- Email: [email protected]
Experimental study of heat transfer enhancement in electrohydrodynamic conduction pumping of liquid film using flush electrodes
Applied Thermal Engineering 50 (2013) 327e333
Contents lists available at SciVerse ScienceDirect
Applied Thermal Engineering journal homepage: www.elsevier.com/locate/apthermeng
Experimental study of heat transfer enhancement in electrohydrodynamic conduction pumping of liquid film using flush electrodes Nader Nourdanesh*, Esmaeil Esmaeilzadeh Heat and Fluid Flow Research Laboratory, Faculty of Mechanical Engineering, 29 Bahman Blvd, Tabriz, Iran
h i g h l i g h t s < Significant difference in flow rates has been observed by variation of temperature. < The effect of temperature increase on enhancement of efficiency is considerable. < The heat transfer coefficient in case2 increases due to vortices above electrodes. < The optimum film thickness for heat transfer coefficient enhancement is 6 mm < Heat transfer and power consumption ratio is decreased up to 10 kV intensively.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 13 December 2011 Accepted 24 August 2012 Available online 31 August 2012
Electrohydrodynamic conduction pumping of free surface dielectric liquid film, using flush electrodes, has been studied experimentally for various film temperatures. Volume flow rate, heat transfer and power consumption ratio and conduction pumping efficiency of free surface liquid film in different film thicknesses and temperatures have been investigated and then the best operating conditions have been presented. Also, the heat transfer coefficient on free surface liquid film passing on flush electrodes is compared with similar liquid film in absence of flush electrodes in different temperatures. Results show that as applied voltage increases, significant differences in volume flow rates have been observed by changing the temperature. Applied voltage related to the highest percentage of heat transfer coefficient enhancement demonstrates the reverse relation with temperature. Results confirm that there is a direct relationship between film thickness and the applied voltage related to the maximum heat transfer per pumps power consumption. Ó 2012 Elsevier Ltd. All rights reserved.
Keywords: EHD pumping Electrical conduction Heat transfer Liquid film Flush electrode
1. Introduction Interaction of electric field with dielectric fluid medium (EHD) can set up a mechanical body-force which can create a flow in the fluid and can be used in various applications such as liquid film pumping, mass transport, heat transfer control, electronic device cooling, and etc. When a dielectric fluid is exposed to an electric field, three electric body forces which induce the fluid to motion can be expressed as [1]:
1 1 v3 r fe ¼ rc Ee Ee2 V3 þ V Ee2 2 2 vr T
* Corresponding author. Tel.: þ98 411 3334025. E-mail address: [email protected] (N. Nourdanesh). 1359-4311/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.applthermaleng.2012.08.038
(1)
The symbols are defined in the nomenclature. The first term is the force exerting on positive and negative free charges, called Coulomb force (electrophoretic force), the second is dielectrophoretic force which mainly depends on electric permittivity gradient and the third one is electrostrictive force related only to compressible fluids. The forces exist together but it is necessary to note that for a single phase, isothermal medium, the only dominant mechanism to generate a permanent EHD motion is Coulomb force [2]. Moreover, in the present study, due to the low temperature gradient, electric permittivity gradient is negligible. In the basis of free charges generating process, EHD pumping mechanisms can be classified into three kinds: ion-drag pumping [3e7], induction pumping [8e12] and conduction pumping [13e 17]. Ion-drag pumping deals with the direct injection of charges into a dielectric fluid. This mechanism is not very desirable since it can degrade the working fluid [2]. Induction pumping basis is the generation of induced charges due to non-uniformity in electrical
328
N. Nourdanesh, E. Esmaeilzadeh / Applied Thermal Engineering 50 (2013) 327e333
0 Tav TN To,av
Nomenclature electric field strength [kV/m] power consumption [Watt] applied voltage [kV] current [mA] electric body force [kg m/s2] friction force [kg m/s2] channel width [mm] average velocity [mm/s] length of test channel [mm] volume flow rate [m3/s] gravitational acceleration [m/s2] specific energy of open channel flow [m] rate of heat transfer [Watt] average heat transfer coefficient [W/(m2 K)] surface area of the film contacted with air [mm2]
Ee Ep V I Fe Ff w uav L Qv g h Qh h A
Ti,av _ m Cp
Greek symbols rc charge density [1/cm] 3 electric permittivity [F/m] d film thickness [mm] r density [kg/m3] m dynamic viscosity [Pa s] h Pumping efficiency of a liquid film
conductivity of liquid. This non-uniformity can be caused by temperature gradient and/or inhomogeneity of the fluid (in the presence of interface between different liquids phases). Therefore this mechanism is not suitable for isothermal and homogeneous liquid. Conduction pumping is achieved by non-equilibrium behavior of dissociation of molecules within the liquid and recombination of the generated ions. The recombination is in dynamic equilibrium: kd ;kr
Aþ B ! Aþ þ B
average temperature of film in case1 and case2 [ C] environment temperature [ C] average output film temperatures of case1 and case2 [ C] average input film temperatures of case1 and case2 [ C] film mass velocity [kg/s] average specific heat capacity at constant pressure [J/(kg K)]
(2)
Where kd and kr are dissociation and recombination rate constants, respectively. This causes heterocharge layers to be formed in the vicinity of the electrodes under DC electric field. The attraction between each electrode and generated charges in heterocharge layer induces a fluid motion from the liquid side toward the electrode side. Pumping of liquid films has many industrial applications, including: heat pipes, heat exchangers, enhancing heat transfer in phase change processes, etc. EHD pumping is one of the best methods for liquid film pumping because of low electrical power consumption, ease in manufacturing and controlling technique. Moreover, it is vibration less, soundless and lightweight due to the absence of moving parts. Induction pumping of liquid films has been studied by few researchers [11,12,18e20]. Pumping of stratified liquid film with electrical conduction phenomenon was introduced by SeyedYagoobi et al. [21]. They used two types of electrode pairs: flush mounted and perforated. Their results indicated that flush electrodes are more suitable for thin liquid films, while perforated electrodes are better for thicker liquid films. In other work, SeyedYagoobi et al. have studied numerically conduction pumping of dielectric liquid film in the presence of evaporation [22]. Ahmad et al. have studied saturated pool boiling of R-123 including the critical heat flux (CHF). It was enhanced by modifying the surface characteristics and applied a high intensity electrostatic field [23]. They reported that EHD produced a further increase in the heat transfer rates particularly at low heat flux values and near the CHF. To the best of the author’s knowledge, to-date there is no experimental study about the heat transfer on electrohydrodynamic conduction pumping of dielectric liquid film. In present study, volume flow rate, heat transfer per power consumption and efficiency of free surface liquid film conduction pumping in different liquid temperature and film thicknesses have been investigated and used to determine the best operating
conditions. The heat transfer coefficient on free surface liquid film passing on flush electrodes is compared with similar liquid film without flush electrodes in different temperatures. 2. Experimental setup The schematic of the experimental setup is shown in Fig. 1. Loop consists of two direct channels, Case1 and Case2, one involves flush electrodes and the other measures the fluid volume flow rate inside the loop and compares the heat transfer coefficient with electrode side. These direct channels are connected by means of two semicircle channels. These semicircle channels are equipped with two flush electric heaters on the floor in order to control the film
Fig. 1. (a). Photograph of experimental setup. (b). Schematic diagram of experimental setup (dimensions in mm).
N. Nourdanesh, E. Esmaeilzadeh / Applied Thermal Engineering 50 (2013) 327e333
temperature. On the other word, there are 2 heater sections located on the bends that control the liquid temperatures entering the straight sections. Heat transfer occurs at each of the two straight sections. The structure of loop heaters that are mounted in the floor of the semicircle channels involves a cupper plate, silicon-paste to contact the heater with cupper plate properly, temperature sensor (PT100), fiberglass, rubber and finally bottom lid. Heaters temperature is constant. A digital temperature controller made by Autonics Company is used to control the heaters temperature in order to provide the required constant temperature for liquid film. Because all temperatures in an applied voltage are registered concurrently in 90 s time duration by time steps of 1 s and then the average of them has been calculated for each channel section, so this temperature controller system is accurate for current research. The obtained optimal arrangement for electrodes, presented by Esmaeilzadeh et al. [24] is considered for this study too. Also they reported that the best thickness for liquid film is 6 mm. The electrode side of the channel involves six pairs flush electrodes by 4 mm width of narrow electrodes and 16 mm width of the wide electrodes. The interval between two electrodes is 2 mm and pair electrodes are in 22 mm distance from each other, as shown in Fig. 2. Positive polarity Dc voltage is applied to high voltage electrodes. Because of flammability and other limitations in selection of fluid, kerosene dielectric fluid is used as an agent fluid in this research. With regarding to the fact that the open channel flow inside the loop is laminar, the surface velocity is a good criterion to show the flow behavior in the open channel flow. It is noteworthy that the parabolic velocity profile in the laminar open channel flow can be changed to the mean velocity by integrating the thickness of liquid film, then calculate the mean velocity and consequently volume flow rate is calculated as it is shown in the Fig. 5. Fig. 3 shows that the film temperature is measured in six points of each channel section by T-type thermocouples, thus their average is obtained for each film section. Thermocouples have been connected to a PC by especial electronic boards made by ADAMs Company. The part of the loop involving electrodes with length of 280 mm is called case2 and the opposite side by the same length without electrodes is called case1. Tav is average film temperature of total loop obtained by average of four temperatures, involving heaters input and output. Thermocouples are T-type (coppereconstantan) with the accuracy of 0.1 C. Measurement accuracy for channel wide and film thickness is 0.2 mm for applied voltage is 1 V and for electrical
329
Fig. 3. Arrangement of thermocouples in the channel section.
current rate is 1 nA. So, the maximum uncertainty of the heat transfer coefficient enhancement in 40, 44 and 47 C is respectively less than 17.9%, 11.3% and 9%. The maximum uncertainty of the volume flow rate in 4 mm, 6 mm and 8 mm film thicknesses is respectively less than 10%, 8.7% and 8.4%. Also, the maximum uncertainty of the efficiency in 4 mm, 6 mm and 8 mm film thicknesses is less than 11.2%, 9.3% and 8.7%, respectively. 3. Results and discussions 3.1. Pumps power consumption The results show that the current consumption versus applied voltages on the flush electrodes pumping the liquid film in the loop is linear in all three of film thicknesses and average temperatures. Moreover, the investigations show that the power consumption is the same in all of three thicknesses and average temperatures of kerosene liquid film. Power consumption is defined as:
Ep ¼ VI
(3)
Fig. 4 indicates power consumption versus applied voltages which shows the variation is quadratic. 3.2. Volume flow rate The variation of fluid volume flow rate pumped in the channel versus applied voltage for different temperatures and film thicknesses is plotted in Fig. 5, showing that in all three of the thicknesses volume flow rate is increased with increasing the film temperature.
power consumption (W)
0.5 0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 0
5
10
15
Applied Voltage (kV) Fig. 2. Strip of flush electrodes (dimensions in mm).
Fig. 4. Pumps power consumption versus applied voltages.
20
330
N. Nourdanesh, E. Esmaeilzadeh / Applied Thermal Engineering 50 (2013) 327e333
Fig. 5. Fluid volume flow rate pumped in the channel versus applied voltage for different temperatures and film thicknesses.
In this case study, the external force in stream wise, are the electric body force (Fe) and the friction force (Ff). The average velocity can be determined by [24]:
rw2 d2 uav þ 2mLðw þ 2dÞ2 uav Fe wd ¼ 0
(4)
Where r is kerosene density, w is the channel width, d is liquid film thickness, uav is average velocity of liquid film and m is kerosene dynamic viscosity. The volumetric flow rate is written as:
Qv ¼ dwuav
(5)
Since m decreases with increases in the temperature, according to the Eqs. (4) and (5) the film volume flow rate is increased by increasing the temperature. As shown in Fig. 5, by increase of applied voltage and consequently increase of electrical field strength, the fluid volume flow rate is increased and it leads to tangible friction force. Hence, as applied voltage increases, the temperature change causes significant differences in volume flow rates. On the other hand, by a decrease in the applied voltage and consequently decrease in the volume flow rate, the variation of temperature doesn’t have tangible effect on volume flow rates and there are almost the same volume flow rates in different temperatures. 3.3. Efficiency
(6)
Where, h is the specific energy of open channel Qh flow and has been defined as [25]:
h ¼ dþ
u2av 2g
3.4. Average heat transfer coefficient Heat transfer from liquid film can be done through the walls, channel floor and film surface which is in contact with air (environment). Since the walls are made of Polyglass and the area of the liquid films which is in contact with the walls is relatively low, heat transfer from the wall can be ignored. In addition, channel floor is made of Polyethylene with a thickness of 20 mm. The Polyethylene with this thickness can be considered as a thermal insulation compared with liquid film surface which is in contact with air. Therefore, the most important heat transfer factor is the surface which is in contact with air. From other point of view, case1 and case2 have the same condition at boundaries, so the film interface can be considered as the only heat transfer factor when the behavior of these two cases is compared. Average heat transfer coefficient of kerosene film is calculated by the following equation:
0
Qh ¼ hA Tav TN
Pumping efficiency of a liquid film can be defined as:
rgQv h pump output power h¼ ¼ power consumption VI
By considering Fig. 6 it can be found that EHD pump efficiency increases with increasing temperature but this enhancement by temperature in efficiency decreases with increasing applied voltage. The effect of temperature increase on efficiency enhancement is considerable in low voltages, while in voltages higher than 13 the effect of temperature can be considered negligible. Therefore, to increase efficiency by variation of temperature, the voltages less than 13 kV should be considered.
(7)
If h has been determined at any section and shows vertically above the channel surface, the locus of resulting points represents the specific energy diagram.
(8)
Where TN is the environment temperature, A is the surface area of 0 is the average temperature of film in the film contacted with air, Tav a part of the channel (case1, case2) which is investigated and calculated as the following: 0 Tav ¼
To;av þ Ti;av 2
(9)
Where Ti,av, To,av are the averages of input and output film temperatures in the study limit of the channel, respectively. Qh is the rate of heat transfer from liquid film to environment in the
N. Nourdanesh, E. Esmaeilzadeh / Applied Thermal Engineering 50 (2013) 327e333
331
Fig. 6. EHD pumping efficiency of kerosene fluid versus applied voltage in average loop temperatures of 40, 44 and 47 C for different film thicknesses.
study limit of the channel (case1, case2) that is calculated by the following equation:
_ p Ti;av To;av Qh ¼ mC
(10)
_ is fluid mass velocity inside of the channel, Cp is kerosene m average specific heat capacity at constant pressure. As shown in Fig. 7 in case1, the heat transfer coefficient is increased by increasing of applied voltage, but it is nearly constant in voltages higher than 9 kV. Also in case2, the positive slope reduction trend is increased from voltages higher than 9 kV but not as intensive as case1. Based on Fig. 5, the rate of volume flow increases up to 9 kV intensively. However, for voltages higher than 9 kV, the increase in volume flow rate decreases intensively which finally becomes
almost constant. So in case1 the heat transfer coefficient in these voltages is nearly constant according to the behavior of the volume flow rate. In case2, in spite of the existence of this behavior in volume flow rate trend, increasing the electrical field strength leads to enhancement of vortices strength above the electrode [26]. Hence, in voltages higher than 9 kV, the heat transfer coefficient is not constant but its enhancement intensity decrease. Also, the heat transfer coefficient is increased in both cases by increasing temperature. This is because of the viscosity decrease and fluid velocity enhancement by increasing the temperature. 3.5. Average heat transfer coefficient enhancement Average heat transfer coefficient enhancement is calculated from the following equation:
Fig. 7. Average heat transfer coefficient of case1 and case2 in three different temperatures versus applied voltage for different film thicknesses.
332
N. Nourdanesh, E. Esmaeilzadeh / Applied Thermal Engineering 50 (2013) 327e333
Table 1 Maximum percentage of average heat transfer coefficient enhancement in case2 by comparison with case1 and related applied voltage in three different temperatures and film thicknesses. Film thickness
4 mm 6 mm 8 mm
40 C
44 C
47 C
Max enhancement
Related voltage (kV)
Max enhancement
Related voltage (kV)
Max enhancement
Related voltage (kV)
286.4% 432.1% 350.1%
19 18 20
277.5% 332.9% 278%
18 16 20
210.3% 349% 164.3%
15 11 20
Fig. 8. Heat transfer and pumps power consumption ratio versus applied voltage for three different film thicknesses and temperatures.
average heat transfer coefficient enhancement 2 3 h 2 ¼ 4 15 100 h1
(11)
Table 1 depicts the maximum percentage of average heat transfer coefficient enhancement in case2 compared with case1 and related applied voltage in three different temperatures and film thicknesses. In 40, 44 and 47 C film average temperatures, the percentage of enhancement rises up respectively to 286.4%, 277.5% and 210.3% for 4 mm film thickness, 432.1%, 332.9% and 349% for 6 mm film thickness and 350.1%, 278% and 164.3% for 8 mm film thickness. As table shows, it can be found that the optimum thickness of the film for heat transfer coefficient enhancement is 6 mm. Applied voltage of the highest percentage of heat transfer coefficient enhancement shows the reverse relation with temperature and it is reduced with increasing the temperature. In average film temperatures of 40, 44 and 47 C, the related voltages of highest percentage of heat transfer coefficient enhancement are respectively 19, 18 and 15 kV for 4 mm film thickness, 18, 16 and 11 kV for 6 mm film thickness and 20 kV for all three of the temperatures in 8 mm film thickness. 3.6. Heat transfer and pumps power consumption ratio Heat transfer and pumps power consumption ratio is defined as (Qh/VI) for case1 and case2. Fig. 8 shows heat transfer per power consumption (Qh/VI) versus applied voltage for three different film thicknesses and temperatures. The results show that there is a direct relationship between applied voltage of the peak and film thickness. It is obvious that by an increase in film thickness, the peak is occurred in a higher
voltage. After peak, increasing of the applied voltage causes the ratio of the heat transfer and power consumption (Qh/VI) decrease up to 10 kV sharply. Finally, in voltages higher than 16 kV it is nearly constant. 4. Conclusions Conduction pumping of various film temperature with different film thickness, using flush electrodes, has been studied experimentally. Results show that: (1) Electrodes power consumption is the same in different temperatures. As applied voltage increases, significant difference in volume flow rates has been observed by change in temperature. On the other hand, by decrease of the applied voltage and consequently decrease of volume flow rate the variation of temperature doesn’t have tangible effect on volume flow rate. (2) The effect of increasing temperature on the enhancement of efficiency is considerable in low voltages. While in voltages higher than 13 kV, the effect of temperature can be considered negligible. Therefore, voltages less than 13 kV should be considered in order to increase efficiency by variation of temperature. (3) In voltages higher than 9 kV, the heat transfer coefficient is almost constant for case1 but in case2, it increases due to the vortices above the electrodes. Also, the heat transfer coefficient is increased in both cases by increasing temperature. (4) Applied voltage related to the highest percentage of heat transfer coefficient enhancement demonstrates the reverse relation with temperature. In 40, 44 and 47 C film average temperatures, the percentage of enhancement rise up
N. Nourdanesh, E. Esmaeilzadeh / Applied Thermal Engineering 50 (2013) 327e333
respectively to 286.4%, 277.5% and 210.3% for 4 mm film thickness, 432.1%, 332.9% and 349% for 6 mm film thickness and 350.1%, 278% and 164.3% for 8 mm film thickness. Also the results show that, the optimum film thickness for heat transfer coefficient enhancement is 6 mm. (5) Results confirm that there is a direct relationship between the film thickness and the applied voltage related to maximum ratio of heat transfer and power consumption (Qh/VI).
Acknowledgements This study was supported by the Mechanical Engineering Faculty Heat and Fluid Flow Research Laboratory of Tabriz University.
References [1] J.R. Melcher, Continuum Electromechanics, MIT Press, Cambridge, 1981. [2] S.J. Jeong, J. Seyed-Yagoobi, Theoretical/numerical study of electrohydrodynamic pumping through conduction phenomenon, IEEE Trans. Ind. Appl. 39 (2003) 355e361. [3] O.M. Stuetzer, Ion drag pressure generation, J. Appl. Phys. 30 (1959) 984e994. [4] W.F. Pickard, Ion drag pumping, I. Theory 34 (1963) 246e250. [5] P. Atten, Haidara, Electrical conduction and EHD motion of dielectric liquids in knife-plain electrode assembly, IEEE Trans. Electr. Insul. 20 (1985) 187e198. [6] J.M. Crowley, G.S. Wright, J.C. Chato, Selecting a working fluid to increase the efficiency and flow rate of an EHD pump, IEEE Trans. Ind. Appl. 26 (1990) 42e49. [7] J.E. Bryan, J. Seyed-Yagoobi, Analysis of 2-dimensional flow field generated by a 1 electrode-pair ion-drag pump, IEEE Trans. Dielectr. Electr. Insul 3 (1994) 459e466. [8] J.R. Melcher, Traveling-wave induced electroconvection, Phys. Fluids 9 (1966) 1548e1555. [9] A.P. Washabaugh, M. Zhan, J.R. Melcher, Electrohydrodynamic traveling-wave pumping of homogeneous semi-insulating liquids, IEEE Trans. Ind. Appl. 24 (1989) 807e834. [10] J. Seyed-Yagoobi, J.C. Chato, J.M. Crowley, P.T. Krein, Induction electrohydrodynamic pump in a vertical configuration: part I theory, ASME J. Heat Trans. 111 (1989) 664e669.
333
[11] M. Wawzyniak, J. Seyed-Yagoobi, Electrohydrodynamic induction pumping of a stratified liquid/vapor medium in the presence of volumetric and interface charges, IEEE Trans. Ind. Appl. 37 (2001) 950e958. [12] K. Brand, J. Seyed-Yagoobi, Effect of electrode configuration on electrohydrodynamic induction pumping of a stratified liquid/vapor medium, IEEE Trans. Ind. Appl. 38 (2002) 389e400. [13] S.J. Jeong, J. Seyed-Yagoobi, Experimental study of electrohydrodynamic pumping throughconductionphenomenon, J. Electrostatics 56 (2002) 123e 133. [14] R. Hanaoka, S. Takata, M. Murakumo, H. Anzi, Properties of liquid jet induced by electrohydrodynamic pumping in dielectric liquids, Elect. Eng. Jpn. 138 (2002) 224e230. [15] P. Atten, J. Seyed-Yagoobi, Electrohydrodynamic induced dielectric liquid flow through pure conduction in point/plane geometry, IEEE Trans. Dielectr. Electr. Insul 10 (2003) 27e36. [16] S.J. Jeong, J. Seyed-Yagoobi, Innovative electrode design for electrohydrodynamic conduction pumping, IEEE Trans. Ind. Appl. 40 (2004) 900e 904. [17] R. Hanaoka, H. Nakamichi, S. Takata, T. Fukami, Distinctive flow properties of liquid jet generated by EHD pump and conical nozzle, Elect. Eng. Jpn. 154 (2006) 399e406. [18] M. Wawzyniak, J. Seyed-Yagoobi, An analytical study of electrohydrodynamic induction pumping of a stratified liquid/vapor medium, IEEE Trans. Ind. Appl. 35 (1999) 231e239. [19] K. Brand, J. Seyed-Yagoobi, Experimental study of electrohydrodynamic induction pumping of a dielectric micro liquid film in external horizontal condensation process, ASME J. Heat Trans. 125 (2003) 1096e1104. [20] S. Aldini, J. Seyed-Yagoobi, Stability of electrohydrodynamic induction pumping of liquid in vertical annular configuration, IEEE Trans. Ind. Appl. 41 (2005) 1522e1530. [21] M. Siddiqui, J. Seyed-Yagoobi, Experimental study of pumping of liquid film with electric conduction phenomenon, IEEE Trans. Ind. Appl. 45 (2009) 3e9. [22] S.W. Ahmad, T.G. Karayiannis, D.B.R. Kenning, A. Luke, Compound effect of EHD and surface roughness in pool boiling and CHF with R-123, Appl. Thermal Eng. 31 (2011) 1994e2003. [23] M. Yazdani, J. Seyed-Yagoobi, Numerical investigation of electrohydrodynamic-conduction pumping of liquid film in the presence of evaporation, ASME J. Heat Trans. 131 (2009) 011602-1e011602-8. [24] R. Gharraei, E. Esmaeilzadeh, M. Hemayatkhah, J. Danaeefar, Experimental investigation of electrohydrodynamic conduction pumping of various liquids film using flush electrodes, J. Electrostatics 69 (2011) 43e53. [25] R. . Street, G. Watters, J. Vennard, Elementary Fluid Mechanics, seventh ed., J. Wiley & Sons, 1996. [26] M. Hemayatkhah, R. Gharraei, E. Esmaeilzadeh, Flow pattern visualization of liquid film conduction pumping using flush mounted electrodes, Exp. Thermal Fluid Sci. 35 (2011) 933e938.
Contents lists available at SciVerse ScienceDirect
Applied Thermal Engineering journal homepage: www.elsevier.com/locate/apthermeng
Experimental study of heat transfer enhancement in electrohydrodynamic conduction pumping of liquid film using flush electrodes Nader Nourdanesh*, Esmaeil Esmaeilzadeh Heat and Fluid Flow Research Laboratory, Faculty of Mechanical Engineering, 29 Bahman Blvd, Tabriz, Iran
h i g h l i g h t s < Significant difference in flow rates has been observed by variation of temperature. < The effect of temperature increase on enhancement of efficiency is considerable. < The heat transfer coefficient in case2 increases due to vortices above electrodes. < The optimum film thickness for heat transfer coefficient enhancement is 6 mm < Heat transfer and power consumption ratio is decreased up to 10 kV intensively.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 13 December 2011 Accepted 24 August 2012 Available online 31 August 2012
Electrohydrodynamic conduction pumping of free surface dielectric liquid film, using flush electrodes, has been studied experimentally for various film temperatures. Volume flow rate, heat transfer and power consumption ratio and conduction pumping efficiency of free surface liquid film in different film thicknesses and temperatures have been investigated and then the best operating conditions have been presented. Also, the heat transfer coefficient on free surface liquid film passing on flush electrodes is compared with similar liquid film in absence of flush electrodes in different temperatures. Results show that as applied voltage increases, significant differences in volume flow rates have been observed by changing the temperature. Applied voltage related to the highest percentage of heat transfer coefficient enhancement demonstrates the reverse relation with temperature. Results confirm that there is a direct relationship between film thickness and the applied voltage related to the maximum heat transfer per pumps power consumption. Ó 2012 Elsevier Ltd. All rights reserved.
Keywords: EHD pumping Electrical conduction Heat transfer Liquid film Flush electrode
1. Introduction Interaction of electric field with dielectric fluid medium (EHD) can set up a mechanical body-force which can create a flow in the fluid and can be used in various applications such as liquid film pumping, mass transport, heat transfer control, electronic device cooling, and etc. When a dielectric fluid is exposed to an electric field, three electric body forces which induce the fluid to motion can be expressed as [1]:
1 1 v3 r fe ¼ rc Ee Ee2 V3 þ V Ee2 2 2 vr T
* Corresponding author. Tel.: þ98 411 3334025. E-mail address: [email protected] (N. Nourdanesh). 1359-4311/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.applthermaleng.2012.08.038
(1)
The symbols are defined in the nomenclature. The first term is the force exerting on positive and negative free charges, called Coulomb force (electrophoretic force), the second is dielectrophoretic force which mainly depends on electric permittivity gradient and the third one is electrostrictive force related only to compressible fluids. The forces exist together but it is necessary to note that for a single phase, isothermal medium, the only dominant mechanism to generate a permanent EHD motion is Coulomb force [2]. Moreover, in the present study, due to the low temperature gradient, electric permittivity gradient is negligible. In the basis of free charges generating process, EHD pumping mechanisms can be classified into three kinds: ion-drag pumping [3e7], induction pumping [8e12] and conduction pumping [13e 17]. Ion-drag pumping deals with the direct injection of charges into a dielectric fluid. This mechanism is not very desirable since it can degrade the working fluid [2]. Induction pumping basis is the generation of induced charges due to non-uniformity in electrical
328
N. Nourdanesh, E. Esmaeilzadeh / Applied Thermal Engineering 50 (2013) 327e333
0 Tav TN To,av
Nomenclature electric field strength [kV/m] power consumption [Watt] applied voltage [kV] current [mA] electric body force [kg m/s2] friction force [kg m/s2] channel width [mm] average velocity [mm/s] length of test channel [mm] volume flow rate [m3/s] gravitational acceleration [m/s2] specific energy of open channel flow [m] rate of heat transfer [Watt] average heat transfer coefficient [W/(m2 K)] surface area of the film contacted with air [mm2]
Ee Ep V I Fe Ff w uav L Qv g h Qh h A
Ti,av _ m Cp
Greek symbols rc charge density [1/cm] 3 electric permittivity [F/m] d film thickness [mm] r density [kg/m3] m dynamic viscosity [Pa s] h Pumping efficiency of a liquid film
conductivity of liquid. This non-uniformity can be caused by temperature gradient and/or inhomogeneity of the fluid (in the presence of interface between different liquids phases). Therefore this mechanism is not suitable for isothermal and homogeneous liquid. Conduction pumping is achieved by non-equilibrium behavior of dissociation of molecules within the liquid and recombination of the generated ions. The recombination is in dynamic equilibrium: kd ;kr
Aþ B ! Aþ þ B
average temperature of film in case1 and case2 [ C] environment temperature [ C] average output film temperatures of case1 and case2 [ C] average input film temperatures of case1 and case2 [ C] film mass velocity [kg/s] average specific heat capacity at constant pressure [J/(kg K)]
(2)
Where kd and kr are dissociation and recombination rate constants, respectively. This causes heterocharge layers to be formed in the vicinity of the electrodes under DC electric field. The attraction between each electrode and generated charges in heterocharge layer induces a fluid motion from the liquid side toward the electrode side. Pumping of liquid films has many industrial applications, including: heat pipes, heat exchangers, enhancing heat transfer in phase change processes, etc. EHD pumping is one of the best methods for liquid film pumping because of low electrical power consumption, ease in manufacturing and controlling technique. Moreover, it is vibration less, soundless and lightweight due to the absence of moving parts. Induction pumping of liquid films has been studied by few researchers [11,12,18e20]. Pumping of stratified liquid film with electrical conduction phenomenon was introduced by SeyedYagoobi et al. [21]. They used two types of electrode pairs: flush mounted and perforated. Their results indicated that flush electrodes are more suitable for thin liquid films, while perforated electrodes are better for thicker liquid films. In other work, SeyedYagoobi et al. have studied numerically conduction pumping of dielectric liquid film in the presence of evaporation [22]. Ahmad et al. have studied saturated pool boiling of R-123 including the critical heat flux (CHF). It was enhanced by modifying the surface characteristics and applied a high intensity electrostatic field [23]. They reported that EHD produced a further increase in the heat transfer rates particularly at low heat flux values and near the CHF. To the best of the author’s knowledge, to-date there is no experimental study about the heat transfer on electrohydrodynamic conduction pumping of dielectric liquid film. In present study, volume flow rate, heat transfer per power consumption and efficiency of free surface liquid film conduction pumping in different liquid temperature and film thicknesses have been investigated and used to determine the best operating
conditions. The heat transfer coefficient on free surface liquid film passing on flush electrodes is compared with similar liquid film without flush electrodes in different temperatures. 2. Experimental setup The schematic of the experimental setup is shown in Fig. 1. Loop consists of two direct channels, Case1 and Case2, one involves flush electrodes and the other measures the fluid volume flow rate inside the loop and compares the heat transfer coefficient with electrode side. These direct channels are connected by means of two semicircle channels. These semicircle channels are equipped with two flush electric heaters on the floor in order to control the film
Fig. 1. (a). Photograph of experimental setup. (b). Schematic diagram of experimental setup (dimensions in mm).
N. Nourdanesh, E. Esmaeilzadeh / Applied Thermal Engineering 50 (2013) 327e333
temperature. On the other word, there are 2 heater sections located on the bends that control the liquid temperatures entering the straight sections. Heat transfer occurs at each of the two straight sections. The structure of loop heaters that are mounted in the floor of the semicircle channels involves a cupper plate, silicon-paste to contact the heater with cupper plate properly, temperature sensor (PT100), fiberglass, rubber and finally bottom lid. Heaters temperature is constant. A digital temperature controller made by Autonics Company is used to control the heaters temperature in order to provide the required constant temperature for liquid film. Because all temperatures in an applied voltage are registered concurrently in 90 s time duration by time steps of 1 s and then the average of them has been calculated for each channel section, so this temperature controller system is accurate for current research. The obtained optimal arrangement for electrodes, presented by Esmaeilzadeh et al. [24] is considered for this study too. Also they reported that the best thickness for liquid film is 6 mm. The electrode side of the channel involves six pairs flush electrodes by 4 mm width of narrow electrodes and 16 mm width of the wide electrodes. The interval between two electrodes is 2 mm and pair electrodes are in 22 mm distance from each other, as shown in Fig. 2. Positive polarity Dc voltage is applied to high voltage electrodes. Because of flammability and other limitations in selection of fluid, kerosene dielectric fluid is used as an agent fluid in this research. With regarding to the fact that the open channel flow inside the loop is laminar, the surface velocity is a good criterion to show the flow behavior in the open channel flow. It is noteworthy that the parabolic velocity profile in the laminar open channel flow can be changed to the mean velocity by integrating the thickness of liquid film, then calculate the mean velocity and consequently volume flow rate is calculated as it is shown in the Fig. 5. Fig. 3 shows that the film temperature is measured in six points of each channel section by T-type thermocouples, thus their average is obtained for each film section. Thermocouples have been connected to a PC by especial electronic boards made by ADAMs Company. The part of the loop involving electrodes with length of 280 mm is called case2 and the opposite side by the same length without electrodes is called case1. Tav is average film temperature of total loop obtained by average of four temperatures, involving heaters input and output. Thermocouples are T-type (coppereconstantan) with the accuracy of 0.1 C. Measurement accuracy for channel wide and film thickness is 0.2 mm for applied voltage is 1 V and for electrical
329
Fig. 3. Arrangement of thermocouples in the channel section.
current rate is 1 nA. So, the maximum uncertainty of the heat transfer coefficient enhancement in 40, 44 and 47 C is respectively less than 17.9%, 11.3% and 9%. The maximum uncertainty of the volume flow rate in 4 mm, 6 mm and 8 mm film thicknesses is respectively less than 10%, 8.7% and 8.4%. Also, the maximum uncertainty of the efficiency in 4 mm, 6 mm and 8 mm film thicknesses is less than 11.2%, 9.3% and 8.7%, respectively. 3. Results and discussions 3.1. Pumps power consumption The results show that the current consumption versus applied voltages on the flush electrodes pumping the liquid film in the loop is linear in all three of film thicknesses and average temperatures. Moreover, the investigations show that the power consumption is the same in all of three thicknesses and average temperatures of kerosene liquid film. Power consumption is defined as:
Ep ¼ VI
(3)
Fig. 4 indicates power consumption versus applied voltages which shows the variation is quadratic. 3.2. Volume flow rate The variation of fluid volume flow rate pumped in the channel versus applied voltage for different temperatures and film thicknesses is plotted in Fig. 5, showing that in all three of the thicknesses volume flow rate is increased with increasing the film temperature.
power consumption (W)
0.5 0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 0
5
10
15
Applied Voltage (kV) Fig. 2. Strip of flush electrodes (dimensions in mm).
Fig. 4. Pumps power consumption versus applied voltages.
20
330
N. Nourdanesh, E. Esmaeilzadeh / Applied Thermal Engineering 50 (2013) 327e333
Fig. 5. Fluid volume flow rate pumped in the channel versus applied voltage for different temperatures and film thicknesses.
In this case study, the external force in stream wise, are the electric body force (Fe) and the friction force (Ff). The average velocity can be determined by [24]:
rw2 d2 uav þ 2mLðw þ 2dÞ2 uav Fe wd ¼ 0
(4)
Where r is kerosene density, w is the channel width, d is liquid film thickness, uav is average velocity of liquid film and m is kerosene dynamic viscosity. The volumetric flow rate is written as:
Qv ¼ dwuav
(5)
Since m decreases with increases in the temperature, according to the Eqs. (4) and (5) the film volume flow rate is increased by increasing the temperature. As shown in Fig. 5, by increase of applied voltage and consequently increase of electrical field strength, the fluid volume flow rate is increased and it leads to tangible friction force. Hence, as applied voltage increases, the temperature change causes significant differences in volume flow rates. On the other hand, by a decrease in the applied voltage and consequently decrease in the volume flow rate, the variation of temperature doesn’t have tangible effect on volume flow rates and there are almost the same volume flow rates in different temperatures. 3.3. Efficiency
(6)
Where, h is the specific energy of open channel Qh flow and has been defined as [25]:
h ¼ dþ
u2av 2g
3.4. Average heat transfer coefficient Heat transfer from liquid film can be done through the walls, channel floor and film surface which is in contact with air (environment). Since the walls are made of Polyglass and the area of the liquid films which is in contact with the walls is relatively low, heat transfer from the wall can be ignored. In addition, channel floor is made of Polyethylene with a thickness of 20 mm. The Polyethylene with this thickness can be considered as a thermal insulation compared with liquid film surface which is in contact with air. Therefore, the most important heat transfer factor is the surface which is in contact with air. From other point of view, case1 and case2 have the same condition at boundaries, so the film interface can be considered as the only heat transfer factor when the behavior of these two cases is compared. Average heat transfer coefficient of kerosene film is calculated by the following equation:
0
Qh ¼ hA Tav TN
Pumping efficiency of a liquid film can be defined as:
rgQv h pump output power h¼ ¼ power consumption VI
By considering Fig. 6 it can be found that EHD pump efficiency increases with increasing temperature but this enhancement by temperature in efficiency decreases with increasing applied voltage. The effect of temperature increase on efficiency enhancement is considerable in low voltages, while in voltages higher than 13 the effect of temperature can be considered negligible. Therefore, to increase efficiency by variation of temperature, the voltages less than 13 kV should be considered.
(7)
If h has been determined at any section and shows vertically above the channel surface, the locus of resulting points represents the specific energy diagram.
(8)
Where TN is the environment temperature, A is the surface area of 0 is the average temperature of film in the film contacted with air, Tav a part of the channel (case1, case2) which is investigated and calculated as the following: 0 Tav ¼
To;av þ Ti;av 2
(9)
Where Ti,av, To,av are the averages of input and output film temperatures in the study limit of the channel, respectively. Qh is the rate of heat transfer from liquid film to environment in the
N. Nourdanesh, E. Esmaeilzadeh / Applied Thermal Engineering 50 (2013) 327e333
331
Fig. 6. EHD pumping efficiency of kerosene fluid versus applied voltage in average loop temperatures of 40, 44 and 47 C for different film thicknesses.
study limit of the channel (case1, case2) that is calculated by the following equation:
_ p Ti;av To;av Qh ¼ mC
(10)
_ is fluid mass velocity inside of the channel, Cp is kerosene m average specific heat capacity at constant pressure. As shown in Fig. 7 in case1, the heat transfer coefficient is increased by increasing of applied voltage, but it is nearly constant in voltages higher than 9 kV. Also in case2, the positive slope reduction trend is increased from voltages higher than 9 kV but not as intensive as case1. Based on Fig. 5, the rate of volume flow increases up to 9 kV intensively. However, for voltages higher than 9 kV, the increase in volume flow rate decreases intensively which finally becomes
almost constant. So in case1 the heat transfer coefficient in these voltages is nearly constant according to the behavior of the volume flow rate. In case2, in spite of the existence of this behavior in volume flow rate trend, increasing the electrical field strength leads to enhancement of vortices strength above the electrode [26]. Hence, in voltages higher than 9 kV, the heat transfer coefficient is not constant but its enhancement intensity decrease. Also, the heat transfer coefficient is increased in both cases by increasing temperature. This is because of the viscosity decrease and fluid velocity enhancement by increasing the temperature. 3.5. Average heat transfer coefficient enhancement Average heat transfer coefficient enhancement is calculated from the following equation:
Fig. 7. Average heat transfer coefficient of case1 and case2 in three different temperatures versus applied voltage for different film thicknesses.
332
N. Nourdanesh, E. Esmaeilzadeh / Applied Thermal Engineering 50 (2013) 327e333
Table 1 Maximum percentage of average heat transfer coefficient enhancement in case2 by comparison with case1 and related applied voltage in three different temperatures and film thicknesses. Film thickness
4 mm 6 mm 8 mm
40 C
44 C
47 C
Max enhancement
Related voltage (kV)
Max enhancement
Related voltage (kV)
Max enhancement
Related voltage (kV)
286.4% 432.1% 350.1%
19 18 20
277.5% 332.9% 278%
18 16 20
210.3% 349% 164.3%
15 11 20
Fig. 8. Heat transfer and pumps power consumption ratio versus applied voltage for three different film thicknesses and temperatures.
average heat transfer coefficient enhancement 2 3 h 2 ¼ 4 15 100 h1
(11)
Table 1 depicts the maximum percentage of average heat transfer coefficient enhancement in case2 compared with case1 and related applied voltage in three different temperatures and film thicknesses. In 40, 44 and 47 C film average temperatures, the percentage of enhancement rises up respectively to 286.4%, 277.5% and 210.3% for 4 mm film thickness, 432.1%, 332.9% and 349% for 6 mm film thickness and 350.1%, 278% and 164.3% for 8 mm film thickness. As table shows, it can be found that the optimum thickness of the film for heat transfer coefficient enhancement is 6 mm. Applied voltage of the highest percentage of heat transfer coefficient enhancement shows the reverse relation with temperature and it is reduced with increasing the temperature. In average film temperatures of 40, 44 and 47 C, the related voltages of highest percentage of heat transfer coefficient enhancement are respectively 19, 18 and 15 kV for 4 mm film thickness, 18, 16 and 11 kV for 6 mm film thickness and 20 kV for all three of the temperatures in 8 mm film thickness. 3.6. Heat transfer and pumps power consumption ratio Heat transfer and pumps power consumption ratio is defined as (Qh/VI) for case1 and case2. Fig. 8 shows heat transfer per power consumption (Qh/VI) versus applied voltage for three different film thicknesses and temperatures. The results show that there is a direct relationship between applied voltage of the peak and film thickness. It is obvious that by an increase in film thickness, the peak is occurred in a higher
voltage. After peak, increasing of the applied voltage causes the ratio of the heat transfer and power consumption (Qh/VI) decrease up to 10 kV sharply. Finally, in voltages higher than 16 kV it is nearly constant. 4. Conclusions Conduction pumping of various film temperature with different film thickness, using flush electrodes, has been studied experimentally. Results show that: (1) Electrodes power consumption is the same in different temperatures. As applied voltage increases, significant difference in volume flow rates has been observed by change in temperature. On the other hand, by decrease of the applied voltage and consequently decrease of volume flow rate the variation of temperature doesn’t have tangible effect on volume flow rate. (2) The effect of increasing temperature on the enhancement of efficiency is considerable in low voltages. While in voltages higher than 13 kV, the effect of temperature can be considered negligible. Therefore, voltages less than 13 kV should be considered in order to increase efficiency by variation of temperature. (3) In voltages higher than 9 kV, the heat transfer coefficient is almost constant for case1 but in case2, it increases due to the vortices above the electrodes. Also, the heat transfer coefficient is increased in both cases by increasing temperature. (4) Applied voltage related to the highest percentage of heat transfer coefficient enhancement demonstrates the reverse relation with temperature. In 40, 44 and 47 C film average temperatures, the percentage of enhancement rise up
N. Nourdanesh, E. Esmaeilzadeh / Applied Thermal Engineering 50 (2013) 327e333
respectively to 286.4%, 277.5% and 210.3% for 4 mm film thickness, 432.1%, 332.9% and 349% for 6 mm film thickness and 350.1%, 278% and 164.3% for 8 mm film thickness. Also the results show that, the optimum film thickness for heat transfer coefficient enhancement is 6 mm. (5) Results confirm that there is a direct relationship between the film thickness and the applied voltage related to maximum ratio of heat transfer and power consumption (Qh/VI).
Acknowledgements This study was supported by the Mechanical Engineering Faculty Heat and Fluid Flow Research Laboratory of Tabriz University.
References [1] J.R. Melcher, Continuum Electromechanics, MIT Press, Cambridge, 1981. [2] S.J. Jeong, J. Seyed-Yagoobi, Theoretical/numerical study of electrohydrodynamic pumping through conduction phenomenon, IEEE Trans. Ind. Appl. 39 (2003) 355e361. [3] O.M. Stuetzer, Ion drag pressure generation, J. Appl. Phys. 30 (1959) 984e994. [4] W.F. Pickard, Ion drag pumping, I. Theory 34 (1963) 246e250. [5] P. Atten, Haidara, Electrical conduction and EHD motion of dielectric liquids in knife-plain electrode assembly, IEEE Trans. Electr. Insul. 20 (1985) 187e198. [6] J.M. Crowley, G.S. Wright, J.C. Chato, Selecting a working fluid to increase the efficiency and flow rate of an EHD pump, IEEE Trans. Ind. Appl. 26 (1990) 42e49. [7] J.E. Bryan, J. Seyed-Yagoobi, Analysis of 2-dimensional flow field generated by a 1 electrode-pair ion-drag pump, IEEE Trans. Dielectr. Electr. Insul 3 (1994) 459e466. [8] J.R. Melcher, Traveling-wave induced electroconvection, Phys. Fluids 9 (1966) 1548e1555. [9] A.P. Washabaugh, M. Zhan, J.R. Melcher, Electrohydrodynamic traveling-wave pumping of homogeneous semi-insulating liquids, IEEE Trans. Ind. Appl. 24 (1989) 807e834. [10] J. Seyed-Yagoobi, J.C. Chato, J.M. Crowley, P.T. Krein, Induction electrohydrodynamic pump in a vertical configuration: part I theory, ASME J. Heat Trans. 111 (1989) 664e669.
333
[11] M. Wawzyniak, J. Seyed-Yagoobi, Electrohydrodynamic induction pumping of a stratified liquid/vapor medium in the presence of volumetric and interface charges, IEEE Trans. Ind. Appl. 37 (2001) 950e958. [12] K. Brand, J. Seyed-Yagoobi, Effect of electrode configuration on electrohydrodynamic induction pumping of a stratified liquid/vapor medium, IEEE Trans. Ind. Appl. 38 (2002) 389e400. [13] S.J. Jeong, J. Seyed-Yagoobi, Experimental study of electrohydrodynamic pumping throughconductionphenomenon, J. Electrostatics 56 (2002) 123e 133. [14] R. Hanaoka, S. Takata, M. Murakumo, H. Anzi, Properties of liquid jet induced by electrohydrodynamic pumping in dielectric liquids, Elect. Eng. Jpn. 138 (2002) 224e230. [15] P. Atten, J. Seyed-Yagoobi, Electrohydrodynamic induced dielectric liquid flow through pure conduction in point/plane geometry, IEEE Trans. Dielectr. Electr. Insul 10 (2003) 27e36. [16] S.J. Jeong, J. Seyed-Yagoobi, Innovative electrode design for electrohydrodynamic conduction pumping, IEEE Trans. Ind. Appl. 40 (2004) 900e 904. [17] R. Hanaoka, H. Nakamichi, S. Takata, T. Fukami, Distinctive flow properties of liquid jet generated by EHD pump and conical nozzle, Elect. Eng. Jpn. 154 (2006) 399e406. [18] M. Wawzyniak, J. Seyed-Yagoobi, An analytical study of electrohydrodynamic induction pumping of a stratified liquid/vapor medium, IEEE Trans. Ind. Appl. 35 (1999) 231e239. [19] K. Brand, J. Seyed-Yagoobi, Experimental study of electrohydrodynamic induction pumping of a dielectric micro liquid film in external horizontal condensation process, ASME J. Heat Trans. 125 (2003) 1096e1104. [20] S. Aldini, J. Seyed-Yagoobi, Stability of electrohydrodynamic induction pumping of liquid in vertical annular configuration, IEEE Trans. Ind. Appl. 41 (2005) 1522e1530. [21] M. Siddiqui, J. Seyed-Yagoobi, Experimental study of pumping of liquid film with electric conduction phenomenon, IEEE Trans. Ind. Appl. 45 (2009) 3e9. [22] S.W. Ahmad, T.G. Karayiannis, D.B.R. Kenning, A. Luke, Compound effect of EHD and surface roughness in pool boiling and CHF with R-123, Appl. Thermal Eng. 31 (2011) 1994e2003. [23] M. Yazdani, J. Seyed-Yagoobi, Numerical investigation of electrohydrodynamic-conduction pumping of liquid film in the presence of evaporation, ASME J. Heat Trans. 131 (2009) 011602-1e011602-8. [24] R. Gharraei, E. Esmaeilzadeh, M. Hemayatkhah, J. Danaeefar, Experimental investigation of electrohydrodynamic conduction pumping of various liquids film using flush electrodes, J. Electrostatics 69 (2011) 43e53. [25] R. . Street, G. Watters, J. Vennard, Elementary Fluid Mechanics, seventh ed., J. Wiley & Sons, 1996. [26] M. Hemayatkhah, R. Gharraei, E. Esmaeilzadeh, Flow pattern visualization of liquid film conduction pumping using flush mounted electrodes, Exp. Thermal Fluid Sci. 35 (2011) 933e938.