prior to the OSV, vapor bubbles that are detached from the heating surfaces collapse and disappear since the flow in the core continues to be subcooled. Following the OSV, the bubbles detached from the heating surfaces survive, coalesce, and form large bubbles, which may lead to flow instability and Critical Heat Flux (Faghri and Zhang, 2006). Therefore, detection of boiling is important to prevent unfavorable flow behaviors. Void fraction measurement is widely used to monitor boiling occurrences. Several researchers measure void fraction using electrical conductivity probes, gages for differential pressure, optic fibers, and electro-magnetic sensors (Hetsroni, 1982). The start ⇑ Corresponding author. E-mail address: [email protected] (D. Jo). http://dx.doi.org/10.1016/j.anucene.2017.06.030 0306-4549/Ó 2017 Elsevier Ltd. All rights reserved.

point of boiling is ONB. However, existing methods to detect boiling occurrence have focused on OSV to predict Critical Heat Flux (CHF) that should be avoided to maintain the fuel integrity. These methods are not appropriate to detect boiling occurrences prior to the OSV since vapor bubbles collapse and disappear due to condensation in this regime. Hence, a new methodology for monitoring boiling is required to detect earlier boiling when compared with conventional techniques. Several researchers examined condensation phenomena to understand detailed processes of condensation. Representatively, Tang et al. (2015a,b) defined regimes that showed that the bubble condensation became significant and rough based on the temperature of sub-cooled water and the flow rate of the injecting gas. Four different condensing characteristics were defined and included a capillary wave, transition, shape oscillation, and smooth bubble regimes. Fig. 2 shows two representative regimes; (1) capillary wave regime and (2) smooth bubble regime. The smooth bubble condensation is observed in a low flow rate condition, and the bubbles possess clear phase interfaces. In contrast, capillary wave condensation occurs in a high flow rate condition, and fine waves are formed on the bubble surfaces. Additionally, Jeon et al. (2011) and Pan et al. (2012) showed numerical simulations of collapsing bubbles using a Volume of

172

H. Jo, D. Jo / Annals of Nuclear Energy 110 (2017) 171–185

Nomenclature d din dpx Fi Fb fd h LSP l lref nd nleft nright q00 r rn rk Ref dBSPL Ref V=lPa SdB SmV=Pa T tc td V _ V

Diameter of injecting nozzle [mm] Inner diameter of nozzle [mm] Number of pixel in the nozzle region Interfacial tension force [mN] Buoyancy force [N] Bubble departure frequency [Hz] Rising distance [mm] Sound pressure level [dB re 1 lPa] Distance of hydrophone from the nozzle [m] Reference distance of sound pressure level [m] Number of bubble departure Pixel number of left boundary Pixel number of right boundary Heat flux [W=m2 ] Radius [mm] Inner radius of nozzle [mm] Radius of unit cylinder generated on the kth row [mm] Reference pressure level in water [dB] Reference voltage at a micro-pascal [V] Sensitivity [dB] Sensitivity [mV=Pa] Temperature [°C] Time distance between detachment and condensation [s] Time distance between an departure of a bubble and that of follow-up bubble [s] Volume [ml] Volume reduction rate [ml/s]

Fluid (VOF) model with various conditions of condensation such as different subcooling degrees, inlet pressures, and initial bubble sizes. In addition to these phenomenological studies, new measurement techniques were studied by a few researchers. Specifically, an acoustic measurement was suggested because sound waves occur when a bubble collapses by cavitation or condensation. Acoustic sensing methods were introduced by Minnaert (1933) who studied sound from occurrence and breakage of a noncondensable bubble. These methodologies were further developed by Leighton (1997). Vanquez et al. (2005) reviewed and applied these measurement techniques to determine bubble size. The acoustically measured results were compared with other data obtained from video imaging and inverted funnel techniques.

Vin y

Indicated voltage [V] Highest pixel number of y-axis in the image

Greeks

a

h

r ql qg

Sound attenuation coefficient [dB/m] Contact angle [deg] Surface tension [N/m] Density of water [kg/m3] Density of gas [kg/m3]

Subscripts sat Saturated sub Subcooled sur Surface Abbrevations CHF Critical Heat Flux DAQ data acquisition system DNB Departure from Nucleate Boiling ONB Onset of Nucleate Boiling OSV Onset of Significant Void PIB Phase Interface Binarization RMS root mean square SPL sound pressure level TRA Tridimensional Reconstructing Assumption VOF Volume of Fluid VRR volume reduction rate

Based on the fore-mentioned techniques, several researchers applied hydrophones to various experiments in which the behaviors of non-condensable bubbles were observed. For example, acoustic measuring was applied to check the effects of antifoam agents (Al-Masry et al., 2006), predict transitions of flow regimes (Al-Masry et al., 2007; Ajbar et al., 2009), detect bubbles from a sediment (Vanquez et al., 2014), and detect bubbles generated in a cross-flow (Chicharro and Vanquez, 2014). Furthermore, acoustic sound measurement was applied to a few experiments related to cavitation. Staudenraus and Elsenmenger (1993) and Osterman et al. (2009) characterized cavitation by measuring ultrasonic waves and shockwaves using a hydrophone. Additionally, sound pressure waves from the collapse of cavitation bubbles were detected in a study by Tinguely (2013).

Fig. 1. Boiling curves in pool boiling.

H. Jo, D. Jo / Annals of Nuclear Energy 110 (2017) 171–185

173

Fig. 2. Condensation processes of capillary wave and smooth bubble condensation regimes.

Monitoring of boiling with acoustic methods was suggested by Nishihara and Bessho (1977) who examined the sound emission with respect to nucleate pool boiling. Following this, Lee (2011) focused on methodologies of detection using hydrophones in engine cooling systems, and Geraldo et al. (2014) applied the same in sodium boiling by using basic signal processes. Recently, Tang et al. (2015a,b) defined condensation regimes and described the characteristics of acoustic signals involved in vapor bubble condensation. Additionally, they identified regimes with the features of sound waves such as dominant frequencies. The results of the experiments indicated that condensation can be defined as a capillary wave or transition regimes if the oscillation includes very high-frequency components. In contrast, the regime is termed as a smooth bubble or shape oscillation regime when the sound frequency is lower when compared with that of the other regimes. As previously mentioned, several previous studies on condensation phenomena covered various aspects. However, in contrast to previous studies, the experiments in the present study focused more on a smooth bubble condensation regime that occurs at a low flow rate of vapor. In order to efficiently respond to the earlier boiling, it is necessary to understand the smooth bubble condensation in further detail because a condensation phenomenon at the beginning of boiling is similar to the fore-mentioned condensation region. In the present study, the condensation phenomena were observed under various subcooling degrees, flow rates of vapor, and bubble sizes, and examined by visual and acoustic analysis methods. The visual analysis results were obtained by two image processing techniques, namely (1) Phase Interface Binarization (PIB) and (2) Tridimensional Reconstructing Assumption (TRA). In order to analyze acoustic signals, analog type signals were converted such that they included a unit of sound pressure level. The analysis methods are described in detail in Section 3.

The flow rates of the vapor bubbles through each nozzle were controlled by needle valves. The needle valves were used to keep the condensation regime in a smooth bubble condensation condition. Beside the needle valves, various valves were applied in system for various purposes as shown in Table 1. The bubble was generated in different ways depending on whether the bubble is condensable or not. For condensable steam bubble, a steam generator with an electrical heater was installed below the test section. An electrical heater of 1200 W was placed at the bottom of the tank. The power of the electrical heater was controlled in the range of 10 W–1200 W such that the amount of the vapor generated could be controlled. For non-condensable bubble, a syringe pump was used with a 50 mL syringe. The air was injected in the range of 10 to 100 mL per an hour. With respect to the visual observations, a high-speed camera (Miro eX4) was used, and the images were captured at 2000– 5000 frames per second. The frame rate was selected to capture the fast process of condensing phenomena in which the steam ejected from the nozzle and exposed to the subcooled water was completely condensed within 10–80 ms (Tang et al., 2015a,b).

2. Experimental apparatus Fig. 3 shows the experimental apparatus. A two-dimensional test section was designed to simulate a reactor pool with 25 mm thickness, 300 mm width, and 500 mm length. The sizes of the test section were selected such that there was sufficient space for rising and condensing bubbles. The front and back of the test section were composed of a transparent poly-carbonate panel for visual observations, and the pool was filled with demineralized water. As shown in Fig. 4, the bottom of the test section included three stainless steel nozzles with the following sizes: 6.35 mm (ID: 4.57 mm), 3.18 mm (ID: 1.78 mm), and 1.59 mm (ID: 0.79 mm).

Fig. 3. Schematics of experimental apparatus.

174

H. Jo, D. Jo / Annals of Nuclear Energy 110 (2017) 171–185

Fig. 4. Various nozzles for injecting gas; (a) 1.59 mm (ID: 0.79 mm), (b) 6.35 mm (ID: 4.57 mm), and (c) 3.18 mm (ID: 1.78 mm).

Table 1 Information of various valves in the experimental system. V1 V2 V3 V4 V5 V6 V7 V8 V9 V10 V11

Needle valve for controlling gas flow rate injected through the 1.59 mm OD nozzle Needle valve for controlling gas flow rate injected through the 6.35 mm OD nozzle Needle valve for controlling gas flow rate injected through the 3.18 mm OD nozzle Ball valve for switching connection between the syringe and the injecting nozzles Ball valve for reloading gas into the syringe by switching connection between outside and the lines Ball valve for switching connection between the steam generator and the injecting nozzles Relief valve for preventing pressure overload in the steam generator Ball valve for draining Ball valve for switching connection between the water reservoir and the test section Ball valve for switching connection between the water reservoir and the steam generator Ball valve for draining

The sound waves causing by condensation were detected by a miniature hydrophone (Piezoelectric transducer, TC4013 of Reson co.) at a position that was at a distance of 0.05 m from the nozzles. The signals were transformed through an pre-amplifier and filter (EC6081 of Reson co.). Following the transformation, the analog data were transmitted and acquired using a data acquisition system (DAQ) (NI9234) with a 24 bit resolution and 51,200 Hz sampling rate. All the experiments were performed under a room temperature of 23 °C and atmospheric pressure. In order to study different behaviors in the smooth bubble condensation regime, various subcooling degrees (1 K, 5 K, 10 K, 15 K, 20 K, and 25 K) and bubble sizes were used.

resents white. In order to obtain binary values along the axial direction, it is necessary to modify the image such that it was 0 for the liquid phase and 1 for the gas phase. From the colorbinarized image, the PIB data could be extracted to check whether a gas-liquid interface existed at any axial location. The image conversion processes are shown in Fig. 5 (a) through (e) in detail. In Fig. 5(b), the original image shown in Fig. 5(a) is adjusted by multiplying the values of each pixel to make a dark part darker and a bright part brighter in the image. The adjusted image in Fig. 5(b) is converted from a grayscale to a black and white scale. With respect to the color-binarizing step, a criterion involves first averaging the values of all pixels in an image. If the value before is lower than the criterion value then the value should be changed to 0, or if not, then the value should be changed to 1. Fig. 5(c) shows the color-binarized result. Following colorbinarization, the black and white image was color-reversed since the most important part corresponded to the gas phase region in the image and not to the liquid region. Thus, as shown in Fig. 5 (d), the interface between gas and liquid becomes white, and the others regions become black. However, the void inside the bubble is colored black same as the water region. To make a clear distinct between gas and liquid, the image needs to be further modified as follows: if a diameter of a black-colored region surrounded by gas is smaller than that of the interface of the bubble, it is considered as gas. As shown in Fig. 5(e), the color-binarized image indicates white as gas-phase and black as liquid-phase. Following the conversion of the original images, it was determined whether the gas-liquid interface existed. If there an interface existed, then a value of the line was expressed as 1, and if not, then it was expressed as 0. Fig. 5(f) shows the binary value along the axial direction for an individual image.


fðxÞ ¼

3. Methodologies 3.1. Visual analyses In order to understand the detailed condensation process in the smooth bubble regime, the images were analyzed using two different image processing techniques, namely (1) Phase Interface Binarization (PIB) and (2) Tridimensional Reconstructing Assumption (TRA). The PIB method is used to obtain bubble departure frequency, rising distance, and condensation time. The TRA method is used to obtain volume and tridimensional shape of the detaching bubbles. 3.1.1. Phase Interface Binarization The PIB method provides a binary value of either 0 or 1 for the existence of the interface between the gas and liquid phases. As shown in Fig. 5(a), the original image corresponds to a grayscale with a range from 0 to 255, where 0 represents black, and 255 rep-

0

ðInterface nonexistenceÞ

1 ðInterface existenceÞ

ð1Þ

This was followed by changing individual images to unit vertical lines with black and white colors. The arrangement of these lines relative to time is shown in Fig. 5(g), and this is referred to as a PIB image. The x-axis of the image shows time information, the y-axis shows height from the nozzle, and the values indicate whether each image includes phase interfaces. Line A-B indicated in Fig. 5(g) corresponds to the interface existence from Fig. 5(f). To explain physical meanings of the PIB image, some parts of PIB data were specified as drawn in Fig. 6. First, the qualitative meanings of PIB data are described as follows: on a PIB image as shown in Fig. 6(a), a single icicle shaped unit indicated a process from occurrence to complete collapse of a condensable bubble. The leftmost point on the bottom of the unit indicates a moment of bubble occurrence, and the rightmost point on the bottom indicates a moment of bubble detachment from the nozzle. Additionally, the top endpoint of the icicle indicates that a condensable bubble completely collapses. This point also includes rising dis-

H. Jo, D. Jo / Annals of Nuclear Energy 110 (2017) 171–185

175

Fig. 5. Process of Phase Interface Binarization (PIB); (a) Original image, (b) Adjusted image, (c) Color-binarized image, (d) Color-reversed image, (e) Filled image, (f) PIB values of individual images, and (g) PIB data.

Fig. 6. Schematics of analyses using PIB data; (a) Physical meaning of PIB data, (b) Modified PIB data.

tance information defined as a length between nozzle and condensation point in the PIB data. For getting more quantitative information from PIB data, the results should be modified as shown in Fig. 6(b). The black fullline consists of the first appeared interfaces from the topmost side of Fig. 6(a). The blue dashed-line consists of the first appeared interfaces from the bottommost side of Fig. 6(a). By this modification, the additional results related to time parameter can be analyzed. A time distance between the occurrence of a bubble and that of a follow-up bubble are measured and denoted as td in

Fig. 6. This was followed by calculating bubble departure frequency by averaging the time distances and reversing the averaged time value as follows:

nd f d ¼ Pt

ð2Þ

d

where f d denotes the bubble departure frequency, and nd denotes the number of bubble occurrences. Furthermore, the time involved from the departure to complete condensation was calculated and denoted as tc .

176

H. Jo, D. Jo / Annals of Nuclear Energy 110 (2017) 171–185

3.1.2. Tridimensional Reconstructing Assumption After obtaining information from the PIB images, individual images at the detaching moments were specified for the Tridimensional reconstructing assumption. Prior to the analysis, it was assumed that the bubble has maximum volume at the detaching moment. The volume at this moment is defined as detaching volume. Fig. 7 shows the TRA process in detail and includes: (a) original image, (b) color-binarized image, (c) reconstructed image, and (d) 3-D view. In the selected images, the horizontal radii and centroids were found on each unit horizontal line. On each horizontal line, the left and right boundary of the bubble were detected. The radius of the bubble region on each unit line can be calculated as follows:

r¼

 d  nright  nleft 2dpx

ð3Þ

where nright and nleft denote the pixel numbers of the right and left boundaries, respectively. d denotes a diameter of the nozzle that is used as the reference length in the image, and dpx denotes the number of pixels in the nozzle region. Using the boundaries information of bubble region and the radius in each unit line, the centroid can be determined. The unit cylinder can be assumed with a height of a pixel, a basal plane of the specified radius and the centroid. Fig. 8 shows an example corresponding to these steps. The unit cylinders were laminated to reconstruct the tridimensional bubbles as shown in Fig. 7(d). The volume of detaching bubble can be assumed as follows:

1 X d pr2 1; 000 k¼1 dpx k y

V¼

ð4Þ

where y denotes the highest pixel number of the y-axis in the image, and rk denotes a radius of unit cylinder generated on the row. The volume reduction rate can be calculated by considering the maximum volume of the bubble and its condensation time. Herein, the volume reduction rate is defined as an averaged volume reduced by phase change per unit time. Prior to the calculation, it was assumed that the evaporation was negligibly small. Since, only condensation occurs after detachment until the bubble collapses, the volume reduction rate after detachment can be calculated with the ratio between the detaching volume determined by TRA and the condensation time(tc ) determined by PIB.

V_ ¼ V=t c

ð5Þ

3.2. Acoustic analysis Additionally, an acoustic analysis was performed to understand the condensation phenomena in the smooth bubble regime. A hydrophone was used to detect a variation of pressure waves that occurred due to condensation. Fig. 9 shows the acoustic signals and visual observations that phenomenologically occurred during the

condensing process. The blue dashed line indicates the acoustic signals in silent when there is no condensation. This waveform was dominated by a 60 Hz hum noises which were not affected by the boiling phenomena. Besides, the red solid line indicates the acoustic signals when there is condensation. The signals included not only the hum noises but some modifications caused by the condensation of bubbles. As shown in this figure, the acoustic signal changed when the condensation occurred. The amplitude and the frequency of the condensation signal was bigger than the waves in the silent condition. In here, the hum noises were not filtered to prevent possible loss of acoustic data. In the present study, a hydrophone TC4013 was used, and it consisted of a piezoelectric transducer that transferred the acoustic data to the analog voltage. In order to analyze the continuous data, acoustic data was acquired for 30 s at a sampling rate of 51,200 Hz. The acquired raw data is shown in Fig. 10. The acoustic signals possessed stronger amplitudes and rougher waveforms when a bigger bubble was injected or when the pool had a lower temperature. These results were expressed quantitatively by calculating the amplitudes with respect to effective values (RMS). However, it was not possible to physically match the analog data with sound pressure. Thus, a conversion process was applied for the units to obtain an appropriate physical meaning. Unit conversion was performed using a simple equation including underwater references, voltages, and distances between the hydrophone and sound source. Additionally, the sensitivity information of the piezoelectric sensor was required for converting units from voltage to sound pressure. In the case of the TC4013 hydrophone, the sensitivity of the piezoelectric sensor is calculated as follows:

SdB ¼ 211 dB re 1 V=lPa ¼ 20 log



SmV=Pa Ref V=lPa



SmV=Pa ¼ 0:0282 mV=Pa ¼ 2:82  1013 V=lPa

ð6Þ ð7Þ

where S denotes the sensitivity of sensor, and Ref V=lPa denotes the reference voltage at a micro-pascal corresponding to 1. Sound pressure level is a commonly used unit for sound that indicates relative values with respect to the response of human hearing. Thus, there are different pressure reference levels between airborne sounds (20 Pa) and underwater sounds (1 Pa) (Sengpiel, 2014). In the present study, the pressure reference level for underwater was selected. Fundamentally, the unit conversion equation from voltage to sound pressure level can be expressed as follows:

 LSP ¼ Ref dB SPL þ 20 log

Vin

SmV=Pa

 ð8Þ

where Ref dBSPL denotes the reference pressure level for atmospheric pressure which corresponds to 120 dB and Vin denotes indicated voltage.

Fig. 7. Process of tridimensional reconstruction; (a) Original image, (b) Color-binarized image, (c) Y-axial view of tridimensional reconstructed bubble, and (d) Diagonal view of tridimensional reconstructed bubble.

H. Jo, D. Jo / Annals of Nuclear Energy 110 (2017) 171–185

177

Fig. 8. Example of tridimensional assumption.

Fig. 9. Acoustic signals compared with the condensation phenomena.

However, the sound pressure level generally includes a reference distance of 1 m that implies that a sound source is at a distance of 1 m from the transducer. In contrast, the hydrophone was fixed at a distance of 0.05 m from the nozzle. Therefore, the reference distance was adjusted by using an additional term. As well as the reference adjustment, the sound attenuation in distance should be considered. The modified conversion equation can be expressed as follow:

 LSP ¼ Ref dB SPL þ 20 log

Vin

SmV=Pa



  20 log

l

lref

 ð9Þ

where l denotes the distance of the hydrophone from the nozzle (0.05 m), lref denotes the reference distance of the sound pressure level (1 m). 4. Results and discussion 4.1. Phenomenological results Figs. 11 through 13 show the images captured by a high-speed camera for non-condensable and condensable bubbles. Fig. 11

shows the rising aspects variation of non-condensable bubbles with different bubble sizes. Figs. 12 and 13 show rising and condensing aspects variations, respectively, of condensable bubbles with different subcooling degrees and bubble sizes. Behaviors of the rising bubbles could be distinguished based on bubble size. With respect to large bubbles, a body oscillation occurred due to the penetration of a tail through the bubble body and a significant shape transformation. With respect to small bubbles, a surface oscillation without any intense transformation occurred. Fig. 14 shows some forces related to these phenomena. In the cases of the non-condensable bubbles, the behaviors of the rising bubbles were examined with three different nozzle sizes. For this experiment, constant flow rate of general air was injected continuously, using a syringe pump. With respect to the nozzle size of 6.35 mm, the tail of the bubble penetrated through a bubble body following detachment from the nozzle. This was caused by elasticity that occurred on the neck between the detaching bubble and remaining steam in the nozzle. Following the penetration, the bubble shape was transformed by surface tension that maintained the bubble in a circular shape. Due to the differences of elasticity and surface tension, the bubble continued to oscillate up and down, and the whole body oscillation grew weaker with the rising

178

H. Jo, D. Jo / Annals of Nuclear Energy 110 (2017) 171–185

Fig. 10. Original acoustic signals.

Fig. 11. Rising aspects variation of non-condensable bubbles.

of the bubble. The nozzle sizes of 3.18 mm and 1.59 mm did not indicate any penetration because the effect of surface tension exceeded that of the elastic force. Thus, the elastic effect gradually disappeared due to surface oscillation on the surface of the bubble, and the bubble rose without any significant transformation. The observations with respect to non-condensable bubbles were also noted in the experiments with condensable bubbles. In

the case of the condensable bubbles, the behaviors of the rising bubbles were examined with three different nozzle sizes and six different subcooling degrees. Figs. 12 and 13 show the behaviors of the rising condensable bubbles. The primary difference between the non-condensable and condensable bubbles was that the volume continued to change for the condensable bubble, but the volume was constant for the non-condensable bubble. This was due to

H. Jo, D. Jo / Annals of Nuclear Energy 110 (2017) 171–185

179

Fig. 12. Rising and condensing aspects variation of condensable bubble at DT sub ¼ 5 K.

Fig. 13. Rising and condensing aspects variation of condensable bubble at DT sub ¼ 25 K.

the phase change and the temperature difference between the inside of the bubble and the surrounding liquid. With respect to the nozzle size of 6.35 mm, the tail of the condensable bubble penetrated through the body, but the bubble that detached from the nozzle size of 1.59 mm did not indicate any penetration. After the condensable bubbles were exposed to the subcooled water, they began condensing and completely collapsed. With respect to

the subcooling degree of 5 K, complete condensation with the bubble injected by a 6.35 mm nozzle corresponded to 38 ms, and complete condensation with the bubble injected by a 1.59 mm nozzle corresponded to 15 ms. With respect to the subcooling degree of 25 K, complete condensation with the bubble injected by 6.35 mm and 1.59 mm nozzles corresponded to 13 ms and 1.3 ms, respectively. The images were further analyzed using

180

H. Jo, D. Jo / Annals of Nuclear Energy 110 (2017) 171–185

indicate that the frequency of the bubble departure increased with decreases in the nozzle size. With the nozzle sizes of 6.35 mm and 3.18 mm, the bubble detaching frequencies remained as constants relative to the subcooling degrees. In contrast, with the nozzle size of 1.59 mm, the departure frequency increased more significantly with higher subcooling degrees. The reason why the bubble departure frequency varied with the different bubble sizes is interfacial tension between the nozzle and the bubble. When the smaller nozzle sizes were applied, the force that hold the bubble to the nozzle became weaker. The interfacial tension force can be determined by attraction of the liquid to the nozzle surface, and this can be calculated by simple equation as follow:

Fi ¼ din pr sin h

Fig. 14. Sketch of forces acting on shape of bubble.

image processing techniques to obtain a volume reduction rate based on subcooling degrees and bubble sizes. 4.2. Analysis results The quantitative results for the departure and the condensation phenomena were described as below. For the bubble detaching phenomena, the frequency of bubble departure and the bubble detaching volume were analyzed. For the bubble condensation phenomena, the bubble condensation time, rising distance, volume reduction rate and sound pressure level were analyzed. The errors for each results were analyzed as mentioned in Appendix A. 4.2.1. Results of bubble detaching phenomena The bubble at the detaching moment is affected by various forces such as buoyancy force, drag force, surface tension force, and so on (Eastman, 1984). The results of bubble departure frequency can be discussed based on these aspect. The variations of bubble departure frequencies are shown in Fig. 15, and the results

ð10Þ

where din denotes the inner diameter of nozzle, r denotes the fluid surface tension and h denotes the contact angle which goes to 90° as the bubble nears detachment. On the 1.59 mm nozzle, since the inner diameter was 0.79 mm, the interfacial force between the nozzle and the liquid can be calculated as 0.15–0.16 mN. On the 6.35 mm and 3.18 mm nozzles, the interfacial forces acted with 0.85–0.91 mN and 0.33–0.36 mN, respectively. Since the force that interrupts the detachment of bubble acted 2.25–5.78 times weaker at the smallest nozzle, the bubble occurred more frequently. Besides, the variation of the bubble departure frequency by the subcooling degree became more significant when the 1.59 mm nozzled was used. This phenomenon can be explained by following reason. Phase change occurs not only after the bubble detachment but during the growing and necking processes of bubble. Especially the phase change at the neck between the bubble and steam in the nozzle affected the detachment of the bubble directly. On the big nozzle, such as 6.35 mm and 3.18 mm nozzles, the additional gas can be injected during the bubble growing and necking processes. Thus, the detachment is delayed and the variation of bubble departure frequency with the change of the subcooling degree became small. However, on the nozzle size of 1.59 mm, the phase change at the neck is faster than the additional gas injection. Because of this reason, the bubble departure frequency varied more significantly on the smallest nozzle. The phase change at the neck of the bubble had an effect on the bubble volume at the detaching moment. As shown in Fig. 16, it was observed that the detaching volume decreased with increases in the subcooling degree. This indicates that the bubble was detached before it grew sufficiently, due to the condensation phenomena during the growing and necking processes of the bubble. On the nozzle size of 1.59 mm, the bubble had about 9.91 times bigger volume when the subcooling degree inceased from 1 to

Fig. 15. Experimental results of bubble departure frequencies.

H. Jo, D. Jo / Annals of Nuclear Energy 110 (2017) 171–185

181

Fig. 16. Experimental results of averaged detaching volumes.

25 K. Beside, on the nozzle size of 6.35 mm, the bubble had about 1.66 times bigger volume on the same condition. In summary, the interfacial tension force had an effect on the variation of the bubble detaching phenomena. In addition, the condensation phenomena on the growing and necking processes of the bubble affected the results.

4.2.2. Results of bubble condensation phenomena Figs. 17 and 18 show the results of the condensation time from detachment to complete condensation and the rising distance, respectively. A lower condensation time and a shorter rising distance were associated with a higher subcooling degree and a smaller bubble size. Besides, Fig. 19 shows the result of the averaged

Fig. 17. Experimental results of averaged condensation times.

Fig. 18. Experimental results of averaged rising distances.

182

H. Jo, D. Jo / Annals of Nuclear Energy 110 (2017) 171–185

Fig. 19. Experimental results of volume reduction rates.

volume disappeared by condensation per unit time, namely the volume reduction rates. The bubble volume was reduced faster when the liquid temperature decreased and the bubble size increased. These results can explained as follow. Since the bubble had the biggest volume at the biggest nozzle, the condensation time was longer than the others. Also, it can be observed that the volume reduction rate was fastest at the biggest nozzle. Naturally,

since the contact surface was biggest when the bubble had the biggest volume, the most mass condensed. Finally, the result of effective amplitudes gained by hydrophone is shown in Fig. 20. As shown in Fig. 20, increases in subcooling degree and larger bubbles resulted in increases in amplitude of acoustic signal. Calculating these results using Equation (9), the variation of the sound pressure level for different bubble sizes

Fig. 20. Experimental results of effective amplitudes of acoustic signals.

Fig. 21. Correlation between the sound pressure levels and the volume reduction rates.

H. Jo, D. Jo / Annals of Nuclear Energy 110 (2017) 171–185

and subcooling degrees can be known. The calculated sound pressure level was closely related to the volume reduction rate shown in Fig. 19. As shown in Fig. 21, the volume reduction rate increased with subcooling degree and bubble size, and thus the sound pressure level increased due to the vapor condensed more significantly in the bigger volume reduction rate. Based on this correlation, at the early moment of boiling, the charateristics of the bubble on the smooth bubble condensation regime can be predicted acoustically. Therefore, the possibilities of acoustically monitoring earlier boiling were obtained.

5. Summary and conclusions In this study, smooth bubble condensation was examined in a subcooled pool to understand condensation phenomenon at the beginning of boiling. The experimental apparatus was designed to investigate the behaviors of condensable and non-condensable bubbles for varying temperatures and bubble sizes. Visual and acoustic analysis methods were used to examine the condensation phenomena. With respect to the visual analyses, the bubble images were captured by a high-speed camera. The captured images were analyzed using Phase Interface Binarization (PIB) and Tridimensional Reconstructing Assumption (TRA) methods. The PIB method determined the bubble departure frequency, condensation time, and rising distance. The TRA method determined the departure bubble size and volume reduction rate. With respect to the acoustic analysis, the sound signals were obtained using a hydrophone. The obtained analog data were converted to sound pressure units. In the cases of the non-condensable bubbles, the variations in the rising behaviors of bubbles were investigated. With respect to the nozzle size of 6.35 mm, body oscillation occurred. The tail of the bubble penetrated through a bubble body following detachment from the nozzle due to elasticity between the detaching bubble and remaining steam in the nozzle. Following the penetration, the bubble shape was transformed by surface tension. The bubble continued oscillating up and down, and this body oscillation grew weaker as the bubble rose. With respect to the nozzle sizes of 1.59 mm and 3.18 mm, surface oscillation was observed. The bubble was not penetrated by its tail because the effect of surface tension exceeded that of elastic force. The elastic effect gradually disappeared due to surface oscillation on the surface of the bubble, and the bubble rose up without any significant transformation. In the cases of condensable bubbles, the variations in the rising behaviors as well as the variations in condensing behaviors were investigated and studied in detail. With respect to the nozzle size of 6.35 mm, the tail of a condensable bubble penetrates through the body. However, the bubble detached from the nozzle size of 1.59 mm did not exhibit any penetration. After a condensable bubble was exposed to the subcooled water, it began to condense and completely collapsed. The detailed condensation phenomena were analyzed by the various previously mentioned methods, including PIB, TRA, and acoustic analysis. The PIB method and the TRA method were used to analyze the results of the bubble departure frequencies and the bubble volumes at the detaching moments. The frequency of the bubble departure increased with decreases in the nozzle size, and the detaching volume decreased with increases in the subcooling degree. This implied that the bubble was detached before it grew sufficiently due to the interfacial tension force had an effect on the variation of the bubble detaching phenomena. Depending on the variation of the subcooling degree and the bubble size, the bubble departure frequency was changed. During these processes, the condensation phenomena on the growing and necking processes of the bubble affected the results.

183

In addition to the departure frequencies and the bubble volumes, the condensation times, rising distances and the volume reduction rates were analyzed. The results revealed that the condensation time and rising distance decreased with increases in the subcooling degree and decreases in the bubble size. This was because the bubble had the biggest volume at the biggest nozzle and thus the condensation time was longer than the others. Besides, the bubble volume was reduced faster when the liquid temperature decreased and the bubble size increased. Naturally, since the contact surface was biggest when the bubble had the biggest volume, the most mass condensed. The acoustic analysis indicated that there were differences in acoustic signals among the various condensation conditions. the volume reduction rate increased with increases in subcooling degree and bubble size, and thus the sound pressure level increased due to the vapor condensed more significantly in the bigger volume reduction rate. As a result, the possibilities of acoustically monitoring the earlier boiling were obtained. The sound signal differences focused on the acoustic amplitudes, it was necessary to consider wave interference to develop monitoring techniques. Several wave interferences occur due to continuous injecting or due to the structures in the test section such as the visualization windows, the nozzles, and the walls. Furthermore, it is necessary to transform the acoustic data from time-based units to frequency-based units to extract the condensing signals. Therefore, in order to increase the accuracy of the present study, future studies will involve an exploration of the following topics: (1) analysis of the acoustic signals under the frequency scale such as Faster Fourier Transform and Discrete Wavelet Transform, (2) a numerical study of the complex relations that exist between the parameters of condensation and rising phenomena, (3) acquiring experimental data including the mechanical noises and extracting the pure boiling signs. Acknowledgement This research was supported by Kyungpook National University Bokhyeon Research Fund, 2016. Appendix A. Uncertainty analyses In the current study, uncertainties of experimental results were analyzed for (1) subcooling condition, (2) PIB results, (3) TRA results, and (4) acoustic analysis results. Subcooling degrees contained errors from accuracy of thermocouple, accuracy of measurement equipments, and significant digit. The K-type thermocouple(KTSS-HH) had absolute error 2.2 °C and measurement system NE7000 had error 0.7 °C. Besides, since the subcooling degrees were accurate to two decimal places, the error 0.1 °C was considered. Therefore, the uncertainty of subcooling degrees were considered as 3.0 °C. PIB analyzing method had errors for length and time parameters. First, the error of length (ulength ) were determined with the pixel-to-pixel distance of captured images (ulength ¼ 2d=dpx ). Regardless of experimental conditions, the error of length in PIB results were approx. 0.06 mm. Using this error, the relative error of bubble rising distance can be calculated as follow:

urising distance ¼

ulength h

ðA:1Þ

The relative errors of bubble rising distance were less than 5.49%. The error of time (utime ) were determined with the reciprocal of the sampling rates of high-speed camera (utime ¼ 2=SR). Using this

184

H. Jo, D. Jo / Annals of Nuclear Energy 110 (2017) 171–185

error, the error of condensation time and that of bubble departure frequency can be calculated as follow:

ucondensation time ¼

utime tc

udeparture frequency ¼

ðA:2Þ

utime td

ðA:3Þ

The relative errors of bubble condensation time and bubble departure frequency were less than 0.80% and 6.55%, respectively. The reliability of TRA analyzing method was determined indirectly comparing experimental data of non-condensable bubble. The volume analyzed from TRA process were compared with the injected volume by syringe pump. The real volume (Vinjected ) was calculated with a volumetric flow rate set in the syringe pump (Q ) and a injected time. When the non-condensable bubble was injected, the time between the departure of a bubble and that of a follow-up bubble (td;air ) was determined by the captured images:

Vinjected ¼ Q  td;air

ðA:4Þ

The error of injecting volumetric flow rate were determined with the accuracy of the syringe pump NE300 (0.1%). The error of time distance was determined with the reciprocal of the sampling rates of high-speed camera as metioned before. By the error propagation, the error of injected volume can be calculated as follow:

uinjected volume

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  2ffi utime 2 ¼ ðupump Þ þ  100 td;air

ðA:5Þ

Comparing the TRA results with the injected volume, the uncertainty of TRA analyzing method can be calculated:

uvolume ¼

  Vinjected  Vanalyzed  100 þ uinjected Vinjected

ðA:6Þ

The relative errors of bubble detaching volume at the nozzle of 6.35 mm, 3.18 mm and 1.59 mm were 0.61%, 1.58% and 4.13%, respectively. Propagating the errors of volume and condensation time, the error of volume reduction rate can be calculated as follow:

uVRR ¼

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðuvolume Þ2 þ ðucondensation time Þ2

ðA:7Þ

The relative errors of VRR were less than 4.93%. All calculations of VRR errors for each variables were shown in Table. 2. According to the table, the error became bigger when the smaller nozzle used, because the ratio between the pixel-to-pixel distance and the noz-

Table 2 Error calculation results. Nozzle

Temp

VRR

uVRR

SPL

uSPL

1.59 mm 1.59 mm 1.59 mm 1.59 mm 1.59 mm 1.59 mm 3.18 mm 3.18 mm 3.18 mm 3.18 mm 3.18 mm 3.18 mm 6.35 mm 6.35 mm 6.35 mm 6.35 mm 6.35 mm 6.35 mm

1.00 5.00 10.00 15.00 20.00 25.00 1.00 5.00 10.00 15.00 20.00 25.00 1.00 5.00 10.00 15.00 20.00 25.00

0.42 1.51 4.14 4.65 5.07 5.50 1.94 3.99 6.84 7.68 8.79 11.31 4.45 7.24 10.07 15.01 17.72 20.21

4.14 4.17 4.32 4.39 4.56 4.93 1.59 1.59 1.61 1.63 1.65 1.68 0.62 0.62 0.62 0.63 0.64 0.65

272.37 273.53 273.65 274.10 274.94 275.91 273.08 273.70 274.04 274.71 275.09 276.33 273.51 273.91 274.67 275.55 275.98 277.54

0.66 0.58 0.57 0.54 0.49 0.44 0.61 0.57 0.55 0.51 0.49 0.42 0.58 0.56 0.51 0.46 0.44 0.37

zle outer-diameter became bigger. In this reason, the relative error of length increased. For accurate result, it is recommended that the captured image should be blowed up for the experiments of the smaller nozzle. The uncertainty of acoustic analysis was calculated with the parameters used in Eq. (9), such as acquired voltage (Vin ), sensitivity of hydrophone (SmV=Pa ) and distance between the nozzle and the hydrophone (l). First, the uncertainty of voltage was determined with the accuracy of DAQ NI 9234 (uvoltage ¼ 1:90%). Besides, the sensitivity of hydrophone varies according to the frequency of sound. Because the effects of frequency were not considered in the present study, the variation of sensitivity affected from frequency regarded as an error (uhydrophone ¼ 1:42%). Finally, the error of distance between the nozzle and the hydrophone was determined. The hydrophone has 9.5U head diameter and a hole that the hydrophone was inserted has 12U diameter. Because of this difference, the error of distance was occurred (udistance ¼ 2:50%). In conclusion, the uncertainty of sound pressure level can be calculated as follow:

uSPL ¼

20  SmV=Pa Vin  l

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðuhydrophone Þ2 þ ðuvoltage Þ2 þ ðudistance Þ2

ðA:8Þ

The relative errors of SPL were less than 0.66%. All calculations of SPL errors for each variables were expressed in Table 2. According to the Table 2, the relative errors of this parameter are fairly small. However, according to the Fig. 21, the error bar seems quite big. When the errors were converted to the absolute errors, the errors in range of 0.37% to 0.66% became in range of 1.02 dB to 1.81 dB. For more accurate experiment, it is recommended that the distance between the hydrophone and the injecting nozzle should be adjusted using additional fixture and the error of hydrophone distance should be reduced.

References Ajbar, A., Al-Masry, W.A., Ali, E.M., 2009. Prediction of flow regimes transitions in bubble columns using passive acoustic measurements. Chem. Eng. Process. Process Intensif. 48 (1), 101–110. Al-Masry, W.A., Ali, E.M., Aqeel, Y.M., 2006. Effect of antifoam agents on bubble characteristics in bubble columns based on acoustic sound measurements. Chem. Eng. Sci. 61 (11), 3610–3622. Al-Masry, W.A., Ali, E.M., Al-Kalbani, M.N., 2007. Prediction of regime transitions in bubble columns using acoustic and differential pressure signals. Chem. Eng. J. 133 (1), 139–149. Chicharro, R., Vanquez, A., 2014. The acoustic signature of gas bubbles generated in a liquid cross-flow. Exp. Thermal Fluid Sci. 55, 221–227. Eastman, R.E., 1984. Dynamics of bubble departure, AIAA 1984. Thermophysics, 1–5. Faghri, A., Zhang, Y., 2006. Transport Phenomena in Multiphase Systems. Elsevier. Geraldo, I.C., Bose, T., Pekpe, K.M., Cassar, J., Mohanty, A.R., Paumel, K., 2014. Acoustic monitoring of sodium boiling in a liquid metal fast breeder reactor from autoregressive modes. Nucl. Eng. Des. 278, 573–585. Hetsroni, G., 1982. Handbook of Multiphase Systems. Hemisphere Publishing Corporation and McGraw-Hill. Jeon, S., Kim, S., Park, G., 2011. Numerical study of condensing bubble in subcooled boiling flow using volume of fluid model. Chem. Eng. Sci. 66, 5899–5909. Lee, H., 2011. Literature study on detection of nucleate boiling and signal process in engine cooling system. BorgWarner Therm. Syst. Leighton, T., 1997. The Acoustic Bubble. Academic Press. Minnaert, M., 1933. On musical air-bubbles and the sounds of running water, 7 16. Philos. Mag. Ser. 7 16 (104), 235–248. Nishihara, H., Bessho, Y., 1977. Acoustic emission in subcooled nucleate pool boiling, 14. J. Nucl. Sci. Technol. 14 (6), 407–415. Osterman, A., Hocevar, M., Širok, B., Dular, M., 2009. Characterization of incipient cavitation in axial valve by hydrophone and visualization. Exp. Therm. Fluid Sci. 33 (4), 620–629. Pan, L., Tan, Z., Chen, D., Xue, L., 2012. Numerical investigation of vapor bubble condensation characteristics of subcooled flow boiling in vertical rectangular channel. Nucl. Eng. Des. 248, 126–136. Sengpiel, E., 2014. Sound pressure level, pressure, and intensity level conversion, Forum für mikrofonaufnahmetechnik und tonstudiotechnik, http://www. sengpielaudio.com/. Staudenraus, J., Elsenmenger, W., 1993. Fibre-optic probe hydrophone for ultrasonic and shockwave measurements in water. Ultrasonics 31 (4), 267–273.

H. Jo, D. Jo / Annals of Nuclear Energy 110 (2017) 171–185 Tang, J., Yan, C., Sun, L., 2015a. A study visualizing the collapse of vapor bubbles in a subcooled pool. Int. J. Heat Mass Transf. 88, 597–608. Tang, J., Yan, C., Sun, L., 2015b. Feature of acoustic sound signals involved in vapor bubble condensation and its application in identification of condensation regime. Chem. Eng. Sci. 137, 384–397. Tinguely, M., 2013, The effect of pressure gradient on the collapse of cavitation bubbles in normal and reduced gravity (Ph.D thesis). École Polytechnique Fédérale De Lausanne.

185

Vanquez, A., Sanchez, R.M., Salinas-Rodriguez, E., Soria, A., Manasseh, R., 2005. A look at three measurement techniques for bubble size determination. Exp. Therm. Fluid Sci. 30 (1), 49–57. Vanquez, A., Manasseh, R., Chicharro, R., 2014. Can acoustic emissions be used to size bubbles seeping from a sediment bed? Chem. Eng. Sci. 131, 187–196.