- Email: [email protected]
Experimental study of direct contact vaporization heat transfer on n-pentane-water flowing interface
Energy 93 (2015) 854e863
Contents lists available at ScienceDirect
Energy journal homepage: www.elsevier.com/locate/energy
Experimental study of direct contact vaporization heat transfer on n-pentane-water flowing interface Yiping Wang a, b, Hailing Fu a, Qunwu Huang a, *, Yong Cui c, Yong Sun b, Lihong Jiang a a
School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, China School of Architecture, Tianjin University, Tianjin 300072, China c Tianjin University Research Institute of Architectural Design, Tianjin 300072, China b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 2 March 2015 Received in revised form 3 August 2015 Accepted 21 September 2015 Available online xxx
The direct contact vaporization heat transfer is studied on a small circular interface which is a direct contact interface between the n-pentane injected in a tubule and the immiscible hot water flowing in the channel at the high velocity turbulent state. The interface water temperature is measured by infrared thermograph to obtain the actual driving temperature difference. The effects of water flow velocity and temperature on heat transfer coefficient have been investigated experimentally. In addition, the vapor bubbles characteristics on interface are investigated by visualization research. The results show that the actual driving temperature difference of 8.92 C is far lower than the traditional temperature difference of 37.9 C, which causes that their heat transfer coefficients have more than 4 times deviation. The heat transfer coefficient increases as the water flow velocity increases, but decreases with the increase of the driving temperature difference. The n-pentane vaporization rate increases gradually with an increase of water flow velocity and the actual driving temperature difference. The bubbles diameters increase as the water temperature increases, which causes that it is easier to form gas film to reduce the heat transfer coefficient. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Direct contact heat transfer Heat transfer coefficient Temperature difference Immiscible liquids Boiling
1. Introduction The worldwide energy problems demand the energy utilization efficiency to be increased. Renewable energy such as solar energy, ocean energy etc. has been put into use [1e3]. Many manufacturing processes and most industrial chemical reactions generate heat that must be removed in order to maintain standard operating parameters [4,5]. Therefore, the cooling problem has become an increasing critical problem in concentration photovoltaic systems, heat pump systems, desalination seawater, geothermal heat recovery, thermal energy conversion and storage systems [6e10]. The phase change heat transfer is the most effective method to deal with those problems [11,12]. In recent years, due to the advantages of direct contact heat transfer with lower driving temperature difference, higher effective heat transfer coefficients, fewer scaling problems, a relatively simpler design and no surface corrosion than the traditional heat transfer [13], much attention is devoted to the application of direct contact heat transfer between two immiscible * Corresponding author. Tel./fax: þ86 22 27404771. E-mail address: [email protected] (Q. Huang). http://dx.doi.org/10.1016/j.energy.2015.09.094 0360-5442/© 2015 Elsevier Ltd. All rights reserved.
liquids [14,15]. The most important parameter of heat transfer is the heat transfer coefficient, which determines the heat transfer efficiency [16,17]. The direct contact vaporization heat transfer of the conventional stratified flow with two layer of immiscible liquids in horizontal and inclined liquideliquid systems has been studied in many literature [18,19]. Gollan [20] presented a study of the pentane thin film surface evaporation while flowing co-currently and in direct contact with a laminar water film down a solid plane. A number of mathematical models including an exact formulation and simpler approximate models were considered. The varying thickness of the upper layer was treated by using finite difference techniques. Bentwich [21] analyzed the temperature distribution in a stratified two-phase laminar flow using a Graetz-type solution, and taking into account the discontinuities in the physical properties at the liquideliquid interface. Nosoko [22] studied the evaporation of single volatile-liquid lenses placed on the surface of a quiescent pool of immiscible, denser, less-volatile liquid. Relatively much work had been done to obtain experimental and theoretical data for stratified flow in laminar systems or stagnant liquid medium because that two layer liquid were easy to be entrained or mixed at
Y. Wang et al. / Energy 93 (2015) 854e863
Nomenclature A D Dd H K Ki Q t T Ts Tp Tw
heat transfer area, m2 vapor bubble diameter, mm average bubble diameter, mm latent heat of n-pentane vaporization, kJ/kg heat transfer coefficient, kW/(m2 C) heat transfer coefficient based on TTsat, kW/(m2 C) total heat transmission, kJ/s complete vaporization time of n-pentane, s bulk water temperature, C actual surface water temperature, C measured surface water temperature, C measured interface water temperature, C
high velocity turbulent state. Some stratified flow models were developed for describing heat transfer across the interface of two independent stirred liquids [23e26], but the agitation interrupted the conduct of experiments and it was difficult to obtain accurate quantitative data [27]. The study of liquideliquid interface vaporization heat transfer is a vast scientific topic, one can cite the liquideliquid interface heat transfer of drobble when dispersed phase n-pentane was injected into continuous phase water because the heat transfer taking place at the liquideliquid interface of drobble mostly had been substantiated by Sideman [28]. In practice, the volatile liquid phase is not always located at the bottom of two phase bubble in the evaporation process as described by theoretical models [29,30]. Simpson [31] observed that the drop oscillated during its rise, which caused the unevaporated liquid to slosh from side to side. Over past years, numerous extensive reviews of experimental and theoretical investigations about the vaporization of an immiscible liquid drop in a continuous liquid were published [32,33]. Sideman [34] also studied the heat transfer coefficient of n-pentane liquid drops evaporating in another immiscible liquid on the assumption that the drop is a sphere with no heat flux across the surface. The driving temperature difference was defined as the difference between the average temperature of the water and n-pentane liquid drop, and the heat transfer area was taken as the area of spheroidal drobble. Simpson [31] defined temperature difference as the difference between the water temperature and butane normal saturation temperature, while the heat transfer area was defined as the equivalent spherical area of the drobble. Kulkarni [35] presented a model to predict the interfacial area and direct contact heat transfer coefficient. The results showed that the volume fraction, mass fraction of the npentane liquid and vapor, interfacial area for heat transfer were governed by initial drop size and initial temperature difference between the hot fluid and n-pentane saturation temperature. The study of interfacial vaporization heat transfer in two immiscible liquid is a significant research topic since key parameters, especially actual driving temperature difference and heat transfer coefficient, are expected to depend on the interfacial heat transfer behavior [36,37]. When n-pentane begins to vaporize by absorbing heat from the liquideliquid interface, the interface water temperature decreases rapidly, and there will be the difference temperature between the interface water and bulk water. The liquideliquid direct contact heat transfer interfacial area varies continuously because of the continuing evaporation, system oscillation and sloshing of the unvaporized n-pentane liquid. On the other hand, most of direct contact heat transfer in stratified flow happened on a large interface where the flow velocities are not
Ti Tsat DT DTm u V
y r
855
average interface water temperature, C n-pentane saturation temperature, C actual driving temperature difference, C traditional driving temperature difference, C water flow velocity, m/s n-pentane vaporization volume, m3 n-pentane vaporization rate, kg/(m2 s) n-pentane liquid density, kg/m3
Abbreviations CMOS complementary metal oxide semiconductor IR infrared PE polyethylene
uniform and unstable especially at the turbulent state. Hence, the change of interface leads to the actual driving temperature difference and heat transfer coefficient cannot be obtained accurately. To solve this problem, a visualized experiment was fabricated to investigate the direct contact vaporization heat transfer between two immiscible liquids at turbulence flow state. In this work, a special circular interface with small unchanged area was designed to make the water flow velocities uniform on the entire interface. The actual driving temperature difference and heat transfer coefficient are obtained by using this n-pentane-water interface which also could be regarded as the liquideliquid heat transfer interface of drobbles. The influence factors of heat transfer coefficient are analyzed and the boiling phenomenon is observed when the contact interface constantly updates in the condition of turbulent flow. It is also anticipated that the results obtained in this study will provide a useful method to research the mechanism of immiscible liquids direct contact heat transfer, which will be used for better understanding the future design of gaseliquideliquid three phase exchangers. 2. Experimental apparatus and design of the test interface 2.1. Experimental apparatus and procedure The experimental apparatus was shown schematically in Fig. 1. The rectangular flow channel (internal dimensions of 10 mm in
Fig. 1. Schematic of experimental system for the direct contact heat transfer.
856
Y. Wang et al. / Energy 93 (2015) 854e863
width, 10 mm in height and 500 mm in length) was made by PE (Polyethylene) plate with a circular hole on the upper plate of the channel. The experiments were carried out in a tubule with an inner diameter of 8 mm and 30 mm in height, which was adhered on the circular hole of the rectangular channel. The lighter volatile npentane liquid added into the tubule was in direct contact with the immiscible hot water below the circular hole and the vaporization heat transfer of n-pentane taking place on the circular pentanewater interface. The pipeline of the system was stainless steel tube with an inner diameter of 20 mm and each two jointing sides were air proofed by air-proof glue in order to avoid entering of gases during the experiment. The entire pipeline was enveloped by thermal insulation layer to reduce heat loss except for the test section. The measurement principle is to calculate the heat flow rate by measuring the complete vaporization time of a certain volume npentane on the small circular interface whose area is not changed. Before each experiment, the water in the pipeline was heated to boiling for three hours in order to remove the gas in it and then was cooled down to the required temperature. Water passed through the shell and tube heat exchanger which must be used to maintain the uniform water temperature, rotor flowmeter (LZB-25, Jiangsu YuYing Automation Instrument Co., Ltd., China), PE rectangle flow channel and was circulated by a shielding pump (GDR32/6, WILO SE, Germany). During the process of experiment, the water flow velocity was controlled using the valve at the inlet of flowmeter. To reduce the fluctuation of water caused by the vibration of pump and intense turbulence of hot water, both sides of the pump were connected with two-stage hoses. A fine-tuning valve was installed to regulate the water layer height in the circular hole. It was necessary to remove the gas in the pipeline because water circulation was required to maintain the steady state. Water tank was used to provide make-up water and a valve in the outlet of it was closed during the experiment. When the fluctuation of liquideliquid interface reached a relatively steady state, with the aid of microsyringe (SPLab01, Baoding ShenChen precision pump Co., Ltd., China), n-pentane in the saturation temperature was injected into the tubule along its inner wall. At the same time, the evaporation time of n-pentane was recorded by a chronometer (DA-8SZ, Timing resolution: 1 ms, Xuchang ChangAn Technology Co., Ltd., China). In order to observe the phenomenon of evaporation, n-pentane was dyed by Oil Soluble Red. The bulk temperature of hot water was measured by a T-type thermocouple with the diameter of 0.3 mm positioned at the inlet of the rectangular flow channel. 2.2. Design of the pentane-water direct contact interface The design of pentane-water interface was important in this experiment because when the area of the interface is not changed and the water flow velocity on the entire interface is uniform everywhere, it is easy to accurately determine the heat transfer coefficient. As shown in Fig. 2b, the upper and lower openings of the circular hole in the upper plate of the rectangular flow channel were both designed in the shape of a circular truncated cone to reduce the influence of the velocity boundary layer. The schematic and dimensions of the heat transfer interface were shown in Fig. 2a and the angle between generatrix and basal diameter of the lower circular truncated cone was relatively small to weaken the influence of the rise of water level at left side of the circular hole on the interface at high flow velocity of water. The cec' was considered as the diameter of n-pentane-water direct contact circular interface.
Fig. 2. Schematic of direct contact interface.
0.25 mm) with spacing of about 5 mm [38]. The conductance probes were put into the tubule and the tips of probes were just in contact with the surface of water filled in the rectangle flow channel. The variable voltage signal in proportion to water layer height was obtained by using a conductometer connected with the parallel wire conductance probes. The analog voltage signal from the conductometer was sampled using a data logger (DI-710, Quatronix Company, USA) with sampling frequency of 1 kHz. Actual water layer height should be obtained by a linear relation between water layer height and voltage. A high-speed CMOS (complementary metal oxide semiconductor) digital camera (EOS70D, Canon Company, Japan) was used to capture the photographs of vaporization phenomenon on the interface. And the acquired images were color images with a resolution of 5472 3648 pixels. As shown in Fig. 1 the camera was positioned at a place of 5 cm away from bottom of the interface with illumination lamp as the light source at the same side. According to the image pixels, the coordinate y-up, y-down in the vertical direction, and coordinate x-left, x-right in horizontal direction of the departure bubbles were calculated. Thus, the diameter of bubble in horizontal and vertical directions was expressed as follows [39]:
Dx ¼ xright xleft
(1)
Dy ¼ yup ydown
(2)
Then, the average bubble diameter was expressed as follow:
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi D2x þ D2y
2.3. Measuring methods
Dd ¼
Variations of water layer height were measured by parallel wire conductance probes which consisted of two wires (diameters of
Liquid temperature profiles of liquideliquid heat transfer interface were analyzed by an non-contact measurement technique
2
(3)
Y. Wang et al. / Energy 93 (2015) 854e863
857
in order to avoid disturbing the fluid interface during the boiling process. Infrared thermography was chosen because of its good signal strength and a large temperature range. A material that was transparent in the visible range and infrared would allow using the optical temperature measurement technique: Infrared thermography to determine the internal water surface temperature [40]. One material that offered the desired characteristics was PE. The infrared thermography was carried out with a thermal infrared imager (SAT-90, SAT Company, China). It had a resolution of 640 480 pixels, a spectral range of 8e12 mm and was used in combination with a 50 mm lens. The average transmittance of IR (infrared) radiation in this wavelength ranges in PE plate and npentane was about 85% and 90%, respectively. The complete vaporization time of n-pentane was recorded by a chronometer. When the n-pentane started to contact with the water and then vaporized completely, some response time was needed to control the chronometer, which would cause some errors in the measurement of evaporation time. But this relative error is less than 0.24%, which is less than temperature and flow velocity measures errors and could be ignored in this experiment. Accuracy of measuring operational data is essential and errors should be reduced to ensure sufficient accuracy of the follow-up calculations [41]. Therefore, all the experiments are repeated 25 times in each group under the same conditions and Grubbs procedure was used to detect outliers [42]. Standard deviation maps were made to analyze the experiment and put into the section of results and discussions. 3. Results and discussions In this section, the experimental results for actual driving temperature difference, heat transfer coefficient and n-pentane vaporization rate are described and discussed in detail. The effects of water flow velocity and temperature difference on heat transfer performance are analyzed. 3.1. Actual driving temperature differences Fig. 3. Calibration curve of water temperature.
3.1.1. Temperature calibration Infrared thermography is a well-established technique for temperature and heat transfer measurements [43]. However, in practical applications, it is difficult to distinguish whether the infrared radiance is emitted from the test article or the contributions of surroundings. Furthermore, the infrared signal emitted from the test article depends on its emissivity, the transmissivity of the optical path and the sensitivity of the detector. Obviously, there is a difference between the actual water temperature and the measurement temperature. Because of that, an in-situ calibration is strongly recommended for accurate temperature measurement. Additionally, the target's emissivity and the reflective temperature had been known for the calibration and accurate results. The water surface temperature in rectangular flow channel is measured by the thermal infrared imager above the PE plate and the calibration curve is fitted to the water surface temperature measurement by a PT100 thermal resistance. Therefore, according to Fig. 3a, the actual water surface temperature Ts in the rectangular flow channel could be calibrated by the following equation:
Ts ¼ 1:0428 Tp
(4)
where Tp is the measured surface water temperature. The n-pentane with the temperature of 35 C was injected into the tubule and the water temperatures varied from 5.5 C to 35.3 C which are always lower than the n-pentane boiling point. The same method should be applied to calibrate the liquideliquid direct
contact interfacial water temperature which is measured by thermal infrared imager through the n-pentane liquid layer. As can be seen from Fig. 3b, the actual interface water temperature Ti could be calibrated by the following equation:
Ti ¼ 1:0246 Tw
(5)
where Tw is the measured interface water temperature. 3.1.2. Actual driving temperature differences When n-pentane is in direct contact with hot water, with the increases of water temperature, there would be n-pentane vapor bubbles on the interface of n-pentane and water. As the vaporization, n-pentane vapor bubbles would become bigger and leave away from interface finally, then the n-pentane liquid nearby is rapidly added to the liquideliquid interface at the same time. In the phase change process of n-pentane liquid, the strong perturbation on the n-pentane-water interface destroys the boundary layer and reduces the heat transfer resistance which greatly enhances the heat transfer rate. There are mainly two types of heat transfer including the convective heat transfer in severely turbulent hot water and n-pentane vaporization heat transfer on the direct contact liquideliquid interface. In order to determine the heat transfer coefficient accurately, the actual driving temperature difference must be measured. Due
858
Y. Wang et al. / Energy 93 (2015) 854e863
to the severe turbulence of hot water, the nucleation sites became very unstable, which caused uneven distribution of temperature on the boiling interface. The temperatures at different locations of the boiling interface were measured by a thermal infrared imager and the average temperature was taken to represent the interface water temperature. N-pentane at the saturation temperature is injected to the interface through a circular tubule. Based on the measured interface water temperature, the actual driving temperature difference DT can be defined by the following equation:
DT ¼ Ti Tsat
(6)
where Tsat is the n-pentane saturation temperature, Ti is the average interface water temperature. This actual driving temperature difference has a large difference from the traditional driving temperature difference defined by the calculating equation:
DTm ¼ T Tsat
(7)
where T is the bulk water temperature. As seen from Table 1, when the bulk water temperature T reaches 74 C, the traditional driving temperature difference DTm is 37.9 C. However, the average interface water temperature Ti is 45.02 C, and the actual driving temperature difference DT is 8.92 C. The actual driving temperature difference DT was lower than traditional driving temperature difference DTm. This is because the vaporization of n-pentane takes away heat from direct contact interface, which reduces the interface water temperature rapidly. The flowing hot water continuously supplies heat for the liquideliquid interface to maintain its temperature relatively stable. The actual temperature differences at different water temperatures and flow velocities are also shown in Table 1. As the bulk water temperature increases, the interface water temperature increases due to higher temperature gradient and convection heat transfer from the hot bulk water to the water of liquideliquid interface, which means the actual driving temperature difference DT increases with the increase of water temperature. But this increasing amplitude is far less than that of traditional driving temperature difference DTm. With the increase of water flow velocity, the interface water temperature do not have a large decrease. This may be because water flow velocity plays a more important role in making heat transfer boundary layer thinner than increasing water temperature gradient. 3.1.3. Water temperature profile Water surface temperatures profile along the center axial line of the circular heat transfer interface is measured at the top of PE plate using a thermal infrared imager. The typical water temperatures distribution at various axial locations are shown in Fig. 4. The zero point of X axis represents the center of circular interface and the upper wall of tubule has the lowest temperature. From Fig. 4, it is clear that a large temperature gradient exists on the left side of
interface close to the outlet of the rectangular flow channel. It shows that the interface water temperature is cooled down by the evaporation of n-pentane and then quickly heated to the bulk water temperature when it leaves way from the interface. The reasons can be related to the intense turbulence and mixing in hot water. As flow velocity increases from 0.56 m/s to 2.78 m/s, the temperature gradients increases from 2.306 С/mm to 3.222 С/mm at 74 С. With water increasing from 44 С to 74 С, the temperature gradient increases from 0.667 С/mm to 2.887 С/mm at 1.67 m/s. Because that thermal interaction between water and n-pentane is enhanced with the increase of flow velocity and bulk water temperature. The temperatures in the edge of circle interface are lower than that in its center because that solid contact angle can make it easier to form nucleation sites and vaporize n-pentane to reduce interface water temperatures. 3.2. Effect of water temperature and flow velocity on heat transfer coefficient When n-pentane temperature reaches saturation temperature it will lose the sensible heat and there will be no further change in npentane temperature [5]. According to the mass and time of npentane vaporization completely, the total heat transmission Q between the n-pentane and hot water on the direct contact interface can be found as follows:
Q¼
rVH t
(8)
where t is the time of n-pentane vaporization completely; V is the volume of n-pentane vaporization. The heat transfer coefficient on the n-pentane-water interface can be written as follows:
K¼
Q A$DT
(9)
where A is the area of n-pentane-water direct contact interface. Fig. 5 shows the comparison of the heat transfer coefficients obtained by two kinds of driving temperature differences which are the actual driving temperature difference DT and the traditional driving temperature difference DTm. It can be seen that there is a large deviation between the heat transfer coefficient obtained by two different kinds of driving temperature difference. When the bulk water temperature is 44 C, the heat transfer coefficient obtained by DT is 2.734 times as large as that by DTm. When the bulk water temperature is 74 C, the heat transfer coefficient obtained by DT is 3.258 times larger than that by DTm. The heat transfer coefficient increases gradually with the increases of water flow velocity can be seen in Fig. 6. At the low temperature difference of 2.86 C, with the increase of flow velocity from 0.56 m/s to 2.22 m/s, the heat transfer coefficient increases from 15.002 kW/(m2 C) to 33.859 kW/(m2 C). Such evolution can
Table 1 Actual temperature differences for different bulk water temperatures. Flow velocity (u, m/s)
N-pentane saturation temperature (Tsat, C)
Bulk water temperature (T, C)
Average interface water temperature (Ti, C)
Traditional temperature difference (DTm, C)
Actual temperature difference (DT, C)
1.67 1.67 1.67 1.67 0.56 2.22 2.78
36.1 36.1 36.1 36.1 36.1 36.1 36.1
44.00 54.00 64.00 74.00 74.00 74.00 74.00
38.96 41.08 42.82 45.02 45.11 44.84 44.81
7.9 17.9 27.9 37.9 37.9 37.9 37.9
2.86 4.98 6.72 8.92 9.01 8.74 8.71
Y. Wang et al. / Energy 93 (2015) 854e863
859
Fig. 4. Temperature profile of interface for different bulk water temperature and flow velocity.
be attributed to an increase frequency of interface update when the flowing hot water go through the interface faster. The violent agitation of interface reduces the heat transfer boundary lay and the heat transfer resistance, which increases the thermal energy transfer and enhances the heat transfer coefficient. In the process of npentane vapor bubbles rising up with the entraining of n-pentane liquid from the bottom layer to the top layer of n-pentane liquid, the increase of bubbles escape and rising velocities promotes the perturbation and enhances the heat transfer coefficient between the two immiscible liquid. When the flow velocity reaches 2.5 m/s, the value of heat transfer coefficient is approximately 35.748 kW/ (m2 C) and more likely to be stable. In addition, the heat transfer coefficient is 22.079 kW/(m2 C) when the water flow velocity reaches 1.94 m/s at the highest temperature difference of 8.92 C and the increase of heat transfer coefficient is not obvious. This may be explained mainly by the balance of heat transfer including the positive effect of the increasing frequency of interface update in promoting the heat supply with the negative effect of excessive interaction between vapor bubbles. On the other hand, the increase of temperature difference weakens the affection of flow velocity.
Heat transfer coefficients are determined for each of the actual driving temperature difference and plotted in Fig. 7, we can see that heat transfer coefficient increases with the decrease of temperature difference, in other words, there is much higher heat transfer efficiency at a lower water temperature. This trend is more obvious with the increase of water flow velocity. During the process of experiment, it has been observed that an increase of temperature is accompanied with the increase of bubbles departure diameter and this might decrease considerably the actual contact area between hot water and n-pentane, which leads to lower values of heat transfer coefficient. Particular emphasis is given to the horizontal coalescence of the departed bubbles close to the interface in this experiment, which is observed to cause the deterioration of the heat transfer coefficient [44], due to the formation and coalescence of large vapor bubbles that prevents the fresh bulk liquid from accessing the heat surface for evaporation heat transfer. The gas film with high thermal resistance also reduces the heat transfer efficiency sharply. This affection should be considered in this paper because of the existence of solid wall around the interface.
860
Y. Wang et al. / Energy 93 (2015) 854e863
Fig. 5. The heat transfer coefficient obtained by DT and DTm for: (a) T ¼ 44 С (b) T ¼ 54 С (c) T ¼ 64 С (d) T ¼ 74 С.
3.3. N-pentane vaporization rate
v¼
rV At
(10)
The influences of natural convection and diffusion on the npentane vaporization rate is far less than that of boiling heat transfer and could be ignored. The vaporization mass of n-pentane on 1 m2 direct contact interface per time unit is considered as the vaporization rate. The n-pentane vaporization rate n can be determined from the following expression:
It can be seen from Fig. 8 that the n-pentane vaporization rate increases gradually with an increase of water flow velocity and the actual driving temperature difference. As flow velocity increases, the turbulence is more serious and the convection between liquideliquid interface and bulk water is more intense, which
Fig. 6. Variations of n-pentane heat transfer coefficient with water flow velocity.
Fig. 7. Variations of n-pentane heat transfer coefficient with actual driving temperature difference.
Y. Wang et al. / Energy 93 (2015) 854e863
861
velocity higher than 1.39 m/s and 1.94 m/s at the temperature difference of 8.92 C and 2.86 C, respectively. Because the liquideliquid interface will be covered by a layer of gas film which isolates the n-pentane and water, which deteriorates the heat transfer. At a lower temperature difference, there is a larger value of heat transfer coefficient but a longer time is needed to complete vaporization of n-pentane liquid. In engineering design, we should find a balance between the heat transfer coefficient and the vaporization rate.
3.4. Visualization research of boiling on interface
Fig. 8. Variations of n-pentane vaporization rate with water flow rate for different temperature differences.
decreases the time of n-pentane vaporization completely. On the other hand, such evolution could be partially explained by the increase of the heat transfer coefficient at the liquideliquid interface with the increase of the water flow velocity. As the actual driving temperature difference increases, the heat flux increases and time of n-pentane vaporization completely decreases, which can show that temperature has a positive effect on vaporization rate. The increase of vaporization rate becomes not obvious when the flow
The boiling of flow interface in direct contact immiscible liquids is different from the boiling with static heat transfer surface. The npentane would appear to have an unusual type of nucleation because a nucleation site at the interface might become displaced by the agitation of boiling and perturbation of the intense turbulent of water. When the water in the rectangular flow channel is in the state of turbulent flow, a single bubble in the nucleate flow boiling process would go through five stages which are formation, growth, sliding, rising up and breakup. In the stages of bubble rising up and breakup, both vaporization of n-pentane and convection would occur in the n-pentane-water interface. The boiling heat transfer phenomena is observed at the bottom of the liquideliquid interface, however, it is difficult to record the process of n-pentane bubble formation and growth up because of the overlap and shelter from other vapor bubbles. Fig. 9 shows examples of the boiling pattern.
Fig. 9. Boiling pattern at u ¼ 0.56 m/s for: (a) T ¼ 44 С (b) T ¼ 54 С (c) T ¼ 64 С (d) T ¼ 74 С.
862
Y. Wang et al. / Energy 93 (2015) 854e863
flow velocity. With an increase of temperature, the number of nucleation sites decrease and the coalescence of large size vapor bubbles increases significantly, which has a negative effect on the heat transfer coefficient. The largest heat transfer coefficient of 35.329 kW/(m2 C) is obtained at the highest water flow velocity of 2.78 m/s and the lowest driving temperature difference of 2.86 C in this experiment. The n-pentane vaporization rate increases gradually with an increase of water flow velocity and driving temperature difference. The new experimental method and the results obtained in this study provide a foundation to research the mechanism of immiscible liquids direct contact vaporization heat transfer. Acknowledgments
Fig. 10. Variation of bubble diameter with bulk water temperature.
Growth and departure diameters of bubbles directly influence the boiling heat transfer efficiency [45]. In the process of boiling, the bubbles are not evenly distributed on the entire heat transfer interface. Due to the nonlinear interaction between the nucleation sites, the distribution of interfacial superheat is very uneven and fluctuating. With the increase of the degree of superheat and flow velocity, the nonlinear interaction between the nucleation sites would become stronger. Because of this nonlinear factors, the boiling system could produce chaos phenomenon and besides the nucleation site or bubble density and bubble diameters have obvious random characteristics. For this reason, the standard deviation of average bubble diameters departing from the heater interface under different water temperature are marked out in Fig. 10. It can be seen that when the flow velocity is 1.67 m/s and the bulk water temperature is 44 С, the pentane vapor bubbles depart from the interface with an average diameter of approximately 0.741 mm. The diameter increases as the water temperature increases, and the average diameter of n-pentane vapor bubble increases to 1.962 mm at a high water temperature of 74 C. The number of n-pentane vapor bubbles significantly decreases with the increase of water temperature, which shows the decrease of nucleation site. The relatively small amounts of nucleation sites and large sizes of vapor bubbles at higher water temperature decreases the heat transfer coefficient. 4. Conclusions A new experimental approach has been brought out to study the n-pentane-water direct contact vaporization heat transfer on a small circular interface whose area is not changed and on which flow velocities are uniform. The following conclusions are drawn after analyzing the experimental results: The actual driving temperature difference DT between interface water temperature and n-pentane saturation temperature is much smaller than the traditional temperature difference DTm, which causes that the actual heat transfer coefficient is larger than that obtained by the traditional algorithm. The actual driving temperature difference DT increases with bulk water temperature increasing, but the increasing amplitude is far less than that of traditional temperature difference DTm. In addition, the water flow velocity has no significant effect on the interface water temperature. Heat transfer coefficient increases with the increase of the water flow velocity, and this increase is no more pronounced at higher
Acknowledgments are given to Dr. Lianwei Qin, Lihong Jiang and Ganchao Xin in Green Energy Laboratory of Tianjin University for their contribution on language editing and results analysis to this paper. References [1] Xie J, Xu J, Xing F, Wang Z, Liu H. The phase separation concept condensation heat transfer in horizontal tubes for low-grade energy utilization. Energy 2014;69:787e800. [2] Li J, Cao W, Chen G. The heat transfer coefficient of new construction e Brick masonry with fly ash blocks. Energy 2015;86:240e6. [3] Ali AHH, Ahmed M, Abdel-Gaied SM. Investigation of heat transfer and fluid flow in transitional regime inside a channel with staggered plates heated by radiation for PV/T system. Energy 2013;59:255e64. [4] Lemouari M, Boumaza M, Kaabi A. Experimental analysis of heat and mass transfer phenomena in a direct contact evaporative cooling tower. Energy Convers Manag 2009;50(6):1610e7. [5] Kiatsiriroat T, Vithayasai S, Vorayos N, Nuntaphan A, Vorayos N. Heat transfer prediction for a direct contact ice thermal energy storage. Energy Convers Manag 2003;44(4):497e508. [6] Peng Z, Yiping W, Cuili G, Kun W. Heat transfer in gaseliquideliquid threephase direct-contact exchanger. Chem Eng J 2001;84(3):381e8. [7] Shimizu Y, Mori YH. Evaporation of single liquid drops in an immiscible liquid at elevated pressures: equipment and preliminary results. Exp Fluids 2004;6(2):73e9. [8] Siqueiros J, Bonilla O. An experimental study of a three-phase, direct-contact heat exchanger. Appl Therm Eng 1999;19(5):477e93. [9] Hamid MOA, Zhang B, Yang L. Application of field synergy principle for optimization fluid flow and convective heat transfer in a tube bundle of a preheater. Energy 2014;76:241e53. [10] Lund H, Werner S, Wiltshire R, Svendsen S, Thorsen JE, Hvelplund F, et al. 4th Generation District Heating (4GDH): integrating smart thermal grids into future sustainable energy systems. Energy 2014;68:1e11. [11] Shin S, Choi G, Kim BS, Cho HH. Flow boiling heat transfer on nanowire-coated surfaces with highly wetting liquid. Energy 2014;76:428e35. [12] Kukulka DJ, Smith R. Heat transfer evaluation of an enhanced heat transfer tube bundle. Energy 2014;75:97e103. [13] Mahood HB, Sharif AO, Al-Aibi S, Hawkins D, Thorpe R. Analytical solution and experimental measurements for temperature distribution prediction of threephase direct-contact condenser. Energy 2014;67:538e47. [14] Ribeiro Jr CP, Lage PLC. Direct-contact evaporation in the homogeneous and heterogeneous bubbling regimes. Part I: experimental analysis. Int J Heat Mass Transf 2004;47(17e18):3825e40. [15] Seetharamu KN, Battya P. Direct contact evaporation between two immiscible liquids in a spray column. J Heat Transf 1989;111(3):780e5. [16] Li Y-Y, Liu Z-H, Zheng B-C. Experimental study on the saturated pool boiling heat transfer on nano-scale modification surface. Int J Heat Mass Transf 2015;84:550e61. [17] Orzechowski T, Wcislik S. Experimental analysis of the drop film boiling at ambient pressure. Energy Convers Manag 2013;76:918e24. [18] Viskanta R, Behnia M. Experimental and analytical study of heat transfer and mixing in thermally stratified buoyant flows. Int J Heat Mass Transf 1982;25(6):847e61. [19] Mukerjee D, Davis EJ. Direct contact heat transfer between immiscible fluid layers in laminar flow. AIChE J 1972;18(1):94e101. [20] Gollan A, Sideman S. Direct contact heat transfer with change of phase: cocurrent flow of immiscible films with surface evaporation. Int J Heat Mass Transf 1968;11(12):1761e70. [21] Bentwich M, Landau U, Sideman S. Direct contact heat transfer with change of phase: evaporation of discrete volatile films from the surface of a stagnant immiscible liquid. Int J Heat Mass Transf 1970;13(6):945e56.
Y. Wang et al. / Energy 93 (2015) 854e863 [22] Nosoko T, Ohyama T, Mori Y. Evaporation of volatile-liquid lenses floating on an immiscible-liquid surface: effects of the surface age and fluid purities in npentane/water system. J Fluid Mech 1985;161:329e46. [23] Szekely J. Notes on the transfer at the interface of two independently stirred liquids. Int J Heat Mass Transf 1963;6(9):833e40. [24] Szekely J. Mathematical model for heat or mass transfer at the bubble-stirred interface of two immiscible liquids. Int J Heat Mass Transf 1963;6(5):417e22. [25] Fortuna G, Sideman S. Direct contact heat transfer between immiscible liquid layers with simultaneous boiling and stirring. Chem Eng Sci 1968;23(9): 1105e19. [26] Sideman S, Uliss I. Direct contact heat transfer between immiscible liquids in stratified flow with simultaneous interfacial boiling. Chem Eng Sci 1972;27(6):1245e56. [27] Iida Y, Takashima T. Direct-contact heat transfer characteristics: evaporation of a drop dropped onto a hot liquid surface. Int J Heat Mass Transf 1980;23(9): 1263e71. [28] Sideman S, Isenberg J. Direct contact heat transfer with change of phase: bubble growth in three-phase systems. Desalination 1967;2(2):207e14. [29] Mori YH. Configurations of gas-liquid two-phase bubbles in immiscible liquid media. Int J Multiph Flow 1978;4(4):383e96. [30] Mori YH. Classification of configurations of two-phase vapor/liquid bubbles in an immiscible liquid in relation to direct-contact evaporation and condensation processes. Int J Multiph Flow 1985;11(4):571e6. [31] Simpson HC, Beggs GC, Nazir M. Evaporation of butane drops in brine. Desalination 1974;15(1):11e23. [32] Sideman S, Hirsch G, Gat Y. Direct contact heat transfer with change of phase: effect of the initial drop size in three-phase heat exchangers. AIChE J 1965;11(6):1081e7. [33] Prakash CB, Pinder KL. Direct contact heat transfer between two immiscible liquids during vaporisation. Can J Chem Eng 1967;45(4):210e4. [34] Sideman S, Taitel Y. Direct-contact heat transfer with change of phase: evaporation of drops in an immiscible liquid medium. Int J Heat Mass Transf 1964;7(11):1273e89.
863
[35] Kulkarni AA, Ranade VV. Direct contact heat transfer via injecting volatile liquid in a hot liquid pool: generation and motion of bubbles. Chem Eng Sci 2013;100:421e32. [36] Tochitani Y, Mori YH, Komotori K. Vaporization of single liquid drops in an immiscible liquid part I: forms and motions of vaporizing drops. Warme Stoffubertragung 1977;10(1):51e9. [37] Tochitani Y, Nakagawa T, Mori YH, Komotori K. Vaporization of single liquid drops in an immiscible liquid part II: heat transfer characteristics. Warme Stoffubertragung 1977;10(2):71e9. [38] Al-Wahaibi T, Angeli P. Experimental study on interfacial waves in stratified horizontal oilewater flow. Int J Multiph Flow 2011;37(8):930e40. [39] Zhang Y, Wei J, Xue Y, Kong X, Zhao J. Bubble dynamics in nucleate pool boiling on micro-pin-finned surfaces in microgravity. Appl Therm Eng 2014;70(1):172e82. [40] Schreivogel P, Pfitzner M. Optical convective heat transfer measurements using infrared thermography and frequency domain phosphor thermometry. Int J Heat Mass Transf 2015;82:299e308. [41] Varbanov PS, Wang Y, Klemes JJ, Lund H. PRES'13 special section: process integration, modelling and optimisation for energy saving and pollution reduction. Energy 2014;75:1e4. [42] Jain RB. A recursive version of Grubbs' test for detecting multiple outliers in environmental and chemical data. Clin Biochem 2010;43(12):1030e3. [43] Kylili A, Fokaides PA, Christou P, Kalogirou SA. Infrared thermography (IRT) applications for building diagnostics: a review. Appl Energy 2014;134: 531e49. [44] Teodori E, Moita AS, Moreira ALN. Characterization of pool boiling mechanisms over micro-patterned surfaces using PIV. Int J Heat Mass Transf 2013;66:261e70. [45] Luke A, Cheng D-C. High speed video recording of bubble formation with pool boiling. Int J Therm Sci 2006;45(3):310e20.
Contents lists available at ScienceDirect
Energy journal homepage: www.elsevier.com/locate/energy
Experimental study of direct contact vaporization heat transfer on n-pentane-water flowing interface Yiping Wang a, b, Hailing Fu a, Qunwu Huang a, *, Yong Cui c, Yong Sun b, Lihong Jiang a a
School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, China School of Architecture, Tianjin University, Tianjin 300072, China c Tianjin University Research Institute of Architectural Design, Tianjin 300072, China b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 2 March 2015 Received in revised form 3 August 2015 Accepted 21 September 2015 Available online xxx
The direct contact vaporization heat transfer is studied on a small circular interface which is a direct contact interface between the n-pentane injected in a tubule and the immiscible hot water flowing in the channel at the high velocity turbulent state. The interface water temperature is measured by infrared thermograph to obtain the actual driving temperature difference. The effects of water flow velocity and temperature on heat transfer coefficient have been investigated experimentally. In addition, the vapor bubbles characteristics on interface are investigated by visualization research. The results show that the actual driving temperature difference of 8.92 C is far lower than the traditional temperature difference of 37.9 C, which causes that their heat transfer coefficients have more than 4 times deviation. The heat transfer coefficient increases as the water flow velocity increases, but decreases with the increase of the driving temperature difference. The n-pentane vaporization rate increases gradually with an increase of water flow velocity and the actual driving temperature difference. The bubbles diameters increase as the water temperature increases, which causes that it is easier to form gas film to reduce the heat transfer coefficient. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Direct contact heat transfer Heat transfer coefficient Temperature difference Immiscible liquids Boiling
1. Introduction The worldwide energy problems demand the energy utilization efficiency to be increased. Renewable energy such as solar energy, ocean energy etc. has been put into use [1e3]. Many manufacturing processes and most industrial chemical reactions generate heat that must be removed in order to maintain standard operating parameters [4,5]. Therefore, the cooling problem has become an increasing critical problem in concentration photovoltaic systems, heat pump systems, desalination seawater, geothermal heat recovery, thermal energy conversion and storage systems [6e10]. The phase change heat transfer is the most effective method to deal with those problems [11,12]. In recent years, due to the advantages of direct contact heat transfer with lower driving temperature difference, higher effective heat transfer coefficients, fewer scaling problems, a relatively simpler design and no surface corrosion than the traditional heat transfer [13], much attention is devoted to the application of direct contact heat transfer between two immiscible * Corresponding author. Tel./fax: þ86 22 27404771. E-mail address: [email protected] (Q. Huang). http://dx.doi.org/10.1016/j.energy.2015.09.094 0360-5442/© 2015 Elsevier Ltd. All rights reserved.
liquids [14,15]. The most important parameter of heat transfer is the heat transfer coefficient, which determines the heat transfer efficiency [16,17]. The direct contact vaporization heat transfer of the conventional stratified flow with two layer of immiscible liquids in horizontal and inclined liquideliquid systems has been studied in many literature [18,19]. Gollan [20] presented a study of the pentane thin film surface evaporation while flowing co-currently and in direct contact with a laminar water film down a solid plane. A number of mathematical models including an exact formulation and simpler approximate models were considered. The varying thickness of the upper layer was treated by using finite difference techniques. Bentwich [21] analyzed the temperature distribution in a stratified two-phase laminar flow using a Graetz-type solution, and taking into account the discontinuities in the physical properties at the liquideliquid interface. Nosoko [22] studied the evaporation of single volatile-liquid lenses placed on the surface of a quiescent pool of immiscible, denser, less-volatile liquid. Relatively much work had been done to obtain experimental and theoretical data for stratified flow in laminar systems or stagnant liquid medium because that two layer liquid were easy to be entrained or mixed at
Y. Wang et al. / Energy 93 (2015) 854e863
Nomenclature A D Dd H K Ki Q t T Ts Tp Tw
heat transfer area, m2 vapor bubble diameter, mm average bubble diameter, mm latent heat of n-pentane vaporization, kJ/kg heat transfer coefficient, kW/(m2 C) heat transfer coefficient based on TTsat, kW/(m2 C) total heat transmission, kJ/s complete vaporization time of n-pentane, s bulk water temperature, C actual surface water temperature, C measured surface water temperature, C measured interface water temperature, C
high velocity turbulent state. Some stratified flow models were developed for describing heat transfer across the interface of two independent stirred liquids [23e26], but the agitation interrupted the conduct of experiments and it was difficult to obtain accurate quantitative data [27]. The study of liquideliquid interface vaporization heat transfer is a vast scientific topic, one can cite the liquideliquid interface heat transfer of drobble when dispersed phase n-pentane was injected into continuous phase water because the heat transfer taking place at the liquideliquid interface of drobble mostly had been substantiated by Sideman [28]. In practice, the volatile liquid phase is not always located at the bottom of two phase bubble in the evaporation process as described by theoretical models [29,30]. Simpson [31] observed that the drop oscillated during its rise, which caused the unevaporated liquid to slosh from side to side. Over past years, numerous extensive reviews of experimental and theoretical investigations about the vaporization of an immiscible liquid drop in a continuous liquid were published [32,33]. Sideman [34] also studied the heat transfer coefficient of n-pentane liquid drops evaporating in another immiscible liquid on the assumption that the drop is a sphere with no heat flux across the surface. The driving temperature difference was defined as the difference between the average temperature of the water and n-pentane liquid drop, and the heat transfer area was taken as the area of spheroidal drobble. Simpson [31] defined temperature difference as the difference between the water temperature and butane normal saturation temperature, while the heat transfer area was defined as the equivalent spherical area of the drobble. Kulkarni [35] presented a model to predict the interfacial area and direct contact heat transfer coefficient. The results showed that the volume fraction, mass fraction of the npentane liquid and vapor, interfacial area for heat transfer were governed by initial drop size and initial temperature difference between the hot fluid and n-pentane saturation temperature. The study of interfacial vaporization heat transfer in two immiscible liquid is a significant research topic since key parameters, especially actual driving temperature difference and heat transfer coefficient, are expected to depend on the interfacial heat transfer behavior [36,37]. When n-pentane begins to vaporize by absorbing heat from the liquideliquid interface, the interface water temperature decreases rapidly, and there will be the difference temperature between the interface water and bulk water. The liquideliquid direct contact heat transfer interfacial area varies continuously because of the continuing evaporation, system oscillation and sloshing of the unvaporized n-pentane liquid. On the other hand, most of direct contact heat transfer in stratified flow happened on a large interface where the flow velocities are not
Ti Tsat DT DTm u V
y r
855
average interface water temperature, C n-pentane saturation temperature, C actual driving temperature difference, C traditional driving temperature difference, C water flow velocity, m/s n-pentane vaporization volume, m3 n-pentane vaporization rate, kg/(m2 s) n-pentane liquid density, kg/m3
Abbreviations CMOS complementary metal oxide semiconductor IR infrared PE polyethylene
uniform and unstable especially at the turbulent state. Hence, the change of interface leads to the actual driving temperature difference and heat transfer coefficient cannot be obtained accurately. To solve this problem, a visualized experiment was fabricated to investigate the direct contact vaporization heat transfer between two immiscible liquids at turbulence flow state. In this work, a special circular interface with small unchanged area was designed to make the water flow velocities uniform on the entire interface. The actual driving temperature difference and heat transfer coefficient are obtained by using this n-pentane-water interface which also could be regarded as the liquideliquid heat transfer interface of drobbles. The influence factors of heat transfer coefficient are analyzed and the boiling phenomenon is observed when the contact interface constantly updates in the condition of turbulent flow. It is also anticipated that the results obtained in this study will provide a useful method to research the mechanism of immiscible liquids direct contact heat transfer, which will be used for better understanding the future design of gaseliquideliquid three phase exchangers. 2. Experimental apparatus and design of the test interface 2.1. Experimental apparatus and procedure The experimental apparatus was shown schematically in Fig. 1. The rectangular flow channel (internal dimensions of 10 mm in
Fig. 1. Schematic of experimental system for the direct contact heat transfer.
856
Y. Wang et al. / Energy 93 (2015) 854e863
width, 10 mm in height and 500 mm in length) was made by PE (Polyethylene) plate with a circular hole on the upper plate of the channel. The experiments were carried out in a tubule with an inner diameter of 8 mm and 30 mm in height, which was adhered on the circular hole of the rectangular channel. The lighter volatile npentane liquid added into the tubule was in direct contact with the immiscible hot water below the circular hole and the vaporization heat transfer of n-pentane taking place on the circular pentanewater interface. The pipeline of the system was stainless steel tube with an inner diameter of 20 mm and each two jointing sides were air proofed by air-proof glue in order to avoid entering of gases during the experiment. The entire pipeline was enveloped by thermal insulation layer to reduce heat loss except for the test section. The measurement principle is to calculate the heat flow rate by measuring the complete vaporization time of a certain volume npentane on the small circular interface whose area is not changed. Before each experiment, the water in the pipeline was heated to boiling for three hours in order to remove the gas in it and then was cooled down to the required temperature. Water passed through the shell and tube heat exchanger which must be used to maintain the uniform water temperature, rotor flowmeter (LZB-25, Jiangsu YuYing Automation Instrument Co., Ltd., China), PE rectangle flow channel and was circulated by a shielding pump (GDR32/6, WILO SE, Germany). During the process of experiment, the water flow velocity was controlled using the valve at the inlet of flowmeter. To reduce the fluctuation of water caused by the vibration of pump and intense turbulence of hot water, both sides of the pump were connected with two-stage hoses. A fine-tuning valve was installed to regulate the water layer height in the circular hole. It was necessary to remove the gas in the pipeline because water circulation was required to maintain the steady state. Water tank was used to provide make-up water and a valve in the outlet of it was closed during the experiment. When the fluctuation of liquideliquid interface reached a relatively steady state, with the aid of microsyringe (SPLab01, Baoding ShenChen precision pump Co., Ltd., China), n-pentane in the saturation temperature was injected into the tubule along its inner wall. At the same time, the evaporation time of n-pentane was recorded by a chronometer (DA-8SZ, Timing resolution: 1 ms, Xuchang ChangAn Technology Co., Ltd., China). In order to observe the phenomenon of evaporation, n-pentane was dyed by Oil Soluble Red. The bulk temperature of hot water was measured by a T-type thermocouple with the diameter of 0.3 mm positioned at the inlet of the rectangular flow channel. 2.2. Design of the pentane-water direct contact interface The design of pentane-water interface was important in this experiment because when the area of the interface is not changed and the water flow velocity on the entire interface is uniform everywhere, it is easy to accurately determine the heat transfer coefficient. As shown in Fig. 2b, the upper and lower openings of the circular hole in the upper plate of the rectangular flow channel were both designed in the shape of a circular truncated cone to reduce the influence of the velocity boundary layer. The schematic and dimensions of the heat transfer interface were shown in Fig. 2a and the angle between generatrix and basal diameter of the lower circular truncated cone was relatively small to weaken the influence of the rise of water level at left side of the circular hole on the interface at high flow velocity of water. The cec' was considered as the diameter of n-pentane-water direct contact circular interface.
Fig. 2. Schematic of direct contact interface.
0.25 mm) with spacing of about 5 mm [38]. The conductance probes were put into the tubule and the tips of probes were just in contact with the surface of water filled in the rectangle flow channel. The variable voltage signal in proportion to water layer height was obtained by using a conductometer connected with the parallel wire conductance probes. The analog voltage signal from the conductometer was sampled using a data logger (DI-710, Quatronix Company, USA) with sampling frequency of 1 kHz. Actual water layer height should be obtained by a linear relation between water layer height and voltage. A high-speed CMOS (complementary metal oxide semiconductor) digital camera (EOS70D, Canon Company, Japan) was used to capture the photographs of vaporization phenomenon on the interface. And the acquired images were color images with a resolution of 5472 3648 pixels. As shown in Fig. 1 the camera was positioned at a place of 5 cm away from bottom of the interface with illumination lamp as the light source at the same side. According to the image pixels, the coordinate y-up, y-down in the vertical direction, and coordinate x-left, x-right in horizontal direction of the departure bubbles were calculated. Thus, the diameter of bubble in horizontal and vertical directions was expressed as follows [39]:
Dx ¼ xright xleft
(1)
Dy ¼ yup ydown
(2)
Then, the average bubble diameter was expressed as follow:
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi D2x þ D2y
2.3. Measuring methods
Dd ¼
Variations of water layer height were measured by parallel wire conductance probes which consisted of two wires (diameters of
Liquid temperature profiles of liquideliquid heat transfer interface were analyzed by an non-contact measurement technique
2
(3)
Y. Wang et al. / Energy 93 (2015) 854e863
857
in order to avoid disturbing the fluid interface during the boiling process. Infrared thermography was chosen because of its good signal strength and a large temperature range. A material that was transparent in the visible range and infrared would allow using the optical temperature measurement technique: Infrared thermography to determine the internal water surface temperature [40]. One material that offered the desired characteristics was PE. The infrared thermography was carried out with a thermal infrared imager (SAT-90, SAT Company, China). It had a resolution of 640 480 pixels, a spectral range of 8e12 mm and was used in combination with a 50 mm lens. The average transmittance of IR (infrared) radiation in this wavelength ranges in PE plate and npentane was about 85% and 90%, respectively. The complete vaporization time of n-pentane was recorded by a chronometer. When the n-pentane started to contact with the water and then vaporized completely, some response time was needed to control the chronometer, which would cause some errors in the measurement of evaporation time. But this relative error is less than 0.24%, which is less than temperature and flow velocity measures errors and could be ignored in this experiment. Accuracy of measuring operational data is essential and errors should be reduced to ensure sufficient accuracy of the follow-up calculations [41]. Therefore, all the experiments are repeated 25 times in each group under the same conditions and Grubbs procedure was used to detect outliers [42]. Standard deviation maps were made to analyze the experiment and put into the section of results and discussions. 3. Results and discussions In this section, the experimental results for actual driving temperature difference, heat transfer coefficient and n-pentane vaporization rate are described and discussed in detail. The effects of water flow velocity and temperature difference on heat transfer performance are analyzed. 3.1. Actual driving temperature differences Fig. 3. Calibration curve of water temperature.
3.1.1. Temperature calibration Infrared thermography is a well-established technique for temperature and heat transfer measurements [43]. However, in practical applications, it is difficult to distinguish whether the infrared radiance is emitted from the test article or the contributions of surroundings. Furthermore, the infrared signal emitted from the test article depends on its emissivity, the transmissivity of the optical path and the sensitivity of the detector. Obviously, there is a difference between the actual water temperature and the measurement temperature. Because of that, an in-situ calibration is strongly recommended for accurate temperature measurement. Additionally, the target's emissivity and the reflective temperature had been known for the calibration and accurate results. The water surface temperature in rectangular flow channel is measured by the thermal infrared imager above the PE plate and the calibration curve is fitted to the water surface temperature measurement by a PT100 thermal resistance. Therefore, according to Fig. 3a, the actual water surface temperature Ts in the rectangular flow channel could be calibrated by the following equation:
Ts ¼ 1:0428 Tp
(4)
where Tp is the measured surface water temperature. The n-pentane with the temperature of 35 C was injected into the tubule and the water temperatures varied from 5.5 C to 35.3 C which are always lower than the n-pentane boiling point. The same method should be applied to calibrate the liquideliquid direct
contact interfacial water temperature which is measured by thermal infrared imager through the n-pentane liquid layer. As can be seen from Fig. 3b, the actual interface water temperature Ti could be calibrated by the following equation:
Ti ¼ 1:0246 Tw
(5)
where Tw is the measured interface water temperature. 3.1.2. Actual driving temperature differences When n-pentane is in direct contact with hot water, with the increases of water temperature, there would be n-pentane vapor bubbles on the interface of n-pentane and water. As the vaporization, n-pentane vapor bubbles would become bigger and leave away from interface finally, then the n-pentane liquid nearby is rapidly added to the liquideliquid interface at the same time. In the phase change process of n-pentane liquid, the strong perturbation on the n-pentane-water interface destroys the boundary layer and reduces the heat transfer resistance which greatly enhances the heat transfer rate. There are mainly two types of heat transfer including the convective heat transfer in severely turbulent hot water and n-pentane vaporization heat transfer on the direct contact liquideliquid interface. In order to determine the heat transfer coefficient accurately, the actual driving temperature difference must be measured. Due
858
Y. Wang et al. / Energy 93 (2015) 854e863
to the severe turbulence of hot water, the nucleation sites became very unstable, which caused uneven distribution of temperature on the boiling interface. The temperatures at different locations of the boiling interface were measured by a thermal infrared imager and the average temperature was taken to represent the interface water temperature. N-pentane at the saturation temperature is injected to the interface through a circular tubule. Based on the measured interface water temperature, the actual driving temperature difference DT can be defined by the following equation:
DT ¼ Ti Tsat
(6)
where Tsat is the n-pentane saturation temperature, Ti is the average interface water temperature. This actual driving temperature difference has a large difference from the traditional driving temperature difference defined by the calculating equation:
DTm ¼ T Tsat
(7)
where T is the bulk water temperature. As seen from Table 1, when the bulk water temperature T reaches 74 C, the traditional driving temperature difference DTm is 37.9 C. However, the average interface water temperature Ti is 45.02 C, and the actual driving temperature difference DT is 8.92 C. The actual driving temperature difference DT was lower than traditional driving temperature difference DTm. This is because the vaporization of n-pentane takes away heat from direct contact interface, which reduces the interface water temperature rapidly. The flowing hot water continuously supplies heat for the liquideliquid interface to maintain its temperature relatively stable. The actual temperature differences at different water temperatures and flow velocities are also shown in Table 1. As the bulk water temperature increases, the interface water temperature increases due to higher temperature gradient and convection heat transfer from the hot bulk water to the water of liquideliquid interface, which means the actual driving temperature difference DT increases with the increase of water temperature. But this increasing amplitude is far less than that of traditional driving temperature difference DTm. With the increase of water flow velocity, the interface water temperature do not have a large decrease. This may be because water flow velocity plays a more important role in making heat transfer boundary layer thinner than increasing water temperature gradient. 3.1.3. Water temperature profile Water surface temperatures profile along the center axial line of the circular heat transfer interface is measured at the top of PE plate using a thermal infrared imager. The typical water temperatures distribution at various axial locations are shown in Fig. 4. The zero point of X axis represents the center of circular interface and the upper wall of tubule has the lowest temperature. From Fig. 4, it is clear that a large temperature gradient exists on the left side of
interface close to the outlet of the rectangular flow channel. It shows that the interface water temperature is cooled down by the evaporation of n-pentane and then quickly heated to the bulk water temperature when it leaves way from the interface. The reasons can be related to the intense turbulence and mixing in hot water. As flow velocity increases from 0.56 m/s to 2.78 m/s, the temperature gradients increases from 2.306 С/mm to 3.222 С/mm at 74 С. With water increasing from 44 С to 74 С, the temperature gradient increases from 0.667 С/mm to 2.887 С/mm at 1.67 m/s. Because that thermal interaction between water and n-pentane is enhanced with the increase of flow velocity and bulk water temperature. The temperatures in the edge of circle interface are lower than that in its center because that solid contact angle can make it easier to form nucleation sites and vaporize n-pentane to reduce interface water temperatures. 3.2. Effect of water temperature and flow velocity on heat transfer coefficient When n-pentane temperature reaches saturation temperature it will lose the sensible heat and there will be no further change in npentane temperature [5]. According to the mass and time of npentane vaporization completely, the total heat transmission Q between the n-pentane and hot water on the direct contact interface can be found as follows:
Q¼
rVH t
(8)
where t is the time of n-pentane vaporization completely; V is the volume of n-pentane vaporization. The heat transfer coefficient on the n-pentane-water interface can be written as follows:
K¼
Q A$DT
(9)
where A is the area of n-pentane-water direct contact interface. Fig. 5 shows the comparison of the heat transfer coefficients obtained by two kinds of driving temperature differences which are the actual driving temperature difference DT and the traditional driving temperature difference DTm. It can be seen that there is a large deviation between the heat transfer coefficient obtained by two different kinds of driving temperature difference. When the bulk water temperature is 44 C, the heat transfer coefficient obtained by DT is 2.734 times as large as that by DTm. When the bulk water temperature is 74 C, the heat transfer coefficient obtained by DT is 3.258 times larger than that by DTm. The heat transfer coefficient increases gradually with the increases of water flow velocity can be seen in Fig. 6. At the low temperature difference of 2.86 C, with the increase of flow velocity from 0.56 m/s to 2.22 m/s, the heat transfer coefficient increases from 15.002 kW/(m2 C) to 33.859 kW/(m2 C). Such evolution can
Table 1 Actual temperature differences for different bulk water temperatures. Flow velocity (u, m/s)
N-pentane saturation temperature (Tsat, C)
Bulk water temperature (T, C)
Average interface water temperature (Ti, C)
Traditional temperature difference (DTm, C)
Actual temperature difference (DT, C)
1.67 1.67 1.67 1.67 0.56 2.22 2.78
36.1 36.1 36.1 36.1 36.1 36.1 36.1
44.00 54.00 64.00 74.00 74.00 74.00 74.00
38.96 41.08 42.82 45.02 45.11 44.84 44.81
7.9 17.9 27.9 37.9 37.9 37.9 37.9
2.86 4.98 6.72 8.92 9.01 8.74 8.71
Y. Wang et al. / Energy 93 (2015) 854e863
859
Fig. 4. Temperature profile of interface for different bulk water temperature and flow velocity.
be attributed to an increase frequency of interface update when the flowing hot water go through the interface faster. The violent agitation of interface reduces the heat transfer boundary lay and the heat transfer resistance, which increases the thermal energy transfer and enhances the heat transfer coefficient. In the process of npentane vapor bubbles rising up with the entraining of n-pentane liquid from the bottom layer to the top layer of n-pentane liquid, the increase of bubbles escape and rising velocities promotes the perturbation and enhances the heat transfer coefficient between the two immiscible liquid. When the flow velocity reaches 2.5 m/s, the value of heat transfer coefficient is approximately 35.748 kW/ (m2 C) and more likely to be stable. In addition, the heat transfer coefficient is 22.079 kW/(m2 C) when the water flow velocity reaches 1.94 m/s at the highest temperature difference of 8.92 C and the increase of heat transfer coefficient is not obvious. This may be explained mainly by the balance of heat transfer including the positive effect of the increasing frequency of interface update in promoting the heat supply with the negative effect of excessive interaction between vapor bubbles. On the other hand, the increase of temperature difference weakens the affection of flow velocity.
Heat transfer coefficients are determined for each of the actual driving temperature difference and plotted in Fig. 7, we can see that heat transfer coefficient increases with the decrease of temperature difference, in other words, there is much higher heat transfer efficiency at a lower water temperature. This trend is more obvious with the increase of water flow velocity. During the process of experiment, it has been observed that an increase of temperature is accompanied with the increase of bubbles departure diameter and this might decrease considerably the actual contact area between hot water and n-pentane, which leads to lower values of heat transfer coefficient. Particular emphasis is given to the horizontal coalescence of the departed bubbles close to the interface in this experiment, which is observed to cause the deterioration of the heat transfer coefficient [44], due to the formation and coalescence of large vapor bubbles that prevents the fresh bulk liquid from accessing the heat surface for evaporation heat transfer. The gas film with high thermal resistance also reduces the heat transfer efficiency sharply. This affection should be considered in this paper because of the existence of solid wall around the interface.
860
Y. Wang et al. / Energy 93 (2015) 854e863
Fig. 5. The heat transfer coefficient obtained by DT and DTm for: (a) T ¼ 44 С (b) T ¼ 54 С (c) T ¼ 64 С (d) T ¼ 74 С.
3.3. N-pentane vaporization rate
v¼
rV At
(10)
The influences of natural convection and diffusion on the npentane vaporization rate is far less than that of boiling heat transfer and could be ignored. The vaporization mass of n-pentane on 1 m2 direct contact interface per time unit is considered as the vaporization rate. The n-pentane vaporization rate n can be determined from the following expression:
It can be seen from Fig. 8 that the n-pentane vaporization rate increases gradually with an increase of water flow velocity and the actual driving temperature difference. As flow velocity increases, the turbulence is more serious and the convection between liquideliquid interface and bulk water is more intense, which
Fig. 6. Variations of n-pentane heat transfer coefficient with water flow velocity.
Fig. 7. Variations of n-pentane heat transfer coefficient with actual driving temperature difference.
Y. Wang et al. / Energy 93 (2015) 854e863
861
velocity higher than 1.39 m/s and 1.94 m/s at the temperature difference of 8.92 C and 2.86 C, respectively. Because the liquideliquid interface will be covered by a layer of gas film which isolates the n-pentane and water, which deteriorates the heat transfer. At a lower temperature difference, there is a larger value of heat transfer coefficient but a longer time is needed to complete vaporization of n-pentane liquid. In engineering design, we should find a balance between the heat transfer coefficient and the vaporization rate.
3.4. Visualization research of boiling on interface
Fig. 8. Variations of n-pentane vaporization rate with water flow rate for different temperature differences.
decreases the time of n-pentane vaporization completely. On the other hand, such evolution could be partially explained by the increase of the heat transfer coefficient at the liquideliquid interface with the increase of the water flow velocity. As the actual driving temperature difference increases, the heat flux increases and time of n-pentane vaporization completely decreases, which can show that temperature has a positive effect on vaporization rate. The increase of vaporization rate becomes not obvious when the flow
The boiling of flow interface in direct contact immiscible liquids is different from the boiling with static heat transfer surface. The npentane would appear to have an unusual type of nucleation because a nucleation site at the interface might become displaced by the agitation of boiling and perturbation of the intense turbulent of water. When the water in the rectangular flow channel is in the state of turbulent flow, a single bubble in the nucleate flow boiling process would go through five stages which are formation, growth, sliding, rising up and breakup. In the stages of bubble rising up and breakup, both vaporization of n-pentane and convection would occur in the n-pentane-water interface. The boiling heat transfer phenomena is observed at the bottom of the liquideliquid interface, however, it is difficult to record the process of n-pentane bubble formation and growth up because of the overlap and shelter from other vapor bubbles. Fig. 9 shows examples of the boiling pattern.
Fig. 9. Boiling pattern at u ¼ 0.56 m/s for: (a) T ¼ 44 С (b) T ¼ 54 С (c) T ¼ 64 С (d) T ¼ 74 С.
862
Y. Wang et al. / Energy 93 (2015) 854e863
flow velocity. With an increase of temperature, the number of nucleation sites decrease and the coalescence of large size vapor bubbles increases significantly, which has a negative effect on the heat transfer coefficient. The largest heat transfer coefficient of 35.329 kW/(m2 C) is obtained at the highest water flow velocity of 2.78 m/s and the lowest driving temperature difference of 2.86 C in this experiment. The n-pentane vaporization rate increases gradually with an increase of water flow velocity and driving temperature difference. The new experimental method and the results obtained in this study provide a foundation to research the mechanism of immiscible liquids direct contact vaporization heat transfer. Acknowledgments
Fig. 10. Variation of bubble diameter with bulk water temperature.
Growth and departure diameters of bubbles directly influence the boiling heat transfer efficiency [45]. In the process of boiling, the bubbles are not evenly distributed on the entire heat transfer interface. Due to the nonlinear interaction between the nucleation sites, the distribution of interfacial superheat is very uneven and fluctuating. With the increase of the degree of superheat and flow velocity, the nonlinear interaction between the nucleation sites would become stronger. Because of this nonlinear factors, the boiling system could produce chaos phenomenon and besides the nucleation site or bubble density and bubble diameters have obvious random characteristics. For this reason, the standard deviation of average bubble diameters departing from the heater interface under different water temperature are marked out in Fig. 10. It can be seen that when the flow velocity is 1.67 m/s and the bulk water temperature is 44 С, the pentane vapor bubbles depart from the interface with an average diameter of approximately 0.741 mm. The diameter increases as the water temperature increases, and the average diameter of n-pentane vapor bubble increases to 1.962 mm at a high water temperature of 74 C. The number of n-pentane vapor bubbles significantly decreases with the increase of water temperature, which shows the decrease of nucleation site. The relatively small amounts of nucleation sites and large sizes of vapor bubbles at higher water temperature decreases the heat transfer coefficient. 4. Conclusions A new experimental approach has been brought out to study the n-pentane-water direct contact vaporization heat transfer on a small circular interface whose area is not changed and on which flow velocities are uniform. The following conclusions are drawn after analyzing the experimental results: The actual driving temperature difference DT between interface water temperature and n-pentane saturation temperature is much smaller than the traditional temperature difference DTm, which causes that the actual heat transfer coefficient is larger than that obtained by the traditional algorithm. The actual driving temperature difference DT increases with bulk water temperature increasing, but the increasing amplitude is far less than that of traditional temperature difference DTm. In addition, the water flow velocity has no significant effect on the interface water temperature. Heat transfer coefficient increases with the increase of the water flow velocity, and this increase is no more pronounced at higher
Acknowledgments are given to Dr. Lianwei Qin, Lihong Jiang and Ganchao Xin in Green Energy Laboratory of Tianjin University for their contribution on language editing and results analysis to this paper. References [1] Xie J, Xu J, Xing F, Wang Z, Liu H. The phase separation concept condensation heat transfer in horizontal tubes for low-grade energy utilization. Energy 2014;69:787e800. [2] Li J, Cao W, Chen G. The heat transfer coefficient of new construction e Brick masonry with fly ash blocks. Energy 2015;86:240e6. [3] Ali AHH, Ahmed M, Abdel-Gaied SM. Investigation of heat transfer and fluid flow in transitional regime inside a channel with staggered plates heated by radiation for PV/T system. Energy 2013;59:255e64. [4] Lemouari M, Boumaza M, Kaabi A. Experimental analysis of heat and mass transfer phenomena in a direct contact evaporative cooling tower. Energy Convers Manag 2009;50(6):1610e7. [5] Kiatsiriroat T, Vithayasai S, Vorayos N, Nuntaphan A, Vorayos N. Heat transfer prediction for a direct contact ice thermal energy storage. Energy Convers Manag 2003;44(4):497e508. [6] Peng Z, Yiping W, Cuili G, Kun W. Heat transfer in gaseliquideliquid threephase direct-contact exchanger. Chem Eng J 2001;84(3):381e8. [7] Shimizu Y, Mori YH. Evaporation of single liquid drops in an immiscible liquid at elevated pressures: equipment and preliminary results. Exp Fluids 2004;6(2):73e9. [8] Siqueiros J, Bonilla O. An experimental study of a three-phase, direct-contact heat exchanger. Appl Therm Eng 1999;19(5):477e93. [9] Hamid MOA, Zhang B, Yang L. Application of field synergy principle for optimization fluid flow and convective heat transfer in a tube bundle of a preheater. Energy 2014;76:241e53. [10] Lund H, Werner S, Wiltshire R, Svendsen S, Thorsen JE, Hvelplund F, et al. 4th Generation District Heating (4GDH): integrating smart thermal grids into future sustainable energy systems. Energy 2014;68:1e11. [11] Shin S, Choi G, Kim BS, Cho HH. Flow boiling heat transfer on nanowire-coated surfaces with highly wetting liquid. Energy 2014;76:428e35. [12] Kukulka DJ, Smith R. Heat transfer evaluation of an enhanced heat transfer tube bundle. Energy 2014;75:97e103. [13] Mahood HB, Sharif AO, Al-Aibi S, Hawkins D, Thorpe R. Analytical solution and experimental measurements for temperature distribution prediction of threephase direct-contact condenser. Energy 2014;67:538e47. [14] Ribeiro Jr CP, Lage PLC. Direct-contact evaporation in the homogeneous and heterogeneous bubbling regimes. Part I: experimental analysis. Int J Heat Mass Transf 2004;47(17e18):3825e40. [15] Seetharamu KN, Battya P. Direct contact evaporation between two immiscible liquids in a spray column. J Heat Transf 1989;111(3):780e5. [16] Li Y-Y, Liu Z-H, Zheng B-C. Experimental study on the saturated pool boiling heat transfer on nano-scale modification surface. Int J Heat Mass Transf 2015;84:550e61. [17] Orzechowski T, Wcislik S. Experimental analysis of the drop film boiling at ambient pressure. Energy Convers Manag 2013;76:918e24. [18] Viskanta R, Behnia M. Experimental and analytical study of heat transfer and mixing in thermally stratified buoyant flows. Int J Heat Mass Transf 1982;25(6):847e61. [19] Mukerjee D, Davis EJ. Direct contact heat transfer between immiscible fluid layers in laminar flow. AIChE J 1972;18(1):94e101. [20] Gollan A, Sideman S. Direct contact heat transfer with change of phase: cocurrent flow of immiscible films with surface evaporation. Int J Heat Mass Transf 1968;11(12):1761e70. [21] Bentwich M, Landau U, Sideman S. Direct contact heat transfer with change of phase: evaporation of discrete volatile films from the surface of a stagnant immiscible liquid. Int J Heat Mass Transf 1970;13(6):945e56.
Y. Wang et al. / Energy 93 (2015) 854e863 [22] Nosoko T, Ohyama T, Mori Y. Evaporation of volatile-liquid lenses floating on an immiscible-liquid surface: effects of the surface age and fluid purities in npentane/water system. J Fluid Mech 1985;161:329e46. [23] Szekely J. Notes on the transfer at the interface of two independently stirred liquids. Int J Heat Mass Transf 1963;6(9):833e40. [24] Szekely J. Mathematical model for heat or mass transfer at the bubble-stirred interface of two immiscible liquids. Int J Heat Mass Transf 1963;6(5):417e22. [25] Fortuna G, Sideman S. Direct contact heat transfer between immiscible liquid layers with simultaneous boiling and stirring. Chem Eng Sci 1968;23(9): 1105e19. [26] Sideman S, Uliss I. Direct contact heat transfer between immiscible liquids in stratified flow with simultaneous interfacial boiling. Chem Eng Sci 1972;27(6):1245e56. [27] Iida Y, Takashima T. Direct-contact heat transfer characteristics: evaporation of a drop dropped onto a hot liquid surface. Int J Heat Mass Transf 1980;23(9): 1263e71. [28] Sideman S, Isenberg J. Direct contact heat transfer with change of phase: bubble growth in three-phase systems. Desalination 1967;2(2):207e14. [29] Mori YH. Configurations of gas-liquid two-phase bubbles in immiscible liquid media. Int J Multiph Flow 1978;4(4):383e96. [30] Mori YH. Classification of configurations of two-phase vapor/liquid bubbles in an immiscible liquid in relation to direct-contact evaporation and condensation processes. Int J Multiph Flow 1985;11(4):571e6. [31] Simpson HC, Beggs GC, Nazir M. Evaporation of butane drops in brine. Desalination 1974;15(1):11e23. [32] Sideman S, Hirsch G, Gat Y. Direct contact heat transfer with change of phase: effect of the initial drop size in three-phase heat exchangers. AIChE J 1965;11(6):1081e7. [33] Prakash CB, Pinder KL. Direct contact heat transfer between two immiscible liquids during vaporisation. Can J Chem Eng 1967;45(4):210e4. [34] Sideman S, Taitel Y. Direct-contact heat transfer with change of phase: evaporation of drops in an immiscible liquid medium. Int J Heat Mass Transf 1964;7(11):1273e89.
863
[35] Kulkarni AA, Ranade VV. Direct contact heat transfer via injecting volatile liquid in a hot liquid pool: generation and motion of bubbles. Chem Eng Sci 2013;100:421e32. [36] Tochitani Y, Mori YH, Komotori K. Vaporization of single liquid drops in an immiscible liquid part I: forms and motions of vaporizing drops. Warme Stoffubertragung 1977;10(1):51e9. [37] Tochitani Y, Nakagawa T, Mori YH, Komotori K. Vaporization of single liquid drops in an immiscible liquid part II: heat transfer characteristics. Warme Stoffubertragung 1977;10(2):71e9. [38] Al-Wahaibi T, Angeli P. Experimental study on interfacial waves in stratified horizontal oilewater flow. Int J Multiph Flow 2011;37(8):930e40. [39] Zhang Y, Wei J, Xue Y, Kong X, Zhao J. Bubble dynamics in nucleate pool boiling on micro-pin-finned surfaces in microgravity. Appl Therm Eng 2014;70(1):172e82. [40] Schreivogel P, Pfitzner M. Optical convective heat transfer measurements using infrared thermography and frequency domain phosphor thermometry. Int J Heat Mass Transf 2015;82:299e308. [41] Varbanov PS, Wang Y, Klemes JJ, Lund H. PRES'13 special section: process integration, modelling and optimisation for energy saving and pollution reduction. Energy 2014;75:1e4. [42] Jain RB. A recursive version of Grubbs' test for detecting multiple outliers in environmental and chemical data. Clin Biochem 2010;43(12):1030e3. [43] Kylili A, Fokaides PA, Christou P, Kalogirou SA. Infrared thermography (IRT) applications for building diagnostics: a review. Appl Energy 2014;134: 531e49. [44] Teodori E, Moita AS, Moreira ALN. Characterization of pool boiling mechanisms over micro-patterned surfaces using PIV. Int J Heat Mass Transf 2013;66:261e70. [45] Luke A, Cheng D-C. High speed video recording of bubble formation with pool boiling. Int J Therm Sci 2006;45(3):310e20.