- Email: [email protected]
Observations of acoustically generated cavitation bubbles within typical fluids applied to a scroll expander lubrication system
Experimental Thermal and Fluid Science 35 (2011) 1544–1554
Contents lists available at SciVerse ScienceDirect
Experimental Thermal and Fluid Science journal homepage: www.elsevier.com/locate/etfs
Observations of acoustically generated cavitation bubbles within typical fluids applied to a scroll expander lubrication system I. Tzanakis a,⇑, M. Hadfield a, I. Henshaw b a b
Sustainable Engineering Research Centre, Bournemouth University, Talbot Campus BH12 5BB, UK Energetix Group Limited, Capenhurst Technology Park, Chester CH1 6EH, UK
a r t i c l e
i n f o
Article history: Received 5 May 2011 Accepted 18 July 2011 Available online 31 July 2011 Keywords: Cavitation Reynolds Weber number Viscosity Surface tension Refrigerant Lubricant Scroll
a b s t r a c t An experimental study to evaluate the dynamic performance of three different types of cavitation bubbles is conducted. An ultrasonic transducer submerged into the working fluids of a scroll expander is utilised to produce cavitation bubbles and a high speed camera device is used to capture their behaviour. Three critical regions around the ultrasonic source, between the source and the solid boundary, and across the solid boundary were observed. Experimental results revealed that refrigerant bubbles sustain a continuous oscillatory movement, referenced as ‘‘wobbling effect’’, without regularly collapsing. Analytical results indicate the influence of several factors such as surface tension/viscosity ratio, Reynolds number and Weber number which interpret that particular behaviour of the refrigerant bubbles. Within the refrigerant environment the bubbles obtain large Reynolds numbers and low Weber numbers. In contrast, within the lubricant and the water environment Weber number is significantly higher and Reynolds number substantially lower. The bubble radius and velocity alterations are accurately calculated during the cavitation process. Lubricant bubbles achieve the highest jet velocity while refrigerant bubbles having the lowest jet velocity are not considered as a destructive mean of cavitation for scroll expander systems. Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction Scroll expanders have been a key component of many industrial systems over the last 20 years (CHP systems, air-conditions, pumps, etc.). A specific scroll expander was tested for 1000 h as a part of a small domestic combined heat and power (CHP) system. Cavitation was identified as a wear mechanism during the investigation of the scroll’s components [1]. The cavitation mechanism developed inside the scroll is caused by the operational fluid environment [2]. This environment includes two fluids; a high molecular organic refrigerant and a synthetic lubricant. The refrigerant, as a gas form, drives the scroll while the lubricant protects the parts of the scroll from excessive wear. Thus the present study focuses on these two scroll fluids and in the dynamic behaviour of their cavitation bubbles investigating their engineering implications to scroll expander systems or similar automotive industrial units. Limited experimental information concerning lubricant and refrigerant bubble formation and their dynamic behaviour [2–6] is available within the published literature. The influence of viscosity and surface tension upon cavitation threshold has led to some debate within published literature, with claims of increased or decreased cavitation activity arising from
⇑ Corresponding author. E-mail address: [email protected] (I. Tzanakis). 0894-1777/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.expthermflusci.2011.07.005
different studies of similar environments. Brujan et al. [7] in polymer aqueous solutions has shown that higher levels of viscosity can substantially mitigate the tendency of cavitational activity to damage. Popinet and Zaleski [8] in his study has found that the amplitude of the oscillations decreases due to viscous damping, while the jet impact velocity decreases as viscosity increases. Williams et al. [9] and Berker et al. [10], reported a mitigating effect of viscosity in the jet propulsion by the bubbles collapse of various fluids. Karunamurthy et al. [11] presents the viscosity influence of various lubricants on erosion rate, showing that as the viscosity increases the erosion decreases. In contrast, Meged et al. [4] after a series of experiments using 20 different types of liquid lubricants, has found that there is no direct correlation between the viscosity of a liquid and the cavitation erosion mechanism. In regards to surface tension in a micro-scale level, the bubble growth and the elongation are influenced by the surface tension forces which are dominant at this scale [12]. According to Iwai and Li [13], experiments conducted in various water solutions with different surface tensions, showed that surface tension increases the liquid-jet impact and the bubble size. In contrast Liu et al. [14] has shown that in a cavitation environment higher surface tension values leads to smaller sized bubbles reducing the collapse duration. In the present study cavitation tests were utilised to analyse and evaluate the dynamic behaviour of the cavitation bubbles generated within the refrigerant and the lubricant environment of a scroll expander system. Distilled water was also used as a
1545
I. Tzanakis et al. / Experimental Thermal and Fluid Science 35 (2011) 1544–1554
reference fluid because of its cavitation bubbles well known behaviour. Bubble behaviour was studied with the use of high speed photography and three distinctive patterns of cavitation bubbles were monitored. In the first pattern the generation mechanism of cavitation bubbles was investigated during their incubation time around the ultrasonic source. Therefore, the bubbles’ behaviour was monitored during their descent to the solid surface and across the solid boundary, revealing two more interesting patterns. Analytical results indicate the cavitation behaviour of all the tested bubbles according to their viscous and surface tension forces. The unusual behaviour of the refrigerant bubble was interpreted using combinations of Reynolds and Weber number while the bubble radius and velocity alterations were calculated during the cavitation process for all the liquid environments. 2. Experimental procedure The experimental apparatus is described in Fig. 1. The experiments were performed based on the ASTM G32-03 standard method. The ASTM method proposes a standard probe of 15.9 mm diameter to be generally used for cavitation tests. However, in this study a probe of 5 mm was deployed to produce cavitation bubbles for more definitive images. Experiments were carried out by using an ultrasonic transducer at a frequency of 20 kHz. The ultrasonic horn has a plane surface oscillating in a simple harmonic motion. Peak to peak amplitude of the horn tip was adjusted to be at 50 lm. The sample consists of a standard chromium steel ball EN-31, polished in a few microns accuracy and implemented on a Bakelite base. The Bakelite base of the sample was mounted on the bottom of a small transparent square beaker with a maximum capacity of 500 ml. The tip of the ultrasonic horn was submerged into the liquid environment of the beaker in a distance of 0.5 mm from the surface of the chromium steel. The liquids used for the tests were distilled water, a synthetic lubricant and a high molecular organic refrigerant. All the liquids were tested in a room temperature of around 23 ± 2 °C and their physical properties used for the experiments are listed in Table 1. The viscosity and surface tension for all the three fluids were measured in the ambient environmental conditions of the laboratory using appropriate experimental devices. A high speed camera (Phantom v7.3) was installed at one of the sides of the beaker. At the beaker’s opposite side a high intensity fibre optic light source (2 kW) was deployed to provide sufficient light for investigation of critical areas. The Phantom v7.3 supplies a full frame (800 600 pixels) recording rate, with a maximum
Table 1 Physical properties of experimental liquids at 23 ± 2 °C. Fluid
Water
Refrigerant
Lubricant
Ultrasonic velocity inside liquid (m/s) Liquid density (kg/m3) Viscosity (mPa s) Surface tension (mN/m) Vapour pressure (Pa) at 20 °C
1476 1000 1 62–65 2230
990 625 0.225 32–35 53,300
1700 995 30 51–54 1300
shutter speed of up to 500,000 frames per second with 1 l exposure time. In this study the speed modes of the camera used for the photography varied between 30.000 fps and 220.000 fps with a 10 ls exposure time. During the tests a Leica mono zoom amplifier was used in front of the camera achieving a resolution of approximately 8 lm/pixel. The images focused on three critical regions; on the tip of the ultrasonic probe where the bubbles are generated, the region across the lower boundary wall where bubbles are destructively collapsed and at the area in between and such methodology was performed for all the liquids tested. An evaluation of various bubble images using the camera software was conducted and a range of graphs was produced. In each of the liquid environments more than 20 different bubble behaviours were monitored. Three distinctive patterns of bubbles’ typical behaviour were finally determined. 3. Experimental results 3.1. Region around the ultrasonic source The bubbles’ generation and the expansion mode on the tip of the horn revealed three different behaviours for each of the liquid environments respectively. A typical isolated incubation bubble was monitored in each liquid environment (Figs. 2–4). When water was used, the deformation of the water bubble started after 189 ls. In the case of the liquid refrigerant it was measured at 90 ls while for the lubricant was at 315 ls. Furthermore the time step where the jet stream was generated, was at 749 ls and at 795 ls for the water and the lubricant respectively. Interestingly, for the refrigerant a similar liquid jet stream was never formed during the first 1350 ls. The refrigerant bubbles maintain a robust shape avoiding the bubbles which are generated to separate, escape and finally collapse. Specifically, in the case of the distilled water Fig. 2, a typical cavitation mechanism is observed. The growth and the implosion
Ultrasonic Processor Computer/Camera software
Experimental Beaker Fiber Optic light-source High-speed Camera
Test Sample Fig. 1. Schematic of experimental apparatus
1546
I. Tzanakis et al. / Experimental Thermal and Fluid Science 35 (2011) 1544–1554
Tip of the ultrasonic horn
Water Jet Stream 189µs
720µs
749µs
Micro bubbles
813µs
1363µs
1135µs
1609µs
Fig. 2. Expansion mode of an acoustic cavitation refrigerant bubble on the tip of the horn. (Frame size is 1 0.5 mm).
Micro-jet Tip of the ultrasonic horn
315µs
Counter jet formation 465µs
450µs
960µs
Pancake Effect
360µs
Lubricant Jet Stream
795µs
1125µs
Toroidal Bubble
405µs
885µs
1305µs
1260µs
Fig. 3. The merging mechanism of the water-lubricant bubbles. (Frame size is 0.5 0.5 mm.)
Tip of the ultrasonic horn
15µs
165µs
135µs
Pancake Effect
Toroidal Bubble
90µs
Counter jet formation
210µs
120µs
225µs
Microbubbles jelly cloud 240µs
370µs
760µs
Fig. 4. The merging mechanism of the refrigerant bubble. (Frame size is 0.5 0.5 mm.)
1350µs
1547
I. Tzanakis et al. / Experimental Thermal and Fluid Science 35 (2011) 1544–1554
of the bubbles due to pressure variations creates a distinctive jet stream of micro-bubbles (frames 4–5, Fig. 2) which in turn can damage the material surface by pressure impact. After a 1000 ls a cloud of micro-bubbles is generated (frames 6–8) with a radius 10–12 times smaller than the primary bubble. The micro-bubble cloud exists continuously while over time becomes asymmetric (frames 5–8, Fig. 2), discharged fairly far downstream from the tip of the horn. Further downstream, as the pressure rapidly increases, the cavitation clouds rapidly divide into small bubbles and then disappeared [15]. In the lubricant environment a ‘‘pancaking’’ effect at the site of the collapsed bubble can be observed (frames 2–3, Fig. 3). The ‘‘pancaking’’ shape is formed only when the initial bubble is very close to the wall during the incubation period of time. After this phase the bubble rapidly obtains a toroidal shape (frame 4, Fig. 3). Then a micro-jet is formed, which is usually difficult to identify in this ‘‘pancaking’’ effect mode, leading eventually to a counter jet generation (frames 5–6, Fig. 3). As the bubble rebounds and grows, in an almost round spherical shape (frame 9, Fig. 3), it then rapidly collapses (frame 11, Fig. 3). A linear jet stream, which comprised of a cloud of micro-bubbles (up to 30 times smaller than the water ones), exerts from the centre of the collapse bubble (frame 7, Fig. 3). Interestingly within the lubricant environment part of the micro-bubbles instead of disappeared during their downstream are reformed to a thin foaming layer and spread along the surface of the solid boundary. The foaming layer plays the role of a cushion, significantly reducing the destructive power of the micro-jet impact [16]. In the case of the refrigerant, the spherical bubble initially is formed to a ‘‘pancaking’’ shape (frame 3, Fig. 4) followed by a counter jet (frames 4–5, Fig. 4). Then it can be clearly seen that loads of micro-bubbles are connected in a ‘‘jelly’’ cloud shape (frames 9–12, Fig. 4). In contrast to the previous cases (water, lubricant) during the cavitation process in the liquid refrigerant there is no bubble diffusion. Molecules cannot overcome the surface tension and other cohesive forces maintaining their bubbly shape. The individual bubbles cannot resist the attractive forces creating strong bonding in a single bubbly body. Regular implosions are unlikely to be observed.
3.2. Region between source and solid surface During their descent to the boundary wall, the lubricant and water bubbles merged or split due to pressure variations (Fig. 5). Thus, when the pressure increases, the bubble implodes and then is divided into smaller bubbles. When the pressure decreases, their radius increases and they eventually merged to a single bubble (frame 4, Fig. 5). In the case of the refrigerant, the adjacent unstable bubbles, present a distorted shape leading to significant alterations in their volume and size. Larger bubbles attract and absorb the smaller ones in the vicinity. The bubbles within the refrigerant solution have a more vigorous behaviour and their tendency to merge is higher than in the other two liquids. Frame 4 in Fig. 6 captures
Substrate
40µs
the attachment mechanism of two amoeboid shaped refrigerant bubbles. The adjacent bubbles are connected leading temporarily to an elongated bubbly shape prior to obtaining a single amoeboid shaped bubble. The pressure variations generate a wobbling behaviour of the refrigerant bubbles, referenced as ‘‘wobbling effect’’ affecting their movement, their morphology (amoeboid shape), their implosion rate (Section 3.3) and their diffusion to the liquid. The frequency of the implosions from the refrigerant bubbles is substantially lower comparing to the other two liquid environments, while a lack of consistency is recorded during their collapse (Fig. 12). Observing the two different bubble merging mechanisms, is evident that lubricant and water bubbles maintain a spherical shape across the bubbles periphery with a significant increment on their size. In the case of the refrigerant, an amoeboid formation is maintained, while the size of the bubble after merging can be further reduced.
3.3. Region near the solid surface boundary The bubbles implosion near the rigid boundary revealed three distinctive behaviours for each of the tested liquids respectively. In the water environment, a typical behaviour of a cavitation bubble is described. The steps of the collapse mechanism are briefly presented in Fig. 7 since previous researchers have analysed the phenomenon and its dynamic performance in close proximity to the boundary wall [17–24]. The life cycle of the bubble can be determined by the acoustic cycles. During one acoustic cycle the bubble grows up reaching its maximum diameter and then collapses. The attractive Bjerknes forces deform the spherical bubble to a toroidal bubble (frame 4, Fig. 7), while a micro-jet is formed and instantly a counter-jet emits a powerful shock wave (frame 5, Fig. 7). Then the bubble becomes unstable and is separated in two parts (frame 6, Fig. 7) which individually generate two micro-bubble clouds, the so-called splashing effect (frame 7, Fig. 7). Thereafter the two clouds merge into one (frame 8, Fig. 7), increasing in size and collapse releasing high velocity destructive pressure impacts to the boundary wall, damaging the material surface. The maximum pressure emitted from the bubbles at the cloud centre becomes much higher than that of a single bubble [25]. Regarding the lubricant, the bubbles have a similar behaviour to that of water (Fig. 8). They were generated in low pressure areas and then they collapsed in regions where the pressure is higher. It is worth noting that the foaming layer generated during the cavitation process (frames 2–3, Fig. 8) can affect the destructive impact of the larger bubbles’ collapse mechanism. The thin foaming layer consists of micro-bubbles with a radius less than 3 lm (frame 2, Fig. 9) and it can protects the material surface by acting as a cushion, absorbing and reducing the impact of the implosion. Thus, despite the fact that the average jet velocity from the lubricant bubbles with an average radius of 50–100 lm is higher than in the water (Section 4), the actual damage is considerably less. The results are in a very good agreement with Tzanakis et al. [16] experimental study where various steel materials under
50µs
120 µs
Fig. 5. Characteristic behaviour of an acoustic cavitation water bubble near the solid boundary. (Frame size is 1 0.5 mm.)
1548
I. Tzanakis et al. / Experimental Thermal and Fluid Science 35 (2011) 1544–1554
Boundary Wall
22µs
44µs
88µs
110µs
66µs
Amoeboid shape 77µs
121µs
Fig. 6. Collapse of an acoustic cavitation lubricant bubble near the solid boundary. (Frame size 0.5 0.5 mm).
Pancake Effect
Spherical Bubble
Boundary Wall Counter-Jet
110µs
Toroidal Bubble
140µs
Micro-Jet
160µs
Shock Wave
170µs
Two Individual Bubbles
200µs
Microbubbles Cloud
234µs
Destructive Collapse
370µs
Fig. 7. Lubrication layer formed across the surface of the rigid boundary, prior to bubble’s implosion. (Frame size is 0.5 0.5 mm.)
Foaming Layer
11µs Substrate
Substrate
Micro-jet formation
22µs
34µs Bubble implosion
89µs
100µs
111µs
Substrate
121µs
Fig. 8. The ceaselessly wobbling movement of an acoustic cavitation refrigerant bubble near the solid boundary. (Frame size 0.5 0.5 mm.)
1549
I. Tzanakis et al. / Experimental Thermal and Fluid Science 35 (2011) 1544–1554
Lubrication Film Thickness
20µs
Subsurface
70µs
50µs
Fig. 9. The different shapes the cavitation refrigerant bubble obtains in its maximum radius. (Frame size 0.5 0.5 mm.)
lubrication, refrigerant and water environment were tested. Tzanakis found that the morphology evolution of the cavities produced across the sample’s surface in a lubricant environment is significantly lower in comparison to the water. The refrigerant bubbles have by far the smallest cavitation pressure impact. When refrigerant was used, bubbles had an interesting behaviour across the surface of the solid boundary (Figs. 10 and 11). The refrigerant bubbles sustain a seemingly random dual motion. Small vigorous oscillations across the bubble surface are accompanied with a wobbling behaviour (‘‘wobbling effect’’); most of the refrigerant bubbles wobbled instead of collapsing. Eventually the refrigerant bubbles obtain an amoeboid shape (frame 2, Figs. 10 and 11). This unusual behaviour of the refrigerant bubbles significantly increases their lifecycle reducing the probability of a destructive impact. The bubbles instead of growing and collapsing at regular time steps are driven by the ‘‘wobbling effect’’ for a prolonged period of time prior to their collapse. A typical series of images in Fig. 10 show a refrigerant bubble prior to its collapse. The bubble after reaching its maximum size (frame 1, Fig. 10), instead of starting the collapse process it follows an oscillating motion. Even after 700 ls the size of the bubble
remains the same and there is a total absence of implosions. The shape slightly changes to similar amoeboid forms. Additionally, Fig. 11 captures another refrigerant bubble with more distinct phases of an amoeboid formation. Immediately after the bubble attained its maximum size, it is moving within the liquid refrigerant without collapsing. In frame 4 a separation mechanism appears. However, the attraction forces of the bubble restrain this phenomenon, keeping the bubble in a robust amoeboid single shape. In frame 10 the bubble is elongated reaching its maximum length and then returns to a bubbly round shape instead of collapsing. The interaction of the surface tension and the viscosity forces of the refrigerant liquid can explain the behaviour of the refrigerant cavitation bubbles. It has been reported in previous studies that viscous forces giving rise to perturbations tend to prevent the breakup of the cavitation bubbles while surface tension tends to prevent bubbles from collapsing [26]. In the present study the surface tension forces prevail the viscous forces within the liquid refrigerant maintaining a bubbly wobbling pattern among the cavities, preventing them from collapsing. The opposite phenomenon observed within the water and lubricant solutions, as viscous forces have a major role leading the bubbles to consecutive
Direction
Bubble’s maximum size
214µs
Amoeboid shape
36µs
72µs
108µs
247µs
377µs
400µs Direction
468µs
514µs
526µs
Boundary Wall
Fig. 10. Variation of water, refrigerant, lubricant cavitation bubbles radius with respect to time.
682µs
1550
I. Tzanakis et al. / Experimental Thermal and Fluid Science 35 (2011) 1544–1554
Direction
Amoeboid shape
140µs
270µs
60µs
30µs
Boundary Wall
170µs
290µs
90µs
210µs
240µs
310µs
340µs
Fig. 11. Variation of water, refrigerant, lubricant bubbles velocity and jet velocity with respect to time.
Bubble Radius ( µ m)
140 120 100
Water Refrigerant Lubricant
80 60 40 20 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170
Time (µsec) Fig. 12. Variation of water, refrigerant, lubricant bubbles distance from the boundary wall with respect to time.
oscillations, separating the bubble body into smaller cavities and this in turn generates regular implosions among the bubbles. Specifically, the refrigerant has the highest surface tension/viscosity ratio overall k = 155 (table 1), indicating a higher resistance to the breaking and collapsing mechanism. The above behaviour also enhances the probability of the refrigerant bubbles to form a ‘‘jelly’’ cloud shape pattern (Fig. 4). A similar behaviour was also noted by Schneider et al. [27] in his latest study, using another type of refrigerant flowing in micro-channels, where the bubbly flow morphology of low surface tension liquid in a fully developed hydrodynamic cavitating flow was highlighted. In contrast, the water and the lubricant have a much smaller surface tension/viscosity ratio of k = 65 and k = 1.8 respectively indicating a prone transition to a toroidal shape. Especially, in the case of the lubricant this is more obvious. The viscous forces affect radically the performance of the lubricant bubbles reducing their momentum (Fig. 13). Similar results were also found by Mishra and Pelesa [28] in his recent work using de-ionised water with substantially higher surface tension and viscosity values to the refrigerant solutions. Mishra showed
the disturbances a water bubble’s body experiences and the tendency of the bubble to implode according the surface tension and viscosity values during the cavitation process. Additionally, the size of the refrigerant bubbles was found to be the largest among the tested bubbles with an average bubble radius around 60–90 lm followed by the water and lubricant bubbles with an average radius of 30–50 lm and 20–45 lm respectively. This implies that the refrigerant bubbles with higher surface tension/viscosity ratio will have more probability to grow to larger sizes by avoiding early breaking-up into much smaller bubbles, during the growing process, in contrast to the lubricant bubbles where their size is substantially smaller. This also explains the hundreds of micro-bubbles which are formed across the foaming protective layer. 4. Analytical results The three characteristic bubble transition patterns, revealed experimentally, are interpreted using analytical methods. Typical
1551
10 9 8 7 6 5 4 3 2 1 0
Water Lubricant Ref Jet
30
Refrigerant Water Jet Lub Jet
25 20 15 10 5 0
10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170
Average Jet Velocity (m/sec)
Bubble Velocity (m/sec)
I. Tzanakis et al. / Experimental Thermal and Fluid Science 35 (2011) 1544–1554
Time (µsec) Fig. 13. Schematic of experimental apparatus.
behaviour of a similar radius cavitation bubble from its rising moment until its final collapse stage in each of the liquid environments is analysed. Fig. 12 presents the fluctuations of the bubble radius over the time for each of the liquids environments. Lubricant and water bubbles after their first implosion take the rebound by re-forming into smaller sizes and afterwards they implode again and reform. At the level of the boundary wall they reach their maximum size before the final collapse. The refrigerant bubbles keep linear fluctuation behaviour on their size while their radius has not been significantly altered since the wobbling behaviour prevails. The only implosion is observed at the level of the boundary wall where refrigerant bubble size is at its maximum. Fig. 13 shows the bubble’s velocity and the average jet velocity during the descent of the bubbles to the boundary wall. The bubble velocity was calculated during the expansion and the shrinkage mode of the bubbles prior to their final collapse onto the solid boundary. The average jet velocity is estimated by the assumption that the jet flow develops during the shrinkage of the bubble from its maximum radius. The average jet velocity of the bubble was estimated, as Claus–Dieter Ohl suggests in his study, using following equation [29]:
ujetðaveÞ ¼
2Rmax Tc
ð1Þ
Distance from the Boundary (µm)
where Rmax is the maximum radius of the bubble and Tc is the time collapse. The lubricant and water bubbles show significant alterations in their velocity following their radius fluctuation patterns. The lubricant bubbles achieve the highest bubble and jet velocity while the refrigerant bubbles implode only once having the lowest jet velocity among the tested bubbles. This happens close to the boundary wall. Refrigerant bubbles keep a constant velocity during their entire route reaching their maximum speed just before the
implosion with the surface. The velocity achieved by the refrigerant bubbles has the lowest impact potential. Fig. 14 shows the variation of boundary distance over the time as each of the bubbles approaches the boundary wall. Water bubbles, according to the inclination of the graph, have the fastest descent to the solid surface achieving an average speed of 2 m/s. The refrigerants bubbles follow with an average descending speed of 1.25 m/s. The lubricant bubbles are the last having the smallest descending speed of 0.6 m/s within the liquid lubricant. The viscosity in the lubricant is the highest within the liquids, decelerating the movement of the bubbles. Thus even if the bubble velocity and the average jet velocity of the lubricant is the highest among the tested liquids (Fig. 13), the motion is seriously affected by the viscous environment reducing bubble’s momentum. In contrast, the refrigerant bubbles having the lowest viscosity move to a faster pace from the lubricant bubbles but not from the water bubbles. The absence of significant oscillations in correlation with the ‘‘wobbling effect’’, highly affect the refrigerant bubbles’ motion reducing their momentum. Concomitantly, it is reasonable to report that the combination of the surface tension with the fluid viscosity properties of the bubbles (table 1) seriously affect the cavitation mechanism. The interpretation of the refrigerant bubbles’ unusual behaviour lies on the values of Reynolds and Weber number for each time step
Re ¼
qud qu2 d and We ¼ l r
where q is the fluid density, u is the bubble’s velocity, d is the diameter of the bubble in each time step, l is the fluid viscosity and r is the surface tension. Investigating the behaviour of a typical single refrigerant bubble according to the experimental analysis and the analytical
350 Water Refrigerant Lubricant
300 250 200 150 100 50 0
10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170
Time (µsec) Fig. 14. Variation of water, refrigerant, lubricant bubbles distance from the boundary wall with respect to time.
1552
I. Tzanakis et al. / Experimental Thermal and Fluid Science 35 (2011) 1544–1554
results three critical Weber Number ranges exist (Table 2). When We 6 2 and Re the bubble will not reach a stationary shape but will maintain a ceaselessly oscillating motion across its periphery (‘‘wobbling effect’’) obtaining an amoeboid shape without splitting (Figs. 10 and 11). When 2 6 We 6 10 and Re the bubble obtains higher velocities (Fig. 13) which in some cases due to prevailing surface tension forces the tendency to be merged with bubbles in the vicinity is high (Fig. 6, frames 2–5). When We P 10 and Re the propagation of the large pressure gradient causes the generation and the development of a distinctive liquid jet. The surface tension force cannot resist to a further increment of the area of the bubble and a liquid jet is formed damaging the material. Interestingly, it can be seen that in the case of lubricant and water (Table 2) the transition patterns of the bubbles are totally different than that of the refrigerant. The bold boxes indicate the moments of the bubbles’ implosion during their descent to the boundary wall. In the lubricant environment when We 6 1 and Re 6 1 the bubble achieves very low velocities (u 6 1) avoiding any significant alterations in its shape (Fig. 13). When We and Re P 30 the bubble achieves the highest velocities and a vigorous jet flow develops (Fig. 8, frame 4). In this range the sample can be severely damaged and it is not protected effectively by the cushioning foaming layer. Finally, when 1 6 We 6 30 and 1 6 Re 6 12 the bubble will rise in a steady spherical shape prior to its collapse (Fig. 8, frames 5–8). In the distilled water the bubble follows similar transition patterns. When We 6 1:5 and Re 6 100 the rising bubble can easily reach a well balanced stationary shape. Then distortions arise, the Weber and Reynolds number gradually increases and the collapsing process begins. Finally, when Weber number and Reynolds number are high enough the bubble instantly collapses and a liquid jet is formed. The determinant between the cavitation bubbles of the tested liquids is that in the refrigerant environment the bubbles obtain large Reynolds numbers and low Weber numbers while in the other two environments the Weber number is significant higher and the Reynolds number substantially lower. The behaviour of the cavitation bubbles is determined by the combination of these two dimensionless parameters in given flow conditions. The results are in a very good agreement with the recent work of Wang et al. [30,31]. Wang using analytical and numerical results has shown the impact the Reynolds and Weber number has on the behaviour of a rising water bubble. Wang combined numerical and analytical results to show these correlations. In this study using experimental and analytical results the behaviour of various cavitation bubbles is determined. Thus from the numerical/analytical results of Refs. [30,31] and the experimental/analytical results of this article it can be deduced that the surface tension and the fluid viscosity can both prevent the development of a liquid jet resisting to the formation of a toroidal bubble. The surface tension will resist the increase of bubble surface energy as seen on the refrigerant bubble avoiding any regular implosions. The fluid viscosity will dissipate the flow kinematic energy as seen on the lubricant and water environment having a less profound impact on the formation of a liquid jet. The combination of surface and viscous forces reveal an optimum ratio which can resists the collapsing mechanism as seen in the refrigerant environment (Section 3.3).
Moreover, the bubbles produced in the refrigerant environment with the lowest surface tension had the largest size among the bubbles while their collapsing time was the lowest reducing the risk of damaging the material to a minimum extent. Thus the probability of damaging the surface by consecutive destructive implosion impacts is unlikely to happen in the refrigerant environment. The results are in good agreement with Liu’s et al. [14] and Benhia’s et al. study [12] where they found that the bubble tends to grow larger in lower surface tension and viscosity solutions reducing the collapse duration time. In regards to viscosity, a theoretical correlation between the viscosity and the jet velocity can be determined. The bubble and jet velocity increases as the viscosity increases. Thus the maximum impact pressure can be found within the lubricant bubbles since the lubricant has the highest viscosity among the tested fluids. However, in reality this cannot be applied since the foaming layer which is formed across the surface of the sample plays the role of cushioning the jet impact. Hence, the actual impact damage would not be severe in comparison to the corresponding jet velocity generated by the bubbles and calculated using the high speed camera images. The results are in a very good agreement with Tzanakis et al. work [16], where he proved this particular behaviour of the lubricant experimentally, and with Meged’s study [4] since no direct correlation between the viscosity and the impact potential of the bubbles was drawn. 5. Discussion Three different bubble transition patterns were experimentally revealed in water, lubricant and refrigerant environment using ultrasonic cavitation techniques. The unusual behaviour of the refrigerant bubbles’ in correlation to the lubricant and the water bubbles were effectively interpreted using analytical methods. During the inception of the ultrasonic horn, the water and lubricant bubbles implode generating hundreds of micro-bubbles, diffused to the liquid as a jet stream. The determinant is that the lubricant creates a foaming layer across the solid boundary which absorbs the destructive impact from the larger imploding bubbles. When a refrigerant bubble collapses, micro-bubbles are generated, creating a ‘‘jelly’’ shaped cloud. Camera observations showed that the attraction forces between the refrigerant micro-bubbles are strong enough to prevent any bubble diffusion into the liquid refrigerant. A bubbly jet stream was never formed in the refrigerant environment during the first 1350 ls. In the water and lubricant environment a jet stream was observed after 749 ls and 795 ls respectively. Additionally, the lubricant and the water bubbles are merged due to pressure variations increasing their size. Among the refrigerant bubbles, the merging mechanism is achieved by the powerful attraction of the individual refrigerant bubbles, when they are adjacent to unstable bubbles. Their size remains the same or is further reduced. Across the solid boundary, the refrigerant bubbles have a very unusual behaviour. The pressure variations affect the shape but not the size of the bubbles. The bubbles maintain a ceaselessly oscillating motion across their periphery (‘‘wobbling effect’’). This behaviour is due to the surface tension forces as they prevail to
Table 2 Calculated Weber and Reynolds numbers for each time step during the descending of a typical refrigerant, lubricant and water bubble to the boundary wall until the final implosion. Time (ls) We (Ref) Re (Ref) We (Lub) Re (Lub) We (H2O) Re (H2O)
10 2.57 3000 22.57 11.61 0.14 22.40
20 0.87 1520 0.14 0.50 0.27 28.80
30 2.10 2100 2.95 2.65 23.08 300
40 4.02 3520 23.59 10.61 88.62 960
50 0.54 1500 11.8 10.61 0.15 20
60 0.49 1375 272.6 57.05 4.81 125
70 0.45 1250 1.32 1.49 17.78 340
80 1.61 2250 9.95 5.97 14.89 403.20
90 0.18 825 8.84 7.96 7.38 240
100 4.75 4160 125 34.49 1.23 80
110 1.56 1980 0.27 0.70 1.12 66
120 1.08 1375 18.7 7.00 0.15 19
130 4.64 3420 10.32 9.29 9.57 172.80
140 12.28 6875 7.46 8.96 17.74 372
150 – – 8.71 10.45 192.89 1510.60
160 – – 399.11 75.62 – –
I. Tzanakis et al. / Experimental Thermal and Fluid Science 35 (2011) 1544–1554
the viscous forces, oscillating the bubble to an amoeboid shape formation. The refrigerant has the highest surface tension/viscosity ratio overall k = 155, followed by water and lubricant bubbles with k = 65 and k = 1.8 respectively, indicating a higher resistance to the breaking and collapsing mechanism. The refrigerant bubbles continue to travel in this shape for a significant period of time prior to their collapse. The lubricant bubble is determined to have the highest bubble velocity and average jet velocity among the testing liquids. However, the foaming layer across the solid boundary takes on the critical role of cushioning the jet impact, protecting the material surface. Thus the damage produced is significantly less than expected. The velocity and the average jet velocity results of the refrigerant bubbles in correlation with the absence of regular implosions indicate a less destructive action. The combination of surface and viscous forces reveals an optimum ratio which can resist against the collapsing mechanism and was analytically interpreted by the correlation of Reynolds and Weber numbers. The analysis of the refrigerant bubbles showed that when Weber number is more than 10 (We P 10) and Reynolds number is high enough (Re) the bubble implodes instantly forming a liquid jet. In contrast when We 6 2 and Re the ‘‘wobbling effect’’ prevails, the bubble obtains an amoeboid shape and there is absence of implosions. Finally, when 2 6 We 6 10 and Re the refrigerant bubble reaches higher velocities with a possibility of the merging mechanism to be occurred. In regards to the lubricant and water bubbles’ their behaviour is totally different in comparison to the refrigerant ones and was initially noticed by analysing high speed camera images. As far as the lubricant is concerned, when We 6 1 and Re 6 1 the bubble is in slow motion avoiding any significant alterations in its shape. When We and Re P 30 a vigorous jet flow develops and the bubble is destructively imploded. When 1 6 We 6 30 and 1 6 Re 6 12 the bubble rises in a steady spherical shape prior to its collapse. In the distilled water environment when Weber number and Reynolds number are high enough the bubble instantly collapses and a liquid jet is formed while when We 6 15 and Re 6 100 the rising bubble reaches a well balanced stationary shape. 6. Conclusions To clarify the cavitation mechanisms of three engineering liquids, a systematic observation of the cavitation bubble’s behaviour by means of high-speed photography is conducted. In all the critical regions, refrigerant bubbles were monitored to maintain a ceaselessly oscillating motion across their periphery (‘‘wobbling effect’’) avoiding regular implosions and obtaining an amoeboid formation. This unusual behaviour of the refrigerant cavitation bubbles is effectively determined by the combination of the Reynolds and Weber numbers in given flow conditions. Analytical results showed that in the refrigerant environment the bubbles obtain large Reynolds numbers and low Weber numbers showing a higher resistance to implosion. In the other two environments the Weber number is significantly higher and the Reynolds number substantially lower indicating a tendency of water and lubricant bubbles towards collapsing. The characteristics of all the tested cavitation bubbles are presented in terms of the velocity and average jet velocity values as calculated and with the variation of the bubble radius over time, as obtained from direct observations using high-speed photography. Lubricant bubble is determined to have the highest bubble velocity and average jet velocity among the testing liquids closely followed by the water bubble. The refrigerant bubble having the lowest bubble velocity and average jet velocity in correlation with the absence of regular implosions, cannot be considered as a destructive mean of cavitation for scroll expander systems or similar automotive industrial units. The camera observations of the interaction cavitation mechanisms, the analytical results and the
1553
interesting behaviour of the refrigerant bubbles give rise to the rich variety of cavitation mechanisms and effects which will be examined in future studies. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.expthermflusci.2011.07.005. References [1] I. Tzanakis. M. Hadfield, Z. Khan, Durability of Domestic Scroll Compressor Systems, Surface Effects and Contact Mechanics, 9th International Conference on Surface Effects and Contact Mechanics – Computational Methods and Experiments, Contact/Surface 2009, WIT Press, 2009, pp. 229–240. doi:10.2495/SECM090211. ISBN:978-184564186-3. [2] I. Tzanakis, M. Hadfield, A. Georgoulas, N. Kotsovinos. Cavitation Damage Observations Within Scroll Expander Lubrication Systems, 3rd International Conference on Tribology and Design, in press. [3] E.A. Brujan, P.R. Williams, Cavitation phenomena in non-newtonian liquids, Chemical Engineering Research and Design 84 (2006) 293–299. [4] P. R Williams, P.M. Williams, S.W.J. Brown, A study of liquid jets formed by bubble collapse under shock waves in elastic and Newtonian liquids, Journal of Non-Newtnian Fluid Mechanics 76 (1998) 307–325. [5] Y. Meged, C.H. Venner, W.E. ten Napel, Classification of lubricants according to cavitation criteria, Wear 186 (1995) 443–453. [6] B. Schneider, A. Kosar, Y. Peles, Hydrodynamic cavitation and boiling in refrigerant (R-123) flow inside microchannels, International Journal of Heat and Mass Transfer 50 (2007) 2838–2854. [7] E.A. Brujan, A.F.H. Al-Hussany, R.L. Williams, P.R. Williams, Cavitation erosion in polymer aqueous solutions, Wear 264 (2008) 1035–1042. [8] S. Popinet, S. Zaleski, Bubble collapse near a solid boundary: a numerical study of the influence of viscosity, Journal of Fluid Mechanics 464 (2002) 137–163. [9] P.R. Williams, P.M. Williams, S.W.J. Brown, A study of liquid jets formed by bubble collapse under shockwaves in elastic and Newtonian liquids, Journal of Non-Newtonian Fluid Mechanics 76 (1998) 307–325. [10] A. Berker, M.G. Bouldin, S.J. Kleis, W.E. Van Arsdale, Effect of polymer on flow in journal bearings, Journal of Non-Newtonian Fluid Mechanics 53 (1995) 333–347. [11] B. Karunamurthy, M. Hadfield, C. Vieillard, G.E. Morales-Espejel, Z. Khan, Erosion and rolling contact wear mechanisms in silicon nitride hybrid bearings, Ceramics International Journal, in press. [12] S. Behnia, A. Jafari, W. Soltanpoor, O. Jahanbakhsh, Nonlinear transitions of a spherical bubble, Chaos, Solions and Fractals 41 (2009) 818–828. [13] Y. Iwai, S.Li. Cavitation, Cavitation erosion in waters having different surface tensions, Wear 254 (2003) 1–9. [14] X.M. Liu, J. He, J. Lu, X.W. Ni, Effects of surface tension on a liquid-jet produced by the collapse of a laser-induced bubble against a rigid boundary, Optics and Laser Technology 41 (2009) 21–24. [15] H. Soyama, Y. Yanauchi, K. Sato, R. Oba, R. Oshima, High-speed observation of ultrahigh-speed submerged water jets, Experimental Thermal and Fluid Science 12 (1996) 411–416. [16] I. Tzanakis, M. Hadfield, N. Garland, Cavitation damage incubation with typical fluids applied to a scroll expander system, Tribology International, in press, Available Online 21 June 2011. [17] R. Etienne, L. Jacques, F. Mohamed, M. Peter, A. François, Cavitation bubble behaviour inside a liquid jet, Physics of Fluids 19 (2007) 067106. [18] X. Chen, R.Q. Xu, Z.H. Shen, J. Lu, X.W. Ni, Optical investigation of cavitation erosion by laser-induced bubble collapse, Optics and Laser Technology 36 (2004) 197–203. [19] Li-Xin Bai, Wei-Lin Xu, Zhong Tian, Nai-Wen Li, A high-speed photographic study of ultrasonic cavitation near rigid boundary, Journal of Hydrodynamics 20 (2008) 637–644. [20] O. Lindau, W. Lauterborn, Investigation of the Countrjet Developed Cavitation Bubble that Collapses Near a Rigid Boundary, CAV 2001, 4th International Symposium on Cavitation, 2001. [21] C. Ohl, T. Kurz, R. Geisler, O. Lindau, W. Lauterborn, Microimplosions: cavitation collapse and shock wave emission on a nanosecond time scale, The Royal Society 357 (1999) 269–294. [22] R. Pecha, Microimplosions: cavitation collapse and shock wave emission on a nanosecond time scale, Physics 84 (2000) 1328–1330. [23] R.P. Tong, W.P. Schiffers, J.S. Shaw, J.R. Blake, D.C. Emmony, The role of splashing in the collapse of the laser-generated cavity near a rigid boundary, Journal of Fluid Mechanics 380 (1999) 339–361. [24] E.A. Brujan, T. Ikeda, Y. Matsumoto, On the pressure of cavitation bubbles, Experimental Thermal and Fluid Science 32 (2008) 1188–1191. [25] Y. Matsumoto, J.S. Allen, S. Yoshizawa, T. Ikeda, Y. Kaneko, Medical ultrasound with microbubbles, Experimental Thermal and Fluid Science 29 (2005) 255– 265. [26] A.H. Lefebvre, Atomization and Sprays, Hemisphere Pub., Corp., New York, 1989. [27] B. Schneider, A. Kosar, Y. Peles, Hydrodynamic cavitation and boiling in refrigerant (R-123) flow inside microchannels, Heat and Mass Transfer 50 (2007) 2838–2854.
1554
I. Tzanakis et al. / Experimental Thermal and Fluid Science 35 (2011) 1544–1554
[28] C. Mishra, Y. Pelesa, Flow visualization of cavitating flows through a rectangular slot micro-orifice ingrained in a microchannel, Physics of Fluids 17 (2005). [29] C. Ohl, M. Arora, R. Ikink, N. Jong, M. Versluis, M. Delius, D. Lohse, Sonoporation from jetting cavitation bubbles, Biophysical Journal 91 (2006) 4285–4295.
[30] H. Wang, Z. Zhang, Y. Yang, H. Zhang, Viscosity effects on the behaviour of a rising bubble, Journal of Hudrodynamics 22 (2010) 81–89. [31] H. Wang, Z. Zhang, Y. Yang, H. Zhang, Surface tension effects on the behaviour of a rising bubble driven by buoyancy force, Chinese Physics B 19 (2010) 026801-1–026801-9.
Contents lists available at SciVerse ScienceDirect
Experimental Thermal and Fluid Science journal homepage: www.elsevier.com/locate/etfs
Observations of acoustically generated cavitation bubbles within typical fluids applied to a scroll expander lubrication system I. Tzanakis a,⇑, M. Hadfield a, I. Henshaw b a b
Sustainable Engineering Research Centre, Bournemouth University, Talbot Campus BH12 5BB, UK Energetix Group Limited, Capenhurst Technology Park, Chester CH1 6EH, UK
a r t i c l e
i n f o
Article history: Received 5 May 2011 Accepted 18 July 2011 Available online 31 July 2011 Keywords: Cavitation Reynolds Weber number Viscosity Surface tension Refrigerant Lubricant Scroll
a b s t r a c t An experimental study to evaluate the dynamic performance of three different types of cavitation bubbles is conducted. An ultrasonic transducer submerged into the working fluids of a scroll expander is utilised to produce cavitation bubbles and a high speed camera device is used to capture their behaviour. Three critical regions around the ultrasonic source, between the source and the solid boundary, and across the solid boundary were observed. Experimental results revealed that refrigerant bubbles sustain a continuous oscillatory movement, referenced as ‘‘wobbling effect’’, without regularly collapsing. Analytical results indicate the influence of several factors such as surface tension/viscosity ratio, Reynolds number and Weber number which interpret that particular behaviour of the refrigerant bubbles. Within the refrigerant environment the bubbles obtain large Reynolds numbers and low Weber numbers. In contrast, within the lubricant and the water environment Weber number is significantly higher and Reynolds number substantially lower. The bubble radius and velocity alterations are accurately calculated during the cavitation process. Lubricant bubbles achieve the highest jet velocity while refrigerant bubbles having the lowest jet velocity are not considered as a destructive mean of cavitation for scroll expander systems. Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction Scroll expanders have been a key component of many industrial systems over the last 20 years (CHP systems, air-conditions, pumps, etc.). A specific scroll expander was tested for 1000 h as a part of a small domestic combined heat and power (CHP) system. Cavitation was identified as a wear mechanism during the investigation of the scroll’s components [1]. The cavitation mechanism developed inside the scroll is caused by the operational fluid environment [2]. This environment includes two fluids; a high molecular organic refrigerant and a synthetic lubricant. The refrigerant, as a gas form, drives the scroll while the lubricant protects the parts of the scroll from excessive wear. Thus the present study focuses on these two scroll fluids and in the dynamic behaviour of their cavitation bubbles investigating their engineering implications to scroll expander systems or similar automotive industrial units. Limited experimental information concerning lubricant and refrigerant bubble formation and their dynamic behaviour [2–6] is available within the published literature. The influence of viscosity and surface tension upon cavitation threshold has led to some debate within published literature, with claims of increased or decreased cavitation activity arising from
⇑ Corresponding author. E-mail address: [email protected] (I. Tzanakis). 0894-1777/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.expthermflusci.2011.07.005
different studies of similar environments. Brujan et al. [7] in polymer aqueous solutions has shown that higher levels of viscosity can substantially mitigate the tendency of cavitational activity to damage. Popinet and Zaleski [8] in his study has found that the amplitude of the oscillations decreases due to viscous damping, while the jet impact velocity decreases as viscosity increases. Williams et al. [9] and Berker et al. [10], reported a mitigating effect of viscosity in the jet propulsion by the bubbles collapse of various fluids. Karunamurthy et al. [11] presents the viscosity influence of various lubricants on erosion rate, showing that as the viscosity increases the erosion decreases. In contrast, Meged et al. [4] after a series of experiments using 20 different types of liquid lubricants, has found that there is no direct correlation between the viscosity of a liquid and the cavitation erosion mechanism. In regards to surface tension in a micro-scale level, the bubble growth and the elongation are influenced by the surface tension forces which are dominant at this scale [12]. According to Iwai and Li [13], experiments conducted in various water solutions with different surface tensions, showed that surface tension increases the liquid-jet impact and the bubble size. In contrast Liu et al. [14] has shown that in a cavitation environment higher surface tension values leads to smaller sized bubbles reducing the collapse duration. In the present study cavitation tests were utilised to analyse and evaluate the dynamic behaviour of the cavitation bubbles generated within the refrigerant and the lubricant environment of a scroll expander system. Distilled water was also used as a
1545
I. Tzanakis et al. / Experimental Thermal and Fluid Science 35 (2011) 1544–1554
reference fluid because of its cavitation bubbles well known behaviour. Bubble behaviour was studied with the use of high speed photography and three distinctive patterns of cavitation bubbles were monitored. In the first pattern the generation mechanism of cavitation bubbles was investigated during their incubation time around the ultrasonic source. Therefore, the bubbles’ behaviour was monitored during their descent to the solid surface and across the solid boundary, revealing two more interesting patterns. Analytical results indicate the cavitation behaviour of all the tested bubbles according to their viscous and surface tension forces. The unusual behaviour of the refrigerant bubble was interpreted using combinations of Reynolds and Weber number while the bubble radius and velocity alterations were calculated during the cavitation process for all the liquid environments. 2. Experimental procedure The experimental apparatus is described in Fig. 1. The experiments were performed based on the ASTM G32-03 standard method. The ASTM method proposes a standard probe of 15.9 mm diameter to be generally used for cavitation tests. However, in this study a probe of 5 mm was deployed to produce cavitation bubbles for more definitive images. Experiments were carried out by using an ultrasonic transducer at a frequency of 20 kHz. The ultrasonic horn has a plane surface oscillating in a simple harmonic motion. Peak to peak amplitude of the horn tip was adjusted to be at 50 lm. The sample consists of a standard chromium steel ball EN-31, polished in a few microns accuracy and implemented on a Bakelite base. The Bakelite base of the sample was mounted on the bottom of a small transparent square beaker with a maximum capacity of 500 ml. The tip of the ultrasonic horn was submerged into the liquid environment of the beaker in a distance of 0.5 mm from the surface of the chromium steel. The liquids used for the tests were distilled water, a synthetic lubricant and a high molecular organic refrigerant. All the liquids were tested in a room temperature of around 23 ± 2 °C and their physical properties used for the experiments are listed in Table 1. The viscosity and surface tension for all the three fluids were measured in the ambient environmental conditions of the laboratory using appropriate experimental devices. A high speed camera (Phantom v7.3) was installed at one of the sides of the beaker. At the beaker’s opposite side a high intensity fibre optic light source (2 kW) was deployed to provide sufficient light for investigation of critical areas. The Phantom v7.3 supplies a full frame (800 600 pixels) recording rate, with a maximum
Table 1 Physical properties of experimental liquids at 23 ± 2 °C. Fluid
Water
Refrigerant
Lubricant
Ultrasonic velocity inside liquid (m/s) Liquid density (kg/m3) Viscosity (mPa s) Surface tension (mN/m) Vapour pressure (Pa) at 20 °C
1476 1000 1 62–65 2230
990 625 0.225 32–35 53,300
1700 995 30 51–54 1300
shutter speed of up to 500,000 frames per second with 1 l exposure time. In this study the speed modes of the camera used for the photography varied between 30.000 fps and 220.000 fps with a 10 ls exposure time. During the tests a Leica mono zoom amplifier was used in front of the camera achieving a resolution of approximately 8 lm/pixel. The images focused on three critical regions; on the tip of the ultrasonic probe where the bubbles are generated, the region across the lower boundary wall where bubbles are destructively collapsed and at the area in between and such methodology was performed for all the liquids tested. An evaluation of various bubble images using the camera software was conducted and a range of graphs was produced. In each of the liquid environments more than 20 different bubble behaviours were monitored. Three distinctive patterns of bubbles’ typical behaviour were finally determined. 3. Experimental results 3.1. Region around the ultrasonic source The bubbles’ generation and the expansion mode on the tip of the horn revealed three different behaviours for each of the liquid environments respectively. A typical isolated incubation bubble was monitored in each liquid environment (Figs. 2–4). When water was used, the deformation of the water bubble started after 189 ls. In the case of the liquid refrigerant it was measured at 90 ls while for the lubricant was at 315 ls. Furthermore the time step where the jet stream was generated, was at 749 ls and at 795 ls for the water and the lubricant respectively. Interestingly, for the refrigerant a similar liquid jet stream was never formed during the first 1350 ls. The refrigerant bubbles maintain a robust shape avoiding the bubbles which are generated to separate, escape and finally collapse. Specifically, in the case of the distilled water Fig. 2, a typical cavitation mechanism is observed. The growth and the implosion
Ultrasonic Processor Computer/Camera software
Experimental Beaker Fiber Optic light-source High-speed Camera
Test Sample Fig. 1. Schematic of experimental apparatus
1546
I. Tzanakis et al. / Experimental Thermal and Fluid Science 35 (2011) 1544–1554
Tip of the ultrasonic horn
Water Jet Stream 189µs
720µs
749µs
Micro bubbles
813µs
1363µs
1135µs
1609µs
Fig. 2. Expansion mode of an acoustic cavitation refrigerant bubble on the tip of the horn. (Frame size is 1 0.5 mm).
Micro-jet Tip of the ultrasonic horn
315µs
Counter jet formation 465µs
450µs
960µs
Pancake Effect
360µs
Lubricant Jet Stream
795µs
1125µs
Toroidal Bubble
405µs
885µs
1305µs
1260µs
Fig. 3. The merging mechanism of the water-lubricant bubbles. (Frame size is 0.5 0.5 mm.)
Tip of the ultrasonic horn
15µs
165µs
135µs
Pancake Effect
Toroidal Bubble
90µs
Counter jet formation
210µs
120µs
225µs
Microbubbles jelly cloud 240µs
370µs
760µs
Fig. 4. The merging mechanism of the refrigerant bubble. (Frame size is 0.5 0.5 mm.)
1350µs
1547
I. Tzanakis et al. / Experimental Thermal and Fluid Science 35 (2011) 1544–1554
of the bubbles due to pressure variations creates a distinctive jet stream of micro-bubbles (frames 4–5, Fig. 2) which in turn can damage the material surface by pressure impact. After a 1000 ls a cloud of micro-bubbles is generated (frames 6–8) with a radius 10–12 times smaller than the primary bubble. The micro-bubble cloud exists continuously while over time becomes asymmetric (frames 5–8, Fig. 2), discharged fairly far downstream from the tip of the horn. Further downstream, as the pressure rapidly increases, the cavitation clouds rapidly divide into small bubbles and then disappeared [15]. In the lubricant environment a ‘‘pancaking’’ effect at the site of the collapsed bubble can be observed (frames 2–3, Fig. 3). The ‘‘pancaking’’ shape is formed only when the initial bubble is very close to the wall during the incubation period of time. After this phase the bubble rapidly obtains a toroidal shape (frame 4, Fig. 3). Then a micro-jet is formed, which is usually difficult to identify in this ‘‘pancaking’’ effect mode, leading eventually to a counter jet generation (frames 5–6, Fig. 3). As the bubble rebounds and grows, in an almost round spherical shape (frame 9, Fig. 3), it then rapidly collapses (frame 11, Fig. 3). A linear jet stream, which comprised of a cloud of micro-bubbles (up to 30 times smaller than the water ones), exerts from the centre of the collapse bubble (frame 7, Fig. 3). Interestingly within the lubricant environment part of the micro-bubbles instead of disappeared during their downstream are reformed to a thin foaming layer and spread along the surface of the solid boundary. The foaming layer plays the role of a cushion, significantly reducing the destructive power of the micro-jet impact [16]. In the case of the refrigerant, the spherical bubble initially is formed to a ‘‘pancaking’’ shape (frame 3, Fig. 4) followed by a counter jet (frames 4–5, Fig. 4). Then it can be clearly seen that loads of micro-bubbles are connected in a ‘‘jelly’’ cloud shape (frames 9–12, Fig. 4). In contrast to the previous cases (water, lubricant) during the cavitation process in the liquid refrigerant there is no bubble diffusion. Molecules cannot overcome the surface tension and other cohesive forces maintaining their bubbly shape. The individual bubbles cannot resist the attractive forces creating strong bonding in a single bubbly body. Regular implosions are unlikely to be observed.
3.2. Region between source and solid surface During their descent to the boundary wall, the lubricant and water bubbles merged or split due to pressure variations (Fig. 5). Thus, when the pressure increases, the bubble implodes and then is divided into smaller bubbles. When the pressure decreases, their radius increases and they eventually merged to a single bubble (frame 4, Fig. 5). In the case of the refrigerant, the adjacent unstable bubbles, present a distorted shape leading to significant alterations in their volume and size. Larger bubbles attract and absorb the smaller ones in the vicinity. The bubbles within the refrigerant solution have a more vigorous behaviour and their tendency to merge is higher than in the other two liquids. Frame 4 in Fig. 6 captures
Substrate
40µs
the attachment mechanism of two amoeboid shaped refrigerant bubbles. The adjacent bubbles are connected leading temporarily to an elongated bubbly shape prior to obtaining a single amoeboid shaped bubble. The pressure variations generate a wobbling behaviour of the refrigerant bubbles, referenced as ‘‘wobbling effect’’ affecting their movement, their morphology (amoeboid shape), their implosion rate (Section 3.3) and their diffusion to the liquid. The frequency of the implosions from the refrigerant bubbles is substantially lower comparing to the other two liquid environments, while a lack of consistency is recorded during their collapse (Fig. 12). Observing the two different bubble merging mechanisms, is evident that lubricant and water bubbles maintain a spherical shape across the bubbles periphery with a significant increment on their size. In the case of the refrigerant, an amoeboid formation is maintained, while the size of the bubble after merging can be further reduced.
3.3. Region near the solid surface boundary The bubbles implosion near the rigid boundary revealed three distinctive behaviours for each of the tested liquids respectively. In the water environment, a typical behaviour of a cavitation bubble is described. The steps of the collapse mechanism are briefly presented in Fig. 7 since previous researchers have analysed the phenomenon and its dynamic performance in close proximity to the boundary wall [17–24]. The life cycle of the bubble can be determined by the acoustic cycles. During one acoustic cycle the bubble grows up reaching its maximum diameter and then collapses. The attractive Bjerknes forces deform the spherical bubble to a toroidal bubble (frame 4, Fig. 7), while a micro-jet is formed and instantly a counter-jet emits a powerful shock wave (frame 5, Fig. 7). Then the bubble becomes unstable and is separated in two parts (frame 6, Fig. 7) which individually generate two micro-bubble clouds, the so-called splashing effect (frame 7, Fig. 7). Thereafter the two clouds merge into one (frame 8, Fig. 7), increasing in size and collapse releasing high velocity destructive pressure impacts to the boundary wall, damaging the material surface. The maximum pressure emitted from the bubbles at the cloud centre becomes much higher than that of a single bubble [25]. Regarding the lubricant, the bubbles have a similar behaviour to that of water (Fig. 8). They were generated in low pressure areas and then they collapsed in regions where the pressure is higher. It is worth noting that the foaming layer generated during the cavitation process (frames 2–3, Fig. 8) can affect the destructive impact of the larger bubbles’ collapse mechanism. The thin foaming layer consists of micro-bubbles with a radius less than 3 lm (frame 2, Fig. 9) and it can protects the material surface by acting as a cushion, absorbing and reducing the impact of the implosion. Thus, despite the fact that the average jet velocity from the lubricant bubbles with an average radius of 50–100 lm is higher than in the water (Section 4), the actual damage is considerably less. The results are in a very good agreement with Tzanakis et al. [16] experimental study where various steel materials under
50µs
120 µs
Fig. 5. Characteristic behaviour of an acoustic cavitation water bubble near the solid boundary. (Frame size is 1 0.5 mm.)
1548
I. Tzanakis et al. / Experimental Thermal and Fluid Science 35 (2011) 1544–1554
Boundary Wall
22µs
44µs
88µs
110µs
66µs
Amoeboid shape 77µs
121µs
Fig. 6. Collapse of an acoustic cavitation lubricant bubble near the solid boundary. (Frame size 0.5 0.5 mm).
Pancake Effect
Spherical Bubble
Boundary Wall Counter-Jet
110µs
Toroidal Bubble
140µs
Micro-Jet
160µs
Shock Wave
170µs
Two Individual Bubbles
200µs
Microbubbles Cloud
234µs
Destructive Collapse
370µs
Fig. 7. Lubrication layer formed across the surface of the rigid boundary, prior to bubble’s implosion. (Frame size is 0.5 0.5 mm.)
Foaming Layer
11µs Substrate
Substrate
Micro-jet formation
22µs
34µs Bubble implosion
89µs
100µs
111µs
Substrate
121µs
Fig. 8. The ceaselessly wobbling movement of an acoustic cavitation refrigerant bubble near the solid boundary. (Frame size 0.5 0.5 mm.)
1549
I. Tzanakis et al. / Experimental Thermal and Fluid Science 35 (2011) 1544–1554
Lubrication Film Thickness
20µs
Subsurface
70µs
50µs
Fig. 9. The different shapes the cavitation refrigerant bubble obtains in its maximum radius. (Frame size 0.5 0.5 mm.)
lubrication, refrigerant and water environment were tested. Tzanakis found that the morphology evolution of the cavities produced across the sample’s surface in a lubricant environment is significantly lower in comparison to the water. The refrigerant bubbles have by far the smallest cavitation pressure impact. When refrigerant was used, bubbles had an interesting behaviour across the surface of the solid boundary (Figs. 10 and 11). The refrigerant bubbles sustain a seemingly random dual motion. Small vigorous oscillations across the bubble surface are accompanied with a wobbling behaviour (‘‘wobbling effect’’); most of the refrigerant bubbles wobbled instead of collapsing. Eventually the refrigerant bubbles obtain an amoeboid shape (frame 2, Figs. 10 and 11). This unusual behaviour of the refrigerant bubbles significantly increases their lifecycle reducing the probability of a destructive impact. The bubbles instead of growing and collapsing at regular time steps are driven by the ‘‘wobbling effect’’ for a prolonged period of time prior to their collapse. A typical series of images in Fig. 10 show a refrigerant bubble prior to its collapse. The bubble after reaching its maximum size (frame 1, Fig. 10), instead of starting the collapse process it follows an oscillating motion. Even after 700 ls the size of the bubble
remains the same and there is a total absence of implosions. The shape slightly changes to similar amoeboid forms. Additionally, Fig. 11 captures another refrigerant bubble with more distinct phases of an amoeboid formation. Immediately after the bubble attained its maximum size, it is moving within the liquid refrigerant without collapsing. In frame 4 a separation mechanism appears. However, the attraction forces of the bubble restrain this phenomenon, keeping the bubble in a robust amoeboid single shape. In frame 10 the bubble is elongated reaching its maximum length and then returns to a bubbly round shape instead of collapsing. The interaction of the surface tension and the viscosity forces of the refrigerant liquid can explain the behaviour of the refrigerant cavitation bubbles. It has been reported in previous studies that viscous forces giving rise to perturbations tend to prevent the breakup of the cavitation bubbles while surface tension tends to prevent bubbles from collapsing [26]. In the present study the surface tension forces prevail the viscous forces within the liquid refrigerant maintaining a bubbly wobbling pattern among the cavities, preventing them from collapsing. The opposite phenomenon observed within the water and lubricant solutions, as viscous forces have a major role leading the bubbles to consecutive
Direction
Bubble’s maximum size
214µs
Amoeboid shape
36µs
72µs
108µs
247µs
377µs
400µs Direction
468µs
514µs
526µs
Boundary Wall
Fig. 10. Variation of water, refrigerant, lubricant cavitation bubbles radius with respect to time.
682µs
1550
I. Tzanakis et al. / Experimental Thermal and Fluid Science 35 (2011) 1544–1554
Direction
Amoeboid shape
140µs
270µs
60µs
30µs
Boundary Wall
170µs
290µs
90µs
210µs
240µs
310µs
340µs
Fig. 11. Variation of water, refrigerant, lubricant bubbles velocity and jet velocity with respect to time.
Bubble Radius ( µ m)
140 120 100
Water Refrigerant Lubricant
80 60 40 20 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170
Time (µsec) Fig. 12. Variation of water, refrigerant, lubricant bubbles distance from the boundary wall with respect to time.
oscillations, separating the bubble body into smaller cavities and this in turn generates regular implosions among the bubbles. Specifically, the refrigerant has the highest surface tension/viscosity ratio overall k = 155 (table 1), indicating a higher resistance to the breaking and collapsing mechanism. The above behaviour also enhances the probability of the refrigerant bubbles to form a ‘‘jelly’’ cloud shape pattern (Fig. 4). A similar behaviour was also noted by Schneider et al. [27] in his latest study, using another type of refrigerant flowing in micro-channels, where the bubbly flow morphology of low surface tension liquid in a fully developed hydrodynamic cavitating flow was highlighted. In contrast, the water and the lubricant have a much smaller surface tension/viscosity ratio of k = 65 and k = 1.8 respectively indicating a prone transition to a toroidal shape. Especially, in the case of the lubricant this is more obvious. The viscous forces affect radically the performance of the lubricant bubbles reducing their momentum (Fig. 13). Similar results were also found by Mishra and Pelesa [28] in his recent work using de-ionised water with substantially higher surface tension and viscosity values to the refrigerant solutions. Mishra showed
the disturbances a water bubble’s body experiences and the tendency of the bubble to implode according the surface tension and viscosity values during the cavitation process. Additionally, the size of the refrigerant bubbles was found to be the largest among the tested bubbles with an average bubble radius around 60–90 lm followed by the water and lubricant bubbles with an average radius of 30–50 lm and 20–45 lm respectively. This implies that the refrigerant bubbles with higher surface tension/viscosity ratio will have more probability to grow to larger sizes by avoiding early breaking-up into much smaller bubbles, during the growing process, in contrast to the lubricant bubbles where their size is substantially smaller. This also explains the hundreds of micro-bubbles which are formed across the foaming protective layer. 4. Analytical results The three characteristic bubble transition patterns, revealed experimentally, are interpreted using analytical methods. Typical
1551
10 9 8 7 6 5 4 3 2 1 0
Water Lubricant Ref Jet
30
Refrigerant Water Jet Lub Jet
25 20 15 10 5 0
10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170
Average Jet Velocity (m/sec)
Bubble Velocity (m/sec)
I. Tzanakis et al. / Experimental Thermal and Fluid Science 35 (2011) 1544–1554
Time (µsec) Fig. 13. Schematic of experimental apparatus.
behaviour of a similar radius cavitation bubble from its rising moment until its final collapse stage in each of the liquid environments is analysed. Fig. 12 presents the fluctuations of the bubble radius over the time for each of the liquids environments. Lubricant and water bubbles after their first implosion take the rebound by re-forming into smaller sizes and afterwards they implode again and reform. At the level of the boundary wall they reach their maximum size before the final collapse. The refrigerant bubbles keep linear fluctuation behaviour on their size while their radius has not been significantly altered since the wobbling behaviour prevails. The only implosion is observed at the level of the boundary wall where refrigerant bubble size is at its maximum. Fig. 13 shows the bubble’s velocity and the average jet velocity during the descent of the bubbles to the boundary wall. The bubble velocity was calculated during the expansion and the shrinkage mode of the bubbles prior to their final collapse onto the solid boundary. The average jet velocity is estimated by the assumption that the jet flow develops during the shrinkage of the bubble from its maximum radius. The average jet velocity of the bubble was estimated, as Claus–Dieter Ohl suggests in his study, using following equation [29]:
ujetðaveÞ ¼
2Rmax Tc
ð1Þ
Distance from the Boundary (µm)
where Rmax is the maximum radius of the bubble and Tc is the time collapse. The lubricant and water bubbles show significant alterations in their velocity following their radius fluctuation patterns. The lubricant bubbles achieve the highest bubble and jet velocity while the refrigerant bubbles implode only once having the lowest jet velocity among the tested bubbles. This happens close to the boundary wall. Refrigerant bubbles keep a constant velocity during their entire route reaching their maximum speed just before the
implosion with the surface. The velocity achieved by the refrigerant bubbles has the lowest impact potential. Fig. 14 shows the variation of boundary distance over the time as each of the bubbles approaches the boundary wall. Water bubbles, according to the inclination of the graph, have the fastest descent to the solid surface achieving an average speed of 2 m/s. The refrigerants bubbles follow with an average descending speed of 1.25 m/s. The lubricant bubbles are the last having the smallest descending speed of 0.6 m/s within the liquid lubricant. The viscosity in the lubricant is the highest within the liquids, decelerating the movement of the bubbles. Thus even if the bubble velocity and the average jet velocity of the lubricant is the highest among the tested liquids (Fig. 13), the motion is seriously affected by the viscous environment reducing bubble’s momentum. In contrast, the refrigerant bubbles having the lowest viscosity move to a faster pace from the lubricant bubbles but not from the water bubbles. The absence of significant oscillations in correlation with the ‘‘wobbling effect’’, highly affect the refrigerant bubbles’ motion reducing their momentum. Concomitantly, it is reasonable to report that the combination of the surface tension with the fluid viscosity properties of the bubbles (table 1) seriously affect the cavitation mechanism. The interpretation of the refrigerant bubbles’ unusual behaviour lies on the values of Reynolds and Weber number for each time step
Re ¼
qud qu2 d and We ¼ l r
where q is the fluid density, u is the bubble’s velocity, d is the diameter of the bubble in each time step, l is the fluid viscosity and r is the surface tension. Investigating the behaviour of a typical single refrigerant bubble according to the experimental analysis and the analytical
350 Water Refrigerant Lubricant
300 250 200 150 100 50 0
10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170
Time (µsec) Fig. 14. Variation of water, refrigerant, lubricant bubbles distance from the boundary wall with respect to time.
1552
I. Tzanakis et al. / Experimental Thermal and Fluid Science 35 (2011) 1544–1554
results three critical Weber Number ranges exist (Table 2). When We 6 2 and Re the bubble will not reach a stationary shape but will maintain a ceaselessly oscillating motion across its periphery (‘‘wobbling effect’’) obtaining an amoeboid shape without splitting (Figs. 10 and 11). When 2 6 We 6 10 and Re the bubble obtains higher velocities (Fig. 13) which in some cases due to prevailing surface tension forces the tendency to be merged with bubbles in the vicinity is high (Fig. 6, frames 2–5). When We P 10 and Re the propagation of the large pressure gradient causes the generation and the development of a distinctive liquid jet. The surface tension force cannot resist to a further increment of the area of the bubble and a liquid jet is formed damaging the material. Interestingly, it can be seen that in the case of lubricant and water (Table 2) the transition patterns of the bubbles are totally different than that of the refrigerant. The bold boxes indicate the moments of the bubbles’ implosion during their descent to the boundary wall. In the lubricant environment when We 6 1 and Re 6 1 the bubble achieves very low velocities (u 6 1) avoiding any significant alterations in its shape (Fig. 13). When We and Re P 30 the bubble achieves the highest velocities and a vigorous jet flow develops (Fig. 8, frame 4). In this range the sample can be severely damaged and it is not protected effectively by the cushioning foaming layer. Finally, when 1 6 We 6 30 and 1 6 Re 6 12 the bubble will rise in a steady spherical shape prior to its collapse (Fig. 8, frames 5–8). In the distilled water the bubble follows similar transition patterns. When We 6 1:5 and Re 6 100 the rising bubble can easily reach a well balanced stationary shape. Then distortions arise, the Weber and Reynolds number gradually increases and the collapsing process begins. Finally, when Weber number and Reynolds number are high enough the bubble instantly collapses and a liquid jet is formed. The determinant between the cavitation bubbles of the tested liquids is that in the refrigerant environment the bubbles obtain large Reynolds numbers and low Weber numbers while in the other two environments the Weber number is significant higher and the Reynolds number substantially lower. The behaviour of the cavitation bubbles is determined by the combination of these two dimensionless parameters in given flow conditions. The results are in a very good agreement with the recent work of Wang et al. [30,31]. Wang using analytical and numerical results has shown the impact the Reynolds and Weber number has on the behaviour of a rising water bubble. Wang combined numerical and analytical results to show these correlations. In this study using experimental and analytical results the behaviour of various cavitation bubbles is determined. Thus from the numerical/analytical results of Refs. [30,31] and the experimental/analytical results of this article it can be deduced that the surface tension and the fluid viscosity can both prevent the development of a liquid jet resisting to the formation of a toroidal bubble. The surface tension will resist the increase of bubble surface energy as seen on the refrigerant bubble avoiding any regular implosions. The fluid viscosity will dissipate the flow kinematic energy as seen on the lubricant and water environment having a less profound impact on the formation of a liquid jet. The combination of surface and viscous forces reveal an optimum ratio which can resists the collapsing mechanism as seen in the refrigerant environment (Section 3.3).
Moreover, the bubbles produced in the refrigerant environment with the lowest surface tension had the largest size among the bubbles while their collapsing time was the lowest reducing the risk of damaging the material to a minimum extent. Thus the probability of damaging the surface by consecutive destructive implosion impacts is unlikely to happen in the refrigerant environment. The results are in good agreement with Liu’s et al. [14] and Benhia’s et al. study [12] where they found that the bubble tends to grow larger in lower surface tension and viscosity solutions reducing the collapse duration time. In regards to viscosity, a theoretical correlation between the viscosity and the jet velocity can be determined. The bubble and jet velocity increases as the viscosity increases. Thus the maximum impact pressure can be found within the lubricant bubbles since the lubricant has the highest viscosity among the tested fluids. However, in reality this cannot be applied since the foaming layer which is formed across the surface of the sample plays the role of cushioning the jet impact. Hence, the actual impact damage would not be severe in comparison to the corresponding jet velocity generated by the bubbles and calculated using the high speed camera images. The results are in a very good agreement with Tzanakis et al. work [16], where he proved this particular behaviour of the lubricant experimentally, and with Meged’s study [4] since no direct correlation between the viscosity and the impact potential of the bubbles was drawn. 5. Discussion Three different bubble transition patterns were experimentally revealed in water, lubricant and refrigerant environment using ultrasonic cavitation techniques. The unusual behaviour of the refrigerant bubbles’ in correlation to the lubricant and the water bubbles were effectively interpreted using analytical methods. During the inception of the ultrasonic horn, the water and lubricant bubbles implode generating hundreds of micro-bubbles, diffused to the liquid as a jet stream. The determinant is that the lubricant creates a foaming layer across the solid boundary which absorbs the destructive impact from the larger imploding bubbles. When a refrigerant bubble collapses, micro-bubbles are generated, creating a ‘‘jelly’’ shaped cloud. Camera observations showed that the attraction forces between the refrigerant micro-bubbles are strong enough to prevent any bubble diffusion into the liquid refrigerant. A bubbly jet stream was never formed in the refrigerant environment during the first 1350 ls. In the water and lubricant environment a jet stream was observed after 749 ls and 795 ls respectively. Additionally, the lubricant and the water bubbles are merged due to pressure variations increasing their size. Among the refrigerant bubbles, the merging mechanism is achieved by the powerful attraction of the individual refrigerant bubbles, when they are adjacent to unstable bubbles. Their size remains the same or is further reduced. Across the solid boundary, the refrigerant bubbles have a very unusual behaviour. The pressure variations affect the shape but not the size of the bubbles. The bubbles maintain a ceaselessly oscillating motion across their periphery (‘‘wobbling effect’’). This behaviour is due to the surface tension forces as they prevail to
Table 2 Calculated Weber and Reynolds numbers for each time step during the descending of a typical refrigerant, lubricant and water bubble to the boundary wall until the final implosion. Time (ls) We (Ref) Re (Ref) We (Lub) Re (Lub) We (H2O) Re (H2O)
10 2.57 3000 22.57 11.61 0.14 22.40
20 0.87 1520 0.14 0.50 0.27 28.80
30 2.10 2100 2.95 2.65 23.08 300
40 4.02 3520 23.59 10.61 88.62 960
50 0.54 1500 11.8 10.61 0.15 20
60 0.49 1375 272.6 57.05 4.81 125
70 0.45 1250 1.32 1.49 17.78 340
80 1.61 2250 9.95 5.97 14.89 403.20
90 0.18 825 8.84 7.96 7.38 240
100 4.75 4160 125 34.49 1.23 80
110 1.56 1980 0.27 0.70 1.12 66
120 1.08 1375 18.7 7.00 0.15 19
130 4.64 3420 10.32 9.29 9.57 172.80
140 12.28 6875 7.46 8.96 17.74 372
150 – – 8.71 10.45 192.89 1510.60
160 – – 399.11 75.62 – –
I. Tzanakis et al. / Experimental Thermal and Fluid Science 35 (2011) 1544–1554
the viscous forces, oscillating the bubble to an amoeboid shape formation. The refrigerant has the highest surface tension/viscosity ratio overall k = 155, followed by water and lubricant bubbles with k = 65 and k = 1.8 respectively, indicating a higher resistance to the breaking and collapsing mechanism. The refrigerant bubbles continue to travel in this shape for a significant period of time prior to their collapse. The lubricant bubble is determined to have the highest bubble velocity and average jet velocity among the testing liquids. However, the foaming layer across the solid boundary takes on the critical role of cushioning the jet impact, protecting the material surface. Thus the damage produced is significantly less than expected. The velocity and the average jet velocity results of the refrigerant bubbles in correlation with the absence of regular implosions indicate a less destructive action. The combination of surface and viscous forces reveals an optimum ratio which can resist against the collapsing mechanism and was analytically interpreted by the correlation of Reynolds and Weber numbers. The analysis of the refrigerant bubbles showed that when Weber number is more than 10 (We P 10) and Reynolds number is high enough (Re) the bubble implodes instantly forming a liquid jet. In contrast when We 6 2 and Re the ‘‘wobbling effect’’ prevails, the bubble obtains an amoeboid shape and there is absence of implosions. Finally, when 2 6 We 6 10 and Re the refrigerant bubble reaches higher velocities with a possibility of the merging mechanism to be occurred. In regards to the lubricant and water bubbles’ their behaviour is totally different in comparison to the refrigerant ones and was initially noticed by analysing high speed camera images. As far as the lubricant is concerned, when We 6 1 and Re 6 1 the bubble is in slow motion avoiding any significant alterations in its shape. When We and Re P 30 a vigorous jet flow develops and the bubble is destructively imploded. When 1 6 We 6 30 and 1 6 Re 6 12 the bubble rises in a steady spherical shape prior to its collapse. In the distilled water environment when Weber number and Reynolds number are high enough the bubble instantly collapses and a liquid jet is formed while when We 6 15 and Re 6 100 the rising bubble reaches a well balanced stationary shape. 6. Conclusions To clarify the cavitation mechanisms of three engineering liquids, a systematic observation of the cavitation bubble’s behaviour by means of high-speed photography is conducted. In all the critical regions, refrigerant bubbles were monitored to maintain a ceaselessly oscillating motion across their periphery (‘‘wobbling effect’’) avoiding regular implosions and obtaining an amoeboid formation. This unusual behaviour of the refrigerant cavitation bubbles is effectively determined by the combination of the Reynolds and Weber numbers in given flow conditions. Analytical results showed that in the refrigerant environment the bubbles obtain large Reynolds numbers and low Weber numbers showing a higher resistance to implosion. In the other two environments the Weber number is significantly higher and the Reynolds number substantially lower indicating a tendency of water and lubricant bubbles towards collapsing. The characteristics of all the tested cavitation bubbles are presented in terms of the velocity and average jet velocity values as calculated and with the variation of the bubble radius over time, as obtained from direct observations using high-speed photography. Lubricant bubble is determined to have the highest bubble velocity and average jet velocity among the testing liquids closely followed by the water bubble. The refrigerant bubble having the lowest bubble velocity and average jet velocity in correlation with the absence of regular implosions, cannot be considered as a destructive mean of cavitation for scroll expander systems or similar automotive industrial units. The camera observations of the interaction cavitation mechanisms, the analytical results and the
1553
interesting behaviour of the refrigerant bubbles give rise to the rich variety of cavitation mechanisms and effects which will be examined in future studies. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.expthermflusci.2011.07.005. References [1] I. Tzanakis. M. Hadfield, Z. Khan, Durability of Domestic Scroll Compressor Systems, Surface Effects and Contact Mechanics, 9th International Conference on Surface Effects and Contact Mechanics – Computational Methods and Experiments, Contact/Surface 2009, WIT Press, 2009, pp. 229–240. doi:10.2495/SECM090211. ISBN:978-184564186-3. [2] I. Tzanakis, M. Hadfield, A. Georgoulas, N. Kotsovinos. Cavitation Damage Observations Within Scroll Expander Lubrication Systems, 3rd International Conference on Tribology and Design, in press. [3] E.A. Brujan, P.R. Williams, Cavitation phenomena in non-newtonian liquids, Chemical Engineering Research and Design 84 (2006) 293–299. [4] P. R Williams, P.M. Williams, S.W.J. Brown, A study of liquid jets formed by bubble collapse under shock waves in elastic and Newtonian liquids, Journal of Non-Newtnian Fluid Mechanics 76 (1998) 307–325. [5] Y. Meged, C.H. Venner, W.E. ten Napel, Classification of lubricants according to cavitation criteria, Wear 186 (1995) 443–453. [6] B. Schneider, A. Kosar, Y. Peles, Hydrodynamic cavitation and boiling in refrigerant (R-123) flow inside microchannels, International Journal of Heat and Mass Transfer 50 (2007) 2838–2854. [7] E.A. Brujan, A.F.H. Al-Hussany, R.L. Williams, P.R. Williams, Cavitation erosion in polymer aqueous solutions, Wear 264 (2008) 1035–1042. [8] S. Popinet, S. Zaleski, Bubble collapse near a solid boundary: a numerical study of the influence of viscosity, Journal of Fluid Mechanics 464 (2002) 137–163. [9] P.R. Williams, P.M. Williams, S.W.J. Brown, A study of liquid jets formed by bubble collapse under shockwaves in elastic and Newtonian liquids, Journal of Non-Newtonian Fluid Mechanics 76 (1998) 307–325. [10] A. Berker, M.G. Bouldin, S.J. Kleis, W.E. Van Arsdale, Effect of polymer on flow in journal bearings, Journal of Non-Newtonian Fluid Mechanics 53 (1995) 333–347. [11] B. Karunamurthy, M. Hadfield, C. Vieillard, G.E. Morales-Espejel, Z. Khan, Erosion and rolling contact wear mechanisms in silicon nitride hybrid bearings, Ceramics International Journal, in press. [12] S. Behnia, A. Jafari, W. Soltanpoor, O. Jahanbakhsh, Nonlinear transitions of a spherical bubble, Chaos, Solions and Fractals 41 (2009) 818–828. [13] Y. Iwai, S.Li. Cavitation, Cavitation erosion in waters having different surface tensions, Wear 254 (2003) 1–9. [14] X.M. Liu, J. He, J. Lu, X.W. Ni, Effects of surface tension on a liquid-jet produced by the collapse of a laser-induced bubble against a rigid boundary, Optics and Laser Technology 41 (2009) 21–24. [15] H. Soyama, Y. Yanauchi, K. Sato, R. Oba, R. Oshima, High-speed observation of ultrahigh-speed submerged water jets, Experimental Thermal and Fluid Science 12 (1996) 411–416. [16] I. Tzanakis, M. Hadfield, N. Garland, Cavitation damage incubation with typical fluids applied to a scroll expander system, Tribology International, in press, Available Online 21 June 2011. [17] R. Etienne, L. Jacques, F. Mohamed, M. Peter, A. François, Cavitation bubble behaviour inside a liquid jet, Physics of Fluids 19 (2007) 067106. [18] X. Chen, R.Q. Xu, Z.H. Shen, J. Lu, X.W. Ni, Optical investigation of cavitation erosion by laser-induced bubble collapse, Optics and Laser Technology 36 (2004) 197–203. [19] Li-Xin Bai, Wei-Lin Xu, Zhong Tian, Nai-Wen Li, A high-speed photographic study of ultrasonic cavitation near rigid boundary, Journal of Hydrodynamics 20 (2008) 637–644. [20] O. Lindau, W. Lauterborn, Investigation of the Countrjet Developed Cavitation Bubble that Collapses Near a Rigid Boundary, CAV 2001, 4th International Symposium on Cavitation, 2001. [21] C. Ohl, T. Kurz, R. Geisler, O. Lindau, W. Lauterborn, Microimplosions: cavitation collapse and shock wave emission on a nanosecond time scale, The Royal Society 357 (1999) 269–294. [22] R. Pecha, Microimplosions: cavitation collapse and shock wave emission on a nanosecond time scale, Physics 84 (2000) 1328–1330. [23] R.P. Tong, W.P. Schiffers, J.S. Shaw, J.R. Blake, D.C. Emmony, The role of splashing in the collapse of the laser-generated cavity near a rigid boundary, Journal of Fluid Mechanics 380 (1999) 339–361. [24] E.A. Brujan, T. Ikeda, Y. Matsumoto, On the pressure of cavitation bubbles, Experimental Thermal and Fluid Science 32 (2008) 1188–1191. [25] Y. Matsumoto, J.S. Allen, S. Yoshizawa, T. Ikeda, Y. Kaneko, Medical ultrasound with microbubbles, Experimental Thermal and Fluid Science 29 (2005) 255– 265. [26] A.H. Lefebvre, Atomization and Sprays, Hemisphere Pub., Corp., New York, 1989. [27] B. Schneider, A. Kosar, Y. Peles, Hydrodynamic cavitation and boiling in refrigerant (R-123) flow inside microchannels, Heat and Mass Transfer 50 (2007) 2838–2854.
1554
I. Tzanakis et al. / Experimental Thermal and Fluid Science 35 (2011) 1544–1554
[28] C. Mishra, Y. Pelesa, Flow visualization of cavitating flows through a rectangular slot micro-orifice ingrained in a microchannel, Physics of Fluids 17 (2005). [29] C. Ohl, M. Arora, R. Ikink, N. Jong, M. Versluis, M. Delius, D. Lohse, Sonoporation from jetting cavitation bubbles, Biophysical Journal 91 (2006) 4285–4295.
[30] H. Wang, Z. Zhang, Y. Yang, H. Zhang, Viscosity effects on the behaviour of a rising bubble, Journal of Hudrodynamics 22 (2010) 81–89. [31] H. Wang, Z. Zhang, Y. Yang, H. Zhang, Surface tension effects on the behaviour of a rising bubble driven by buoyancy force, Chinese Physics B 19 (2010) 026801-1–026801-9.