boiling2009-itp-09-09 nucleate boiling heat transfer

Upload: agnotts09

Post on 04-Jun-2018

217 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/13/2019 Boiling2009-Itp-09-09 Nucleate Boiling Heat Transfer

    1/6

    Interdisciplinary Transport Phenomena VI, Volterra, Italy, 2009 1 Paper No: ITP-09-09

    Proceedings of ITP2009Interdisciplinary Transport Phenomena VI:

    Fluid, Thermal, Biological, Materials and Space SciencesOctober 4-9, 2009, Volterra, Italy

    ITP-09-09

    NUCLEATE BOILING HEAT TRANSFERIN ALCOHOL AQUEOUS SOLUTIONS

    Takashi SakaiDepartment of Aeronautics and Astronautics,

    Kyushu University, Japan(JSPS Research Fellow DC)

    Yuuki TsukinariDepartment of Aeronautics and Astronautics,

    Kyushu University, Japan

    Shinsaku YoshiiDepartment of Aeronautics and Astronautics,

    Kyushu University, Japan

    Kazutoshi KajimotoDepartment of Aeronautics and Astronautics,

    Kyushu University, Japan

    Yasuhisa ShinmotoDepartment of Aeronautics and Astronautics,

    Kyushu University, Japan

    Haruhiko OhtaDepartment of Aeronautics and Astronautics,

    Kyushu University, Japan

    ABSTRACTSaturated pool boiling experiments were conducted by using

    alcohol aqueous solutions, i.e. 1-Propanol/Water, 2-

    Propanol/Water and Water/Ethylene glycol mixtures at0.1MPa. The heating surface is a horizontal upward-facing

    circular flat plate of 40mm in diameter. The increase in the

    heat transfer coefficients was observed for 1-Propanol/Water

    and 2-Propaonl/Water mixtures at very low alcohol

    concentration range, where Marangoni convection towards the

    three-phase interline is induced by the surface tension gradient

    along the vapor-liquid interface due to preferential evaporation

    of alcohol component. The heat transfer enhancement due to

    Marangoni effect seems to overcome the heat transfer

    deterioration due to mass transfer resistance. In the moderate

    concentration range, however, the heat transfer enhancement

    was turned to the heat transfer deterioration with increasing in

    the alcohol concentration. For Water/Ethylene glycol mixture,on the contrary, no significant increase and decrease in the

    heat transfer coefficients was observed in the entire

    concentration range tested. On the other hand, the critical heat

    flux was markedly decreased in the low alcohol concentration

    range of 1-Propanol/Water and 2-Propanol/Water mixtures.

    Marangoni effect is also suggested as a possible reason for the

    peculiar trend of the critical heat flux. Marangoni effect seems

    to promote the extension of drypatches underneath coalesced

    bubbles and the critical heat flux becomes decreased.

    NOMENCLATURED mass diffusion coefficient

    g gravity accelerationLa Laplace constant

    Ma Marangoni number

    P pressure

    q heat flux

    T temperaturewx weight fraction of liquid

    wy weight fraction of vapor

    x mole fraction of liquid

    y mole fraction of vapor

    Greek symbols

    heat transfer coefficient

    viscosity density surface tension

    Subscripts

    1 low-boiling component or more volatile component

    2 high-boiling component or less volatile component

    CHF critical heat flux

    g gas

    l liquid

    m mixture

    sat saturation

  • 8/13/2019 Boiling2009-Itp-09-09 Nucleate Boiling Heat Transfer

    2/6

    Interdisciplinary Transport Phenomena VI, Volterra, Italy, 2009 2 Paper No: ITP-09-09

    1. INTRODUCTIONNucleate boiling heat transfer is one of the most effective

    means for the heat removal from thermal devices because of

    its high performance of heat transfer and heat transportation.For this reason, boiling heat transfer and critical heat flux

    (CHF) of several single components and mixtures have been

    studied by many investigators.

    In nucleate boiling of binary mixtures, mass transfer

    resistance causes the variation of liquid concentration and itssaturation temperature underneath bubbles along a vapor-

    liquid phase equilibrium diagram. Thus Marangoni force is

    induced along the vapor-liquid interface by the surface tension

    gradient due to the variation of concentration and temperature.

    The Marangoni force induces the flow whose direction is

    determined by the sign of surface tension gradient along the

    interface.When the surface tension of mixture increases with

    decreasing the concentration of more volatile component

    (positive mixtures), Marangoni force is expected to induce the

    flow which supplies liquid towards the three-phase interline

    and prevents the extension of drypatch and the heat transfer

    enhancement is expected. On the other hand, when the surface

    tension of mixture decreases with decreasing the concentration

    of more volatile component (negative mixtures), Marangoni

    force is expected to induce the flow towards opposite direction

    and promote the heat transfer deterioration. For alcohol

    aqueous solutions, the value of surface tension changes

    significantly in the low alcohol concentration range and the

    existence of strong Marangoni force is expected.

    There exist a number of experimental and theoretical

    researches for the boiling heat transfer of binary mixtures. It is

    a well known fact from numerous boiling experiments that thenucleate boiling heat transfer coefficients of mixtures areusually lower than those of ideal or, more precisely,

    hypothetical single components with the same thermophysical

    properties as the mixtures. Van Wijk et al.[1] explained for the

    lower heat transfer coefficient of binary mixtures. They noted

    that the reduction of local liquid concentration of more volatile

    component increases the local saturation temperature

    underneath bubbles because more volatile component

    preferentially evaporates to maintain equilibrium between two

    phases. Consequently, the wall temperature rises and the heat

    transfer coefficient based on the saturation temperature of bulk

    liquid is lowered. Stephan and Krner[2] proposed a simple

    correlation based on the explanation given by Van Wijk etal.[1]. They defined the ideal heat transfer coefficients for

    mixtures which are given by the interpolation of wall

    superheats for both pure single components to represent the

    mixture boiling heat transfer coefficients in the absence of

    mixture effects. And they evaluated the heat transfer

    deterioration rate from the ideal heat transfer coefficients by

    using the difference between the molar fractions of both

    phases on the vapor-liquid equilibrium diagram. Calus et

    al.[3] considered the reduction of the bubble growth rate to

    evaluate the deterioration rate by using Scriven[4] and Van

    Stralen[5] analysis of the bubble growth. Thome[6] tried to

    calculate the deterioration rate by only using phase

    equilibrium diagram. He proposed a simple correlation which

    uses only temperature difference between dew and bubbling

    point temperatures at the bulk concentration. Kandlikar[7]

    theoretically analyzed the mixture property effects on nucleate

    boiling heat transfer. He introduced a new pseudo-singlecomponent heat transfer coefficient as the ideal heat transfer

    coefficient to reflect mixture property effects more accurately.

    And the mixture boiling heat transfer coefficient was derived

    theoretically by estimating the effects of heat and mass

    transfer at the vapor-liquid interface of a glowing bubble. Inthe above researches, only heat transfer deterioration of binary

    mixtures was discussed and Marangoni effect on the heat

    transfer coefficients of binary mixtures was ignored.

    On the other hand, there also exist a lot of studies for critical

    heat flux of mixtures. Several geometries of heating surface

    were tested and the result of either increasing or decreasing in

    the critical heat flux was observed. Hovestreijdt[8] first

    speculated that the Marangoni force affects the critical heat

    flux of binary mixtures. McGillis and Carey[9] suggested a

    correlation based on the speculation given by Hovestreijdt.

    They quantitatively estimated the Marangoni effect as an

    additional liquid restoring force caused by the surface tension

    gradient with a modified model derived from hydrodynamics.

    Fujita et al.[10] also considered the Marangoni effect to

    evaluate the increase in the critical heat flux of binary

    mixtures and proposed a correlation using the Marangoni

    number defined by using thermal diffusivity. McEligot[11]

    attributed the increase in the critical heat flux of binary

    mixtures to the increase in effective subcooling because the

    interfacial temperature is increased due to the preferential

    evaporation of more volatile component. Leddy and

    Lienhard[12] quantitatively evaluated the increase in theeffective subcooling and suggested a correlation. The aboveresearches, however, considered no geometry effect on the

    critical heat flux of mixtures.

    3. EXPERIMENTAL APPARATUS AND PROCEDUREThe experimental setup shown in Fig.1 is composed of three

    main parts, a boiling vessel, condensers and a heating section.

    The boiling vessel is made from a stainless steel pipe with

    inner diameter of 200mm and a volume of 0.023m3. Liquid

    and vapor temperature are measured by four thermocouples

    placed in the vessel. The condensers are installed in the upper

    side of the boiling vessel and control the levels of system

    pressure and saturation temperature. The detailed structure ofheating section is also illustrated in Fig.1. A heating surface,

    which is the upper edge of the copper heating block, is heated

    by the cartridge heaters inserted in the bottom of the heating

    block. The horizontal upward-facing heating surface polished

    by sandpaper (No. 600) has a dimension of 40mm in diameter

    and is surrounded by a circular thin fin in order to preventpreferential bubble generation at the edge and minimize the

    circumferential heating loss from the periphery. In the heating

    block, eight thermocouples are inserted at the depth of 1, 7,

    13, 19mm from the heating surface along two axes at the

    center and 14mm from the center. A surface heat flux and

    surface temperature are evaluated from the thermal conduction

  • 8/13/2019 Boiling2009-Itp-09-09 Nucleate Boiling Heat Transfer

    3/6

    Interdisciplinary Transport Phenomena VI, Volterra, Italy, 2009 3 Paper No: ITP-09-09

    across the copper heating block by the indicated temperatures

    of thermocouples.Pool boiling experiments were conducted under saturated

    conditions. After the enough time is elapsed to confirm the

    steady state of indicated temperatures of all thermocouples

    and pressure, heat flux is increased. The condition of critical

    heat flux is detected by the excursion of temperature in the

    copper heating block. The heat flux one step before the value

    resulting temperature excursion is defined as a critical heat

    flux. The increment of input heat flux near the critical heat

    flux is 1.0105 W/m2 and the value gives the resolution ofmeasured critical heat flux.

    The experiments were performed at 0.1MPa on ground.

    Three alcohol aqueous solutions, 1-Propanol/Water, 2-

    Propanol/Water and Water/Ethylene glycol mixtures were

    employed as test fluids. Concentration ranges tested are shown

    in Table1 and the phase equilibrium diagrams of these

    mixtures are shown in Fig.2. In Fig.3, surface tension behavior

    of these mixtures is shown. 1-Propanol/Water and 2-

    Propanol/Water mixtures have an azeotropic concentration, so

    they act as the positive mixture at the lower concentration of

    1-Propanol and 2-Propanol than the azeotrope while they act

    as the negative mixture at the higher concentration of alcohol.

    For Water/Ethylene glycol mixture, it acts as the negative

    mixture in the entire concentration range.

    In the present research, Marangoni effect was evaluated by

    the Marangoni number using mass diffusion coefficient

    defined by Eqs. (1) and (2).

    Fig.1 Schematic diagram of experimental apparatus

    Fig.2 Phase equilibrium diagrams at 0.1MPa

    0 0.2 0.4 0.6 0.8 170

    80

    90

    100

    110

    120

    Weight fractionwx1

    Tsat

    C

    Bubbling Line Dew Line

    2-Propanol / Water

    Positive system

    Negative system

    Azeotrope

    (b) 2-Propanol/Water

    0 0.2 0.4 0.6 0.8 170

    80

    90

    100

    110

    120

    Weight fractionwx1

    Tsat

    C

    Bubbling Line

    Dew Line

    1-Propanol / Water

    Positive system

    Negative system

    Azeotrope

    (a) 1-Propanol/Water

    0 0.2 0.4 0.6 0.8 1

    100

    200

    Weight fractionwx1

    Tsat

    C

    Bubbling Line Dew Line

    Water / Ethylene glycol

    Negative system

    (c) Water/Ethylene glycol

    Table1 Test concentration range

    Test fluids Test bulk liquid

    concentrations of alcohol

    1-Propanol/Water 0 ~ 90wt%

    2-Propanol/Water 0 ~ 89wt%

    Water/Ethylene glycol 0 ~ 90wt%

  • 8/13/2019 Boiling2009-Itp-09-09 Nucleate Boiling Heat Transfer

    4/6

    Interdisciplinary Transport Phenomena VI, Volterra, Italy, 2009 4 Paper No: ITP-09-09

    Fig.5 Heat transfer coefficients

    (c) Water/Ethylene glycol

    0 0.2 0.4 0.6 0.8 10

    1

    2

    3[104]

    Ideal heat transfer coefficientsStephan and Krner correlationThome correlation

    2-Propanol/Water

    Weight fraction of alcoholwx

    W/m2K

    Azeotrope

    q=1.0105W/m

    2

    q=2.0105W/m

    2

    q=4.0105W/m

    2

    (b) 2-Propanol/Water

    (a) 1-Propanol/Water

    0 0.2 0.4 0.6 0.8 10

    1

    2

    3[104]

    Ideal heat transfer coefficientsStephan and Krner correlationThome correlation

    1-Propanol/Water

    Weight fraction of alcoholwx

    W/m2K

    Azeotrope

    q=1.0105W/m

    2

    q=2.0105W/m

    2

    q=4.0105W/m

    2

    0 0.2 0.4 0.6 0.8 10

    1

    2

    3[104]

    Ideal heat transfer coefficientsStephan and Krner correlation

    Thome correlation

    Water/Ethylene glycol

    Weight fraction of alcoholwx

    W/m2K

    q=1.0105W/m

    2

    q=2.0105W/m

    2

    q=4.0105W/m

    2

    Fig.3 Surface tension under phase equilibriumcondition at 0.1MPa

    0 0.2 0.4 0.6 0.8 10

    0.02

    0.04

    0.06

    0.08

    Weight fraction of alcohol

    N/m

    1-Propanol/Water2-Propanol/Water

    Water/Ethylene glycol

    Fig.4 Marangoni number

    0 0.2 0.4 0.6 0.8 1-1

    0

    1

    2

    3[108]

    1-Propanol/Water2-Propanol/WaterWater/Ethylene glycol

    Weight fraction of alcohol

    Ma

    ( )

    D

    Laxyx

    Mal

    mm

    11

    1

    = (1)

    )( ,, mgml

    mm

    gLa

    (2)

    The calculated Marangoni number for the entire

    concentration range is shown in Fig.4. The larger heat transfer

    enhancement is expected for the larger Marangoni number.

    4.EXPERIMENTAL RESULTS AND DISCUSSIONSFigure 5 shows the effect of concentration on the heat

    transfer coefficients at selected heat fluxes (1.0105, 2.0105,4.0105 W/m2). For 1-Propanol/Water and 2-Propanol/Watermixtures, the increase in the heat transfer coefficients was

  • 8/13/2019 Boiling2009-Itp-09-09 Nucleate Boiling Heat Transfer

    5/6

    Interdisciplinary Transport Phenomena VI, Volterra, Italy, 2009 5 Paper No: ITP-09-09

    (a) 1-Propanol/Water

    Fig.6 Critical heat flux

    (c) Water/Ethylene glycol

    (b) 2-Propanol/Water

    0 0.2 0.4 0.6 0.8 1105

    106

    107

    Water/Ethylene glycol

    Weight fraction of alcoholwx

    qCHFW/m

    2

    Experimental resultsZuber correlation

    0 0.2 0.4 0.6 0.8 1105

    106

    10

    Experimental resultsZuber correlation

    1-Propanol/Water

    Weight fraction of alcoholwx

    qCHFW/m2

    0 0.2 0.4 0.6 0.8 1105

    106

    107

    2-Propanol/Water

    Weight fraction of alcoholwx

    qCHFW/m2

    Experimental resultsZuber correlation

    observed at the very low concentration of alcohol. In the

    moderate concentration range, however, the heat transfer

    enhancement was turned to the heat transfer deterioration with

    increasing in the alcohol concentration. And at the azeotropicconcentration, the heat transfer coefficient has a local

    maximum. For Water/Ethylene glycol mixture, on the other

    hand, no marked heat transfer enhancement or deterioration

    was observed and the heat transfer coefficients gradually

    decreased with increasing in the alcohol concentration. In thefigure, predicted heat transfer coefficients by the existing

    correlations are also shown. Stephan and Krner[2] correlation

    and Thome[6] correlation are used, and the heat transfer

    coefficients of each pure alcohol are predicted by the

    Kutateladze[13] correlation because the experiments for the

    pure alcohol were not conducted due to the safety problem

    inherent in the experimental apparatus. Here, multiplying

    factors are introduced to adjust the predicted values to

    coincide with the experimental data for pure water and at the

    azeotropic concentration. The multiplying factors in the

    intermediate concentration range between zero (pure water)

    and the azeotropic concentration and between the azeotropic

    concentration and unity (pure alcohol) was evaluated by the

    linear interpolation of these multipliers, where a multiplier of

    unity was introduced for pure alcohols. As can be seen from

    Fig.5, both Stephan and Krner[2] correlation and Thome[6]

    correlation can reproduce the trends of measured heat transfer

    coefficients in the moderate concentration range for 1-

    Propanol/Water and 2-Propanol/Water mixtures. For

    Water/Ethylene glycol mixture, however, the prediction of the

    heat transfer coefficients by Thome correlation is

    overestimated because its temperature difference between thedew and bubbling temperature at the bulk liquid concentrationis too large. And the heat transfer enhancement at the very low

    concentration range can be reproduced by neither of these

    correlations. The heat transfer enhancement seems to be

    caused by the Marangoni effect, and the heat transfer

    enhancement overcomes the heat transfer deterioration due to

    mass transfer resistance because the Marangoni number has a

    peak value in the low concentration range and is markedly

    reduced with further increasing in the alcohol concentration

    as shown in Fig.4.

    Figure 6 shows the effect of concentration on the critical heat

    flux. The critical heat flux predicted by the Zuber

    correlation[14] for hypothetical single component liquids withthe same properties as mixtures is also shown in Fig.6. In the

    moderate concentration range, the measured critical heat flux

    value was gradually decreased with increasing in the alcohol

    concentration. Zuber correlation can reproduce the critical

    heat flux trend. In the low alcohol concentration range,

    however, the critical heat flux was significantly decreased.

    The peculiar trend of critical heat flux was also reported by

    using a flat plate heating surface[15] while the markedly

    increase in the entire concentration range was observed by

    using a wire heating surface[10,12]. Thus the trend of the

    critical heat flux depends on the geometry of heating surfaces,

    but the logical explanation or correlations which consider the

    geometry effect has not be provided. The low concentration

  • 8/13/2019 Boiling2009-Itp-09-09 Nucleate Boiling Heat Transfer

    6/6

    Interdisciplinary Transport Phenomena VI, Volterra, Italy, 2009 6 Paper No: ITP-09-09

    range, where the peculiar decrease in the critical heat flux was

    observed, corresponds to the peak Marangoni number.

    Consequently, Marangoni effect is suggested again as a

    possible cause for the peculiar trend of the critical heat flux.How Marangoni effect influences to the critical heat flux

    mechanism is deduced as follows. Marangoni force is directed

    towards the three-phase interline underneath primary bubbles

    because of the preferential evaporation of alcohol for positive

    mixtures. The Marangoni effect supplies liquid spontaneouslyto the microlayer of the primary bubbles. As a result,

    evaporation of microlayer is enhanced and heat transfer

    coefficient is increased. On the other hand, the enhanced

    evaporation causes finally the shortage of liquid stored in

    macrolayer and drypatches are extended from some locations

    where primary bubbles were existed. The extension of

    drypaches is continued under a coalesced bubble, and the

    excursion of surface temperature occurs before entire part of

    liquid in macrolayer is consumed. Thus the critical heat flux

    condition seems to be realized independent of the period for

    the detachment of a coalesced bubble and it varies with the

    intensity of Marangoni force induced around primary bubbles.

    This mechanism might be true for the flat plate heating surface

    where liquid supply to primary bubbles is indirectly via the

    formation of macrolayer during the detachment of coalesced

    bubbles and not directly by bulk liquid.

    5. CONCLUTIONSThe boiling heat transfer and the critical heat flux of alcohol

    aqueous solutions were investigated through the saturated pool

    boiling experiments using 1-Propanol/Water, 2-

    Propanol/Water and Water/Ethylene glycol mixtures at0.1MPa. The following conclusions were derived.

    (1) The increase in the heat transfer coefficients was observed

    at the very low alcohol concentration of 1-Propanol/Water

    and 2-Propanol/Water mixtures while the critical heat flux

    was markedly deceased in the same alcohol concentration

    range.

    (2) The above trends of the heat transfer and the critical heat

    flux are deduced to be caused by the Marangoni effect,

    which accelerates the evaporation of the primary bubbles

    in macrolayer underneath coalesced bubbles.

    ACKNOWLEDGMENTS

    This work was supported by Grant-in-Aid for JSPS Fellows(211862) from Japan Society for the Promotion of Science

    (JSPS). The authors express the appreciation for the support.

    REFERENCES[1] W.R. Van Wijk, A.S. Vos, & S.J.D. Van Stralen, Heat

    transfer to boiling binary liquid mixtures. Chem. Engng. Sci.,

    5: 68-80 (1956).

    [2] K. Stephan & M. Krner, Calculation of heat transfer in

    evaporating binary liquid mixtures. Chem. Ing. Tech., 41:

    409-417 (1969).

    [3] W.F. Calus & D.J. Leonidopoulos, Pool boiling - Binary

    liquid mixtures.Int. J. Heat Mass Transfer, 17: 249-256

    (1974).

    [4] L.E. Scriven, On the dynamics of phase growth. Chem.

    Eng. Sci., 10: 1-13 (1959).

    [5] S.J.D. Van Stralen, Bubble growth rates in boiling binary

    mixtures.Br. Chem. Eng., 12: 390-394 (1967).

    [6] J.R. Thome, Prediction of binary mixture boiling heat

    transfer coefficients using only phase equilibrium data.Int. J.

    Heat Mass Transfer, 26: 965-974 (1983).

    [7] S.G. Kandlikar, Boiling heat transfer with binary

    mixtures : Part I - A theoretical model for pool boiling. Trans.

    ASME, 120: 380-387 (1998).

    [8] J. Hovestreijdt, The influence of the surface tension

    difference on the boiling of mixtures. Chem. Eng. Sci., 18:

    631-639 (1963).

    [9] W.R. McGillis & V.P. Carey, On the role of marangoni

    effects on the critical heat flux for pool boiling of binary

    mixtures. Trans. ASME, 118: 103-109 (1996).

    [10] Y. Fujita & Q. Bai, Critical heat flux of binary mixturesin pool boiling and its correlation in terms of Marangoni

    number.Int. J. Refrig., 20: 616-622 (1997).

    [11] D.M. McEligot, Generalized peak heat flux for dilute

    binary mixtures.AIChE J., 10: 130-131 (1964).

    [12] R.P. Reddy & J.H. Lienhard, The peak boiling heat flux

    in saturated ethanol-water mixtures. Trans. ASME, 111: 480-

    486 (1989).

    [13] S.S. Kutateladze, Heat transfer in condensation and

    boiling. USAEC Report, AEC-tr-3770 (1952).

    [14] N. Zuber, Hydrodynamic aspects of boiling heat

    transfer. USAEC Report, AECU-4439 (1959).

    [15] G.I. Bobrovich, I.I. Gogonin, S.S. Kutateladze, & V.N.

    Moskvicheva, Critical heat fluxes with the boiling of binary

    mixtures.Zh.Prikl. Mekhan. I Tekh. Fiz., 4: 108-111 (1962).