Jinfeng Wu Vijay K.

Dhir1
e-mail: vdhir@seas.ucla.edu Department of Mechanical and Aerospace Engineering, Henry Samueli School of Engineering and Applied Science, University of California, Los Angeles, Los Angeles, CA 90095

Numerical Simulations of the Dynamics and Heat Transfer Associated With a Single Bubble in Subcooled Pool Boiling
In this study, numerical simulations of a vapor bubble in subcooled pool boiling have been performed. The applied numerical procedure coupling level-set function with moving mesh method has been validated by comparing the results of bubble-growth history including bubble departure diameter with data from experiments. The predictions of bubble dynamics and heat transfer for various subcoolings as well as different gravity levels are presented. DOI: 10.1115/1.4002093 Keywords: subcooled nucleate boiling, level-set method, moving mesh method, numerical simulation, condensation

Introduction

Nucleate boiling is a liquid-vapor phase-change process associated with bubble formation on the heated surface. As it is a very efcient mode of heat transfer, the boiling process has attracted the attention of many researchers in the past. Subcooled nucleate pool boiling exists when the bulk temperature of the liquid pool is below the saturation temperature of the liquid at the given system pressure, and the temperature of the heating surface exceeds the nucleation temperature, which is higher than the saturation temperature. A key parameter in determining the heat transfer coefcient under subcooled nucleate boiling conditions is the liquid subcooling itself. Gunther and Kreith 1 reported that bubbles ceased to detach as the liquid subcooling was increased. Ramanujapu 2 experimentally studied the dynamics of a single bubble under subcooled boiling conditions up to 5.5 C of subcooling. He observed that a bubble grew to a maximum size and then shrank a little before it detached. In some of the tests, it was observed that the bubble oscillated expanded and contracted for a while before departing. Based on those experiments, Ramanujapu and Dhir 3 also reported dynamics of contact angle during bubble growth and departure. Singh 4 numerically conrmed that for certain wall superheats and low liquid subcoolings, a bubble does depart at a size that is smaller than its maximum value during its growth process. However, for high subcoolings, bubbles do not detach from the wall. Various studies have also been carried out to examine reduced gravity effects on boiling heat transfer. Siegel and Keshock 5 employed a drop tower to study bubble dynamics under gravity elds in the range from 1.4% to 100% of earth-normal gravity. They observed that bubbles became quite large and the growth times were longer in reduced gravity in comparison to those at terrestrial conditions. However, the test duration of about 1 s could not guarantee that test conditions, including the ow eld and temperature distribution, were identical to that for time independent low gravity conditions. Straub et al. 6 carried out a series of boiling experiments with heaters of different geometries at low gravity in ballistic rocket ights TEXUS program, g / ge 104 and in parabolic ights of KC 135 aircraft. The authors
1 Corresponding author. Contributed by the Heat Transfer Division of ASME for publication in the JOURNAL OF HEAT TRANSFER. Manuscript received June 16, 2009; nal manuscript received June 21, 2010; published online August 10, 2010. Assoc. Editor: Satish G. Kandlikar.

reported several bubble-growth histories for R-113 at g / ge 104 and departure diameters for R-12 bubbles at both earthnormal gravity and g / ge 102. The authors proposed that evaporation at the base of a bubble was the primary heat transfer mechanism. Oka et al. 7 reported pool boiling experiments of n-pentane at gravity levels of 0.005 to 0.05ge during parabolic ights. According to their observation, the bubble size could range from 5 mm to 50 mm after bubble coalescence so as to almost cover the entire heating surface for Tw = 19 C and Tsub = 9 C. The authors also noted that the bubble oscillation enabled continuous supply of the liquid from the bulk to the wall. Qiu and Dhir 8 presented an experimental study on growth and detachment of a single bubble on a heated surface with water as the test uid for 0.04ge. Bubble growth time, bubble size, and shape from nucleation to lift-off were measured under subcooled and saturated conditions. In the limited range of wall superheats and liquid subcoolings, the effects of wall superheat and liquid subcooling on bubble lift-off diameter were found to be small; however, the growth periods were very sensitive to liquid subcooling at a given wall superheat. They also showed that for saturated liquid, bubble departure diameter approximately varied as g1/2 and growth period changed about inversely with the level of gravity. Son et al. 9 presented a bubble-growth model based on the level-set method, which has been proven to be quite effective in correlating the experimental data. The model developed by Son et al. 9 laid a foundation for the current work. In the present study, a numerical procedure, which is capable of redistributing the mesh and sustains a high node density around the interface as the bubble grows, is used to model the process during subcooled nucleate boiling under various gravity levels. This approach is used to obtain a higher degree of accuracy in calculating the liquid side heat transfer.

Numerical Formulation

2.1 Model Description. To analyze the growth of a single bubble in subcooled nucleate boiling, we extend the numerical model originally developed by Son et al. 9. In that model, the computational domain was divided into two parts: a microregion and a macroregion, as shown in Fig. 1. The microregion is a thin lm that lies underneath the bubble, whereas the macroregion consists of the bubble and the liquid surrounding the bubble. In the present work, numerical simulations of uid ow and heat transfer are carried out in both micro- and macroregions for a NOVEMBER 2010, Vol. 132 / 111501-1

Journal of Heat Transfer

Copyright 2010 by ASME

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 03/27/2014 Terms of Use: http://asme.org/terms

y=Y Macro region

Liquid

Vapor g T=Tsat

Wall

r r=R

Micro region

0 r = R0 Wall

h/2 r = R1

Fig. 1 Macro- and microregions in numerical simulation

given contact angle. The computed shape of the interface in the microregion and the macroregion is matched at the outer edge of the microlayer. In the numerical simulation, a level-set function is used to represent the shape of the interface in the macroregion. 2.2 Assumptions. The following assumptions are made in this study: the process is two-dimensional and axis-symmetric; the ows are laminar; the wall temperature remains constant; water at atmospheric pressure is used as the test uid; vapor remains at its saturation temperature; the thermodynamic properties of the individual phases are assumed to be insensitive to small changes in temperature and pressure except for surface tension.

pressure and saturation temperature of 100 C. 2.4 Governing Equations for the Macroregion. To numerically analyze the macroregion, we use the level-set formulation developed by Son et al. 9 for nucleate boiling of pure liquid. The interface separating the two phases is captured by solving the following equation for the level-set function, :

= uint t
where uint = u + m

For a constant heat ux surface, the temperature of the solid wall will vary because of the spatial and temporal dependences of the liquid side heat transfer coefcient 10. To resolve the variation in temperature, one must solve conjugate heat transfer problem in the solid. In the present work, the temperature of the solid is assumed to remain constant with the premise that effect of temperature variation on bubble dynamics will be of second order. 2.3 Thermal and Physical Properties. The thermal and physical properties used in carrying out the computation are listed in Table 1. All properties are evaluated for water at atmospheric
Table 1 Thermal and physical properties Property Unit kg / m3 kJ / kg K W/m K m2 / s kJ/kg K K1 N/m Liquid 958 4.212 0.68 2.85 104 2257 7.5 104 0.0589 Vapor 0.598 2.02 0.0248 1.2 105 373.15

and m is the evaporation/condensation-rate vector through the bubble interface see Ref. 9 for details. Reinitialization equation for is solved until steady state is reached to ensure that = 1:

= signo1 t

In the above equation, 0 is the solution of Eq. 1. The material properties are assumed to be constant in the individual phases, except near the interface and in a thin region around the interface. To describe such an interface, we dene the Heaviside function, H, as follows: H=1 =0 if if

+ 1.5h 1.5h
if 1.5h 4

cp k h fg Tsat

=0.5 + /3h + sin2/3h/2

where h is equal to the grid spacing on a uniform grid, and H is 1 in the liquid phase and 0 in the vapor phase. The interface is spread over an interval of 3h, so that the material properties change continuously at the interface. In terms of H, the properties of interest are dened as follows:

= v + l v H

111501-2 / Vol. 132, NOVEMBER 2010

Transactions of the ASME

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 03/27/2014 Terms of Use: http://asme.org/terms

1 1 1 = v + l1 v H

6 7

k1 = kl1H

where , , and k denote the density, the uid viscosity, and the thermal conductivity, respectively. The subscripts v and l represent the vapor and uid phases, respectively. Also, Eq. 7 is consistent with the assumption that the vapor temperature remains constant at Tsat. The interfacial curvature is expressed in terms of the level-set function as follows:

outer end, where it joins with the macroregion. Lay and Dhir 11 modeled and numerically solved for the shape of the microlayer underneath a bubble using lubrication theory. In carrying out the analysis, is taken to be the thickness of the microlayer measured from the wall, and r is the radial coordinate. The quasistatic mass conservation, momentum, and energy equations in the microlayer are given as q/lh fg = 1 r r

ruldy

18

2u l pl = l 2 y r
q = klTw Tint/ = hevTint Tv + pl pvTv/lh fg

19 20

The governing equations of continuity, momentum, and energy conservation for the macroregion are

u + u u = p + g TT Tsatg H t

+ u = 0 t

where Tw is the wall temperature and Tv is the vapor temperature. Tint is the interface temperature, which is approximately Tw at the inner end and Tv at the outer end. A typical prole of Tint has been shown in Ref. 9. Here pv is the vapor pressure and hev is the evaporation heat transfer coefcient. The evaporation heat transfer coefcient is obtained from kinetic theory as RTv0.5vh2 hev = 2 M / fg/Tv, Tv = Tsat pv 21

c pl

+ u + uT

T + u T = k T t
T = Tsat pv for

10 H0 11

for

The pressures in the vapor and liquid phases satisfy the following relation 11: pl = pv A

H=0

Since the vapor in the bubble was assumed to remain at the saturation temperature, the energy equation in the vapor is not solved. The mass conservation equation Eq. 9 can be rewritten, while noting that / t = uint , as, u= kT m + V + V micro = micro 2 h fg2 12

q2

vh 2 fg

22

where V micro is the volume addition attributed to the heat transfer from the microlayer, which is V micro =

where surface tension, , is a function of temperature, and A is the dispersion constant in the disjoining pressure and its magnitude can be related to the contact angle, which is prespecied for a given liquid-solid combination. In Eq. 22, the second term on the right hand side accounts for the capillary pressure, the third term accounts for the disjoining pressure, and the last term accounts for the recoil pressure. The curvature of the interface is dened as

R1

R0

klTw Tint rdr vh fgVmicro

13

1 r / r r r

1+

23

Vmicro is a vapor-side control volume near the microregion. Equations 1012 are nondimensionalized using the characteristic length, time, and velocity scales, l0, t0, and u0, respectively: l0 =

The combination of the mass, momentum, and energy equations for the microlayer yields

= f , , ,

24

u0 = gl0 =

t0 =

l0 = 3 u0 g l v

g l v g l v

14

where denotes / r. The boundary conditions for the above equation are imposed as follows: At r = R0 inner end of microlayer,

1/4

= o,
15 16

= = 0

25

1/4

where o is of the order of molecular size and it can be obtained from Ref. 12 at the junction of the evaporating and nonevaporating regions. At r = R1 outer end of microlayer,

The temperature is nondimensionalized such that the wall temperature is 1 and the subcooled liquid temperature is 0, i.e.,

= h/2,

= 0

26

T Tl T Tl T Tl = = Tw Tl Tw Tsat + Tsat Tl Tw + Tsub

17

The governing equations Eqs. 1012 are solved throughout the domain to obtain the velocity, temperature, and pressure in each cell. The detailed computational framework is discussed later. 2.5 Governing Equations for the Microregion. This region is illustrated in Fig. 1. The thickness of the microlayer varies from a few molecules at the inner end to a few micrometers near the Journal of Heat Transfer

where h is the grid spacing and h / 2 is the vertical distance to the rst computational node for the level-set function, , on uniform grids from the wall. In implementing the above boundary conditions, the radius R1 is determined from the solution of the macroregion. For a given dispersion constant, the microlayer formulation, Eq. 24. and R0 are solved with the ve boundary conditions Eqs. 25 and 26. In this work, an apparent contact angle is dened as tan = 0.5h/R1 R0 27

is measurable experimentally and used as boundary conditions in level-set function. Equation 24 is numerically integrated using
NOVEMBER 2010, Vol. 132 / 111501-3

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 03/27/2014 Terms of Use: http://asme.org/terms

Equivalent diameter [mm]

a RungeKutta method with a separate code. An expression for the rate at which vapor is produced from the microlayer is given in Eq. 13. Within moving meshes, the distance to the rst computational node from the wall will no longer be equal to h / 2 on a uniform grid. It also may vary slightly from time to time. The above equations are still valid when nonuniform grids are used.

3.5 3 2.5 2 1.5 1 0.5 0 0 10 20 Numerical, 6th cycle Numerical, 7 cycle th Numerical, 8 cycle Experimental, cycle 1 Experimental, cycle 2 Experimental, cycle 3 30 40 Time [ms] 50 60 70 80
th

The Computational Framework

In the current study, a numerical procedure in conjunction with moving mesh method is applied to compute the level-set function. Contravariant velocity components in curvilinear coordinates , are taken as primary variables. For detailed information concerning moving mesh generation and related equations, the reader is referred to Ref. 13. The resulting equations contain convection terms of the rst-order derivatives with respect to , , and diffusion terms of the second-order derivatives with respect to , , , and . Taking advantage of the bubble symmetry, we only need to compute one-half of the bubble. We use a staggeredgrid nite difference scheme. The scalar parameters are dened at the centers of cells and velocity components are stored at the edges of cells. To easily obtain the discretized forms of various quantities, both Cartesian velocity components and contravariant ones are stored in memory. We use upwind differencing for advection terms and central differencing for diffusion terms. A projection method is used to solve for velocities and pressure. Because the computation for pressure takes most of the computational time, we combine multigrid and conjugate gradient methods so that the numerical solver for pressure converges in less than ten iterations under most circumstances. This in turn results in signicant computational savings. Overall we have the following computational framework: 1. Initialize mesh distribution by solving the mesh equation to steady state. Incorporate level-set function into mesh equation. The resulting mesh would t well with the initial conditions. It also should be noted that the changes in other variables, such as temperature and temperature gradient, can be easily taken into account in the mesh adaptation process. 2. Update mesh. Using t in the mesh equations, solve for one time step to evolve the mesh in forward direction. 3. Solve the level-set advection equation, reinitialize the levelset function, and determine the properties (density, viscosity, and thermal conductivity). The second order essential nonoscillatory (ENO) scheme is applied to discretize and . A few iteration steps are used in the reinitialization procedure. We use the interface width of 3h to bridge the property difference across the interface. 4. Solve the energy equation for temperature in the liquid. The upwind differencing is used for T and T. The diffusion terms, T and T, are implicitly discretized, and T and T are continuously updated by the current iteration until convergence. 5. Solve the momentum equation for intermediate velocities using the pressure at the previous time step. 6. Solve the Poisson equation for pressure. The governing equation for pressure contains V micro. Under the assumptions of constant wall temperature and constant contact angle, V micro is a function of the distance between the rst computational node and the wall. This function is obtained from solution of microlayer beforehand and is incorporated into this numerical procedure. During each time step, the average distance of several rst computational nodes along the wall at the interface region is used to retrieve the corresponding microlayer heat transfer. 7. Correct the velocities corresponding to the updated pressure. This ensures that the continuity equation is satised. 8. Go to step 2 for the next time step. 111501-4 / Vol. 132, NOVEMBER 2010

(a)
14 12 10 Nu 8 6 4 2 0 0 200 400 600 Time [ms] 800 1000 1200

(b)
Fig. 2 a Growth rate. b Nu versus time for wall superheat = 7 C, liquid subcooling= 1.5 C, contact angle= 54 deg, pressure= 1.013 105 Pa, and g / ge = 1.

3.1

Boundary Conditions. At the wall y = 0, u = 0,

= cos , y = 0, y = 0, r

T = Tw

28

At the top of computational domain y = Y ,

u = 0, y

v = 0, y v = 0, r

T = Tl

29

At the planes of symmetry r = 0 , R, u = 0,

T =0 r

30

3.2 Initial Conditions. Initially, the uid velocity is set to zero. The temperature prole is taken to be linear in the natural convection thermal boundary layer, and its thickness, T, is given by Kays and Crawford 14:

T = 7.14ll/gTTw Tl1/3

31

Validation

To validate this moving mesh method coupled with level-set function, we studied the case of bubbles rising in a quiescent liquid and compared the results to those given by Ryskin and Leal 15 and Son 16. A phase-change problem with an analytical solution described in Ref. 17 was also used to test the capability of our method to include heat transfer. The numerical results for these two cases are reported in Ref. 13. The results of numerical simulations reported here are also validated by comparing the results with the experimental data. Transactions of the ASME

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 03/27/2014 Terms of Use: http://asme.org/terms

t = 0ms
2 2

t = 845.3ms

1.5

1.5

y / l0

y / l0

1 99oC o 100 C 98.6 C 100.2oC 99.5 C o 101 C

0.5

0.5

-1

-0.5

0 r / l0

0.5

-1

-0.5

0 r / l0

0.5

t = 862.9ms
2 2

t = 872.5ms

1.5

1.5 98.6oC 99.5oC o 101 C

y / l0

1 99oC 0.5 100 C

100.2oC

98.6oC 99.5oC 101oC

y / l0

1 99 C 0.5 100 C
o o

100.2 C

-1

-0.5

0 r / l0

0.5

-1

-0.5

0 r / l0

0.5

t = 891.7ms
2 2

t = 917.3ms
100.2oC 98.6oC

1.5 100.2oC y / l0 1 99oC 100 C 0.5

1.5 98.6oC 99.5 C 101oC 0.5

y / l0

99 C 100oC

99.5oC 101 C
o

-1

-0.5

0 r / l0

0.5

-1

-0.5

0 r / l0

0.5

Fig. 3 Flow eld and temperature distribution for wall superheat= 7 C, liquid subcooling= 1.5 C, contact angle= 54 deg, pressure= 1.013 105 Pa, and g / ge = 1

In carrying out numerical simulation, the computational domain is chosen to be R / l0 , Y / l0 = 1 , 3 for a series of bubbles and R / l0 , Y / l0 = 1 , 2 for a single cycle in most cases except for g / ge = 0.0001, to minimize the effects of the computational boundary and save computation time. The initial bubble size is determined by the size of grid spacing, which is D / l0 = 0.01 for normal gravity and g / ge = 0.01. The details for g / ge = 0.0001 are provided in Sec. 5. Considering the complexity involved in boiling phenomenon, it is impossible to simulate every aspects of bubble dynamics and heat transfer during the process, and some simplications have to be made. The constant wall temperature is one of them, although in reality it can vary with time and position due to the high evaporation rate near the contact line. In addition, the advancing and receding contact angles differ and depend on the interface velocity. In the present work, a static contact angle is used, as the advancing and receding contact angles are expected to deviate only about 5 deg from the static contact angle 3. Because Journal of Heat Transfer

the conjugate heat transfer problem in the solid is not solved, waiting time between successive bubble-growth cycles cannot be predicted and must be specied empirically. 4.1 Case 1: Subcooled Nucleate Boiling Under EarthNormal Gravity. The prediction from the numerical model was compared with the experimental data of Ramanujapu 2. The growth of an isolated bubble on a silicon wafer was studied in those experiments. A square cavity, 10 m in width and 20 m in depth, served as the nucleation site. The wafer was heated from below by controlling power through strain gauge heaters while maintaining surface temperature nearly constant. In the experiments, degassed and de-ionized water was used as the test uid. The heated region around the cavity was much larger than the radius of the computational domain. Further experimental details are given in Ref. 2. Figure 2a shows the comparison for the case where Tw = 7 C, Tsub = 1.5 C, contact angle of 54 deg, NOVEMBER 2010, Vol. 132 / 111501-5

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 03/27/2014 Terms of Use: http://asme.org/terms

3 2.5 Equivalent diameter [mm]

3 2.5 Equivalent diameter [mm] 2 1.5 1 0.5 0 Bubble Departure

Tsub = 0 C o Tsub = 1 C o Tsub = 5 C o Tsub = 7.5 C

2 1.5 1 0.5 0

Computation 1 Computation 2 Experiment 1 Experiment 2 Experiment 3 0 20 40 60 Time [ms] 80 100 120

0.05

0.1

0.15

0.2 Time [s]

0.25

0.3

0.35

0.4

Fig. 4 Growth rate for wall superheat= 6.5 C, liquid subcooling= 4 C, contact angle= 54 deg, pressure= 1.013 105 Pa, and g / ge = 1

(a)
14 12 10 8 Nu 6 4 2 0

14 12 10 8 6 4 2 0

Experimental data (Qiu and Dhir [7]) Numerical simulation

Equivalent diameter [mm]

Tsub = 0oC o Tsub = 1 C o Tsub = 5 C o Tsub = 7.5 C 0 0.05 0.1 0.15 0.2 Time [s] 0.25 0.3 0.35 0.4

(b)
Bubble Departure 0 2 4 6 Time [s] 8 10 12

Fig. 7 a Growth rate. b Nu versus time for wall superheat = 8 C, contact angle= 38 deg, pressure= 1.013 105 Pa, and g / ge = 1

Fig. 5 Growth history for wall superheat= 2.5 C, liquid subcooling= 0.4 C, contact angle= 54 deg, pressure= 1.013 105 Pa, g / ge = 0.045, and water as test uid

7 Departure Diameter [mm] 6 5 4 3 2 1 0 0 0.1

Fritz (Straub et al. [5]) Experimental data (Straub et al. [5]) Current study

0.2

0.3

0.4

0.5 0.6 P/Pc

0.7

0.8

0.9

Fig. 6 Comparison of departure diameter of R-12 versus pressure for g / ge = 0.01, contact angle= 25 deg, wall superheat = 8 C, and liquid subcooling= 0 C

and pressure of 1.013 105 Pa. This value of static contact angle has been measured in experiments for water on a clean polished silicon surface. The data from three different bubble release cycles are compared with the results for three sequential bubbles from simulation. It can be seen that there is a good agreement with the data throughout the growth cycle. Although the model overpredicts the departure diameter by about 10% under these conditions, the predicted growth rate and growth periods match well with those found in the experiments. Figure 2b shows the variation of predicted Nusselt number based on the heater area 19.6 mm2 averaged heat ux at the wall with time for eight bubble release cycles from the beginning until quasisteady state condition is achieved. In each cycle, the minimum Nusselt number occurs after the bubble lifts off and just prior to the birth of a new bubble. It is fortuitous that, under quasisteady state conditions represented by sixth, seventh, and eighth cycles, the minimum Nusselt numbers in the absence of bubbles comparable to the initial Nu assumed to correspond to steady state natural convection. It should be noted that Nusselt number values higher than the minimum indicate the enhanced heat transfer rate caused by the formation of bubbles. During the rst cycle, the thin thermal boundary layer shown in Fig. 3 at t = 0 ms provides the least heat transfer to the growing bubble. As a result, the rst bubble takes much longer time to grow and depart than subsequent bubbles. The thin thermal boundary layer implies a steep temperature gradient close to the wall everywhere Transactions of the ASME

111501-6 / Vol. 132, NOVEMBER 2010

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 03/27/2014 Terms of Use: http://asme.org/terms

0.6 Total Heat Transfer Rate [w] 0.5 0.4 0.3 0.2 0.1 0 -0.1 -0.2 0 0.05 0.1 0.15 0.2

o Tsub = 0 C Tsub = 1oC Tsub = 5oC Tsub = 7.5oC

t = 0.009 s
2.5E+05 2.0E+05 1.5E+05 q [w/m ] 1.0E+05 5.0E+04 0.0E+00
2

Tsub = 0 C Tsub = 1oC o Tsub = 5 C Tsub = 7.5oC

0.25

0.3

0.35

0.4

-5.0E+04 -1.0E+05 0 10 20 30 40 50 60 70 80 90 Angular Position, degrees

Time [s]

Fig. 8 Total heat transfer rate through the interface and microlayer for of wall superheat= 8 C, contact angle= 38 deg, pressure= 1.013 105 Pa, and g / ge = 1
2.0E+05

t = 0.27 s
o Tsub = 5 C Tsub = 7.5oC

in the computational domain and yields a larger value of Nu. With time, the ow eld that is created causes the thermal layer to thicken around the bubble generation site. As a result, Nu decreases and more heat is stored in the liquid. This leads to a higher rate of evaporation and reduces growth period. When the bubble base shrinks, cold liquid from the sides of the domain ows in and lls the volume. This ow pattern in turn compresses the thermal boundary layer and intensies the heat transfer. For example, the maximum heat transfer rate in seventh cycle occurs at 910.5 ms as the base radius decreases to 0.57 mm from the maximum value of 0.89 mm prior to departure. Once quasisteady state is achieved, Nu does not alter much from cycle to cycle. This trend is indicated by sixth, seventh, and eighth cycle in Fig. 2b. The waiting time is not provided with the experiment data. In numerical simulations, too small waiting time will lead to bubble merger in vertical direction and a large waiting time will increase computational time considering many cycles of bubbles. A waiting time of 16.6 ms was empirically used in between successive bubble cycles. The initial conditions for the rst and seventh bubble are shown in Fig. 3 at t = 0 and 845.3 ms, respectively. The much thicker boundary layer for the seventh cycle, which is similar to sixth and eighth cycled, gives rise to higher growth rate and smaller growth period 70 ms in comparison to the rst bubble, which has a growth period of 250 ms. Figure 3 provides a sequence of the time-dependent isotherm distributions and velocity elds for the seventh cycle. From Fig. 3, it can be observed that the superheated plume represented by the contour of 100.2 C stays above the top of the bubble throughout this period and the bubble is exposed to the subcooled liquid only during the late portion of the growth period. This process is typical for sixth and eighth cycles as well. It should be noted that as reported in Ref. 9 for saturated case, microlayer contributes about 20% to the total heat transfer rate into the bubble. The impact of initial conditions on bubble-growth rate is further demonstrated for a higher subcooling case in Fig. 4, which compares the data from experiments with prediction from numerical simulations for Tw = 6.5 C, Tsub = 4 C, contact angle of 54, and pressure of 1.013 105 Pa. The data are from Ref. 2. Numerical computation 1 is based on initial thermal layer thickness given by Eq. 31 and on the assumption that temperature varies linearly in the thermal layer. For this case, bubble reaches a quasistatic size of 1.86 mm and does not depart up to 120 ms. Numerical computation 2 is based on the dimensionless temperature distribution that will exist after several bubble release cycles e.g. Fig. 3 at t = 845.3 ms. The prediction from the second set of calculation is in more agreement with the data from experiments, where a bubble of about 2.4 mm in size departs at about 110 ms. Thus the initial condition during subcooled boiling signicantly affects the bubble dynamics. Normally initial conditions in the Journal of Heat Transfer

1.5E+05 1.0E+05 q [w/m2] 5.0E+04 0.0E+00 -5.0E+04 -1.0E+05 0 10

20 30 40 50 60 70 Angular Position, degrees

80

90

Fig. 9 Heat ux as a function of location along bubble interface for wall superheat= 8 C, contact angle= 38 deg, pressure = 1.013 105 Pa, and g / ge = 1

experiments after several bubble nucleation cycles are not known and as a result differences may exist between results of experiments and numerical simulations. To avoid inconsistencies, all the calculations in this work were performed for the rst cycle based on the initial linear distribution of temperature in the thermal layer. In carrying out the calculations, the domain size was varied parametrically and little effect on ow eld and bubble-growth history was observed for domains larger than those used in the reported work. Also bubbles were allowed to escape from the free surface with liquid, lling the vacated space at the top of the domain. 4.2 Case 2: Subcooled Nucleate Boiling Under Microgravity. This section presents the results for the cases where the gravity levels are much lower than the earth-normal gravity. Under these conditions, not only is the size of the bubble larger but also the growth periods are extended. The data from the KC-135 ights are used to validate results for the subcooled cases. The detailed information related to the experimental setup is given by Qiu and Dhir 8. The initial condition is different in the microgravity model, as compared to that at earth-normal gravity in term of the thickness of the thermal layer. The low gravity in the KC-135 ight lasts about 20 s. This relatively short period of time does not allow for the thermal layer to grow and develop over a period of several cycles. Therefore, the conditions at the start of the experiments are a relatively static pool and a growing thermal layer. For the microgravity simulations, we take the results from the rst cycle of the bubble growth itself. Assuming heating was started 2 s prior to bubble initiation, the initial thermal boundary layer thickness from NOVEMBER 2010, Vol. 132 / 111501-7

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 03/27/2014 Terms of Use: http://asme.org/terms

t = 0.1ms
2 2

t = 1.4ms

1.5

1.5

y / l0

y / l0

0.5

100.08oC

0.5

100.08 C

-1

-0.5

0 r / l0

0.5

-1

-0.5

0 r / l0

0.5

t = 9.4ms
2 2

t = 25.3ms

1.5

1.5

y / l0

y / l0

0.5

100.08oC

0.5

100.08oC

-1

-0.5

0 r / l0

0.5

-1

-0.5

0 r / l0

0.5

Fig. 10 Temperature distribution and velocity eld for wall superheat= 8 C, liquid subcooling= 0 C, contact angle= 38 deg, pressure= 1.013 105 Pa, and g / ge = 1 temperature increment between isotherms is 1.58 C

t = 0.1ms
2 2

t = 1.4ms

1.5

1.5

y / l0

y / l0

0.5 99.09oC 0 -1 -0.5 0 r / l0 0.5 1

0.5

99.09 C 0

-1

-0.5

0 r / l0

0.5

t = 9.4ms
2 2

t = 25.3ms

1.5

1.5

y / l0

y / l0

0.5

99.09oC

0.5

99.09oC

-1

-0.5

0 r / l0

0.5

-1

-0.5

0 r / l0

0.5

Fig. 11 Temperature distribution and velocity eld for wall superheat= 8 C, liquid subcooling= 1 C, contact angle= 38 deg, pressure= 1.013 105 Pa, and g / ge = 1 temperature increment between isotherms is 1.78 C

111501-8 / Vol. 132, NOVEMBER 2010

Transactions of the ASME

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 03/27/2014 Terms of Use: http://asme.org/terms

t = 0.12ms
2 2

t = 121.3ms

1.5

1.5

y / l0

y / l0

0.5 95.13 C 0 -1 -0.5 0 r / l0 0.5 1

0.5

95.13oC

-1

-0.5

0 r / l0

0.5

t = 265.1ms
2 2

t = 361.0ms

1.5

1.5

y / l0

1 95.13oC

y / l0

1 95.13oC

0.5

0.5

-1

-0.5

0 r / l0

0.5

-1

-0.5

0 r / l0

0.5

Fig. 12 Temperature distribution and velocity eld for wall superheat= 8 C, liquid subcooling= 5 C, contact angle= 38 deg, pressure= 1.013 105 Pa, and g / ge = 1 temperature increment between isotherms is 2.57 C

transient conduction analysis has been set to 1 mm. Figure 5 displays the comparison of the bubble-growth history between the data and results of simulation for wall superheat = 2.5 C, liquid subcooling= 0.4 C, contact angle= 54 deg, pressure= 1.013 105 Pa, and g / ge = 0.045. The predicted departure diameter is in very good agreement with the data, although the predicted time period is longer by about 20%. Model overpredicts the bubble diameter during the midrange of the growth history of the bubble. The reason for this could be that gravitational acceleration in KC-135 is rarely constant. The variation in acceleration in ight affects the thickness of the thermal layer around the bubble and hence the growth rate. Computed departure diameters for R-12 bubbles at gravity level of g / ge = 0.01 are compared with experimental data of Straub et al. 6 in Fig. 6. From the calculated departure sizes by Straub et al.
t = 1.4 ms t = 9.4 ms

6 using Fritz equation and the properties of R-12, contact angles ranging from 18 deg to 30 deg at different p / pc were expected in their experiments. Therefore, a contact angle of 25 deg was assumed in carrying out the numerical simulation for all p / pc. Overall the calculated departure diameters agree with the experimental data as p / pc ranges from 0.1 to 0.6. In the worst case of p / pc = 0.1, the numerical simulation underpredicts the departure diameter by 13%. Based on the above comparisons with the benchmark test cases, we are condent about the validity of our numerical simulations.

Results and Discussions

In this section, the numerical results for saturated and subcooled boiling of water as test uid under various gravity levels are presented. The saturated cases are limiting cases. All the numerical results correspond to the computations of bubble growth and departure for the rst cycle. 5.1 Earth-Normal Gravity, g ge = 1 . Numerical simulations of single bubble dynamics during saturated and subcooled nucleate boiling of water for a wall superheat of 8 C, a contact angle of 38 deg, and a system pressure of 1.013 105 Pa were performed. This contact angle is considered for water on a mildly oxidized copper surface. The growth history of the bubble is used as the primary means of comparing the effect of various parameters. The actual volume of the bubble is rst computed; thereafter it is converted into an equivalent diameter of a complete sphere. Figure 7a shows the growth histories of the bubble with four different subcoolings of 0 C, 1 C, 5 C, and 7.5 C, respectively. For the saturated case, the bubble continues to increase in size until it nally departs from the wall after attaining a diameter of 2.43 mm. For liquid subcooling of 1 C, the bubble is predicted to detach from the wall with a diameter of 2.22 mm. For subcoolings of 5 C and 7.5 C, the bubble grows to a maximum size 1.79 NOVEMBER 2010, Vol. 132 / 111501-9

1.5

1.5

y / l0

y / l0 0 0.5 r / l0 1

0.5

0.5

0.5 r / l0

Fig. 13 Evolving grid distribution

Journal of Heat Transfer

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 03/27/2014 Terms of Use: http://asme.org/terms

25

Equivalent diameter [mm]

20

Tsub = 0oC o Tsub = 1 C o Tsub = 5 C Tsub = 7.5oC

t = 0.7 s
3.0E+04 2.0E+04
o Tsub = 0 C o Tsub = 1 C Tsub = 5oC o Tsub = 7.5 C

15

10

q [w/m ]

1.0E+04 0.0E+00 -1.0E+04 -2.0E+04

Bubble Departure

4 Time [s]

10

10

20

30

40

50

60

70

80

90

Angular Position, degrees

(a)
25

t = 9.9 s
3.0E+04
o

20

2.0E+04 1.0E+04 0.0E+00 -1.0E+04 -2.0E+04 0

Tsub = 5 C Tsub = 7.5oC

10 Tsub = 0oC o Tsub = 1 C o Tsub = 5 C o Tsub = 7.5 C 0 2 4 Time [s] 6 8

q [w/m2]

15 Nu

10

20 30 40 50 60 70 Angular Position, degrees

80

90

(b)
Fig. 14 a Growth rate. b Nu versus time for wall superheat= 8 C, contact angle= 38, pressure= 1.013 105 Pa, and g / ge = 0.01

Fig. 16 Heat ux as a function of location along bubble interface for wall superheat= 8 C, liquid subcooling= 0 C, contact angle= 38 deg, pressure= 1.013 105 Pa, and g / ge = 0.01

mm for 5 C, 1.49 mm for 7.5 C and then begins to shrink. This shrinkage is due to the large condensation heat transfer that occurs at the upper portions of the bubble, which becomes larger than the sum of evaporative heat transfer around the bubble and from the microlayer. Eventually a balance is reached between the evaporation and condensation rates, and the bubble acquires a quasistatic diameter while remaining attached to the heated surface 1.70 mm
5 Total Heat Transfer Rate [w] 4 3 2 1 0 -1 -2 -3 0 2 4 Time [s] 6 8 10
o Tsub = 0 C Tsub = 1oC Tsub = 5oC Tsub = 7.5oC

Fig. 15 Total heat transfer rate through the interface and microlayer for wall superheat= 8 C, contact angle= 38 deg, pressure= 1.013 105 Pa, and g / ge = 0.01

for 5 C subcooling and 1.36 mm for 7.5 C subcooling. The bubble in these cases never achieves the bubble departure size. Nusselt number based on the wall area average heat transfer rate from the wall over the computational domain is shown in Fig. 7b. The heat transfer coefcient is based on the wall superheat. Initially Nu is higher because of the higher subcooling and the thinner thermal boundary layer. It is found that Nu for 5 C subcooling slightly exceeds that for 7.5 C subcooling after 0.31 s. Possible reason could be the role played by bubble base size: More heat is transmitted into the vapor from the microlayer for the bubble with the larger base area; on the other hand, a larger bubble would have a smaller area through which heat is transferred from the wall to the liquid. Figure 8 shows the total heat transfer rates into bubble including the contribution from the microlayer and the bubble interface for the four subcoolings. The initial high rate of heat transfer accounts for the rapid bubble growth in the early stages. As the bubble approaches its departure size, the heat transfer rate decreases. For subcoolings of 5 C and 7.5 C, heat transfer rate rapidly approaches zero as condensation balances evaporation and bubble size becomes almost static. Figure 9 presents the interfacial heat ux as a function of location along the interface at two times during the bubble growth for various subcoolings. The evaporation contribution from microlayer is not considered here. As expected, larger subcooling results in higher condensation rates. At t = 0.009 s, the bubbles are in the process of rapid growth. The top of the bubbles are exposed to the subcooled liquid. This yields a steep temperature gradient, which Transactions of the ASME

111501-10 / Vol. 132, NOVEMBER 2010

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 03/27/2014 Terms of Use: http://asme.org/terms

t = 0.01s
2 2

t = 0 .7 s

1.5

1.5

y / l0

y / l0

0.5 100.08oC 0 -1 -0.5 0 r / l0 0.5 1

0.5 100.08 C 0 -1 -0.5 0 r / l0 0.5 1

t = 1.5s
2 2

t = 2.2s

1.5

1.5

y / l0

y / l0

0.5 100.08 C 0 -1 -0.5 0 r / l0 0.5 1

0.5 100.08oC 0 -1 -0.5 0 r / l0 0.5 1

Fig. 17 Temperature distribution and velocity eld for wall superheat= 8 C, liquid subcooling= 0 C, contact angle= 38 deg, pressure= 1.013 105 Pa, and g / ge = 0.01 temperature increment between isotherms is 1.58 C

t = 0.01s
2 2

t = 1.2s

1.5

1.5

y / l0

y / l0

0.5 99.09 C 0 -1 -0.5 0 r / l0 0.5 1

0.5 99.09 C 0 -1 -0.5 0 r / l0 0.5 1

t = 2.7s
2 2

t = 4.3s

1.5

1.5

y / l0

y / l0

0.5 99.09oC 0 -1 -0.5 0 r / l0 0.5 1

0.5 99.09 C 0 -1 -0.5 0 r / l0 0.5 1

Fig. 18 Temperature distribution and velocity eld for wall superheat= 8 C, liquid subcooling= 1 C, contact angle= 38 deg, pressure= 1.013 105 Pa, and g / ge = 0.01 temperature increment between isotherms is 1.78 C

Journal of Heat Transfer

NOVEMBER 2010, Vol. 132 / 111501-11

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 03/27/2014 Terms of Use: http://asme.org/terms

t = 0.01s
1.5 1.5

t = 2.2s

1 y / l0 y / l0 0.5 95.13 C 0 -0.75 -0.5 -0.25 0 r / l0 0.25 0.5 0.75

0.5 95.13 C 0 -0.75 -0.5 -0.25 0 r / l0 0.25 0.5 0.75

t = 6 .0 s
1.5 1.5

t = 9.8s

1 y / l0 y / l0 0.5 95.13 C 0 -0.75 -0.5 -0.25

0.5

95.13 C

0 r / l0

0.25

0.5

0.75

0 -0.75 -0.5 -0.25

0 r / l0

0.25

0.5

0.75

Fig. 19 Temperature distribution and velocity eld for wall superheat= 8 C, liquid subcooling= 5 C, contact angle= 38 deg, pressure= 1.013 105 Pa, and g / ge = 0.01 temperature increment between isotherms is 2.57 C

results in high condensation rates. At t = 0.27 s, the bubble size remains relatively constant and lower condensation rates occur as thermal layer develops on the liquid side. Figures 1012 show the calculated ow eld and isotherm distribution for subcoolings of 0 C, 1 C, and 5 C, respectively. In each one of those gures, the top isotherm represents a dimensionless temperature of 0.01 and the rest is divided by ve equal increments. From Fig. 10, it can be seen that the growing bubble initially pushes the liquid radially out. The location where the vapor-liquid interface contacts the wall is observed to move outward and then inward as the bubble grows and departs. The highest heat transfer rate occurs around the base of the bubble and this is reected by the packing of isotherms. The nonuniform vapor velocity inside the bubble results in a noticeable clockwise vortex.

For liquid subcoolings of 1 C and 5 C, the isotherm that terminates at the bubble interface represents the saturation temperature. It separates the areas where evaporation takes place from the areas where condensation occurs. We observe that the condensation areas dominate over the evaporation areas during most of the growth period. For 5 C subcooling, vapor owing upward from the bubble base condenses over most of the interface. The liquid that has just condensed around the interface ows downward toward the wall and thins down the thermal layer near the base of the bubble. This in turn leads to saddle points in the isotherms. Figure 13 represents the grid structure corresponding to Fig. 10 at t = 1.4 and 9.4 ms, respectively. For clarity, only every other point is plotted. 5.2 Microgravity. The results for the cases where the gravity level is 1% of earth-normal gravity are included here. At this gravity level, not only is the size of the bubble an order of magnitude larger, but also the growth time periods are much longer. Figure 14a shows the bubble-growth rates for four different liquid subcoolings for a wall superheat of 8 C, a contact angle of 38 deg, and a system pressure of 1.013 105 Pa. The trends of bubble growth are similar to those for earth-normal gravity. At microgravity, subcooling has a more pronounced effect on bubble size. For example, at g / ge = 1.0 as quasisteady state is reached, the ratios of Dsub / Dd,sat are 0.7 and 0.56 for liquid subcoolings of 5 C and 7.5 C, respectively. However, at g / ge = 0.01, they are 0.49 and 0.38, respectively. The corresponding area 1963 mm2 averaged heat transfer coefcient from the wall during this process is shown in Fig. 14b. Although Nu values at 0.01ge are higher than those at 1ge, the actual heat transfer rate is lower. Here one should note that characteristic length used in dening Nu is ten times larger at 0.01ge than at 1ge. Figure 15 depicts the total heat transfer rate across the interface for various liquid subcoolings. At subcooling of 1 C, the maximum heat transfer rate, corresponding to the maximum bubble Transactions of the ASME

250

Equivalent diameter [mm]

200

Tsub = 0 C o Tsub = 5 C o Tsub = 7.5 C

150

100 Computation Stopped

Bubble Departure

50

20

40

60

80

100 120 Time [s]

140

160

180

200

Fig. 20 Growth rate for wall superheat= 8 C, contact angle = 38 deg, pressure= 1.013 105 Pa, and g / ge = 0.0001

111501-12 / Vol. 132, NOVEMBER 2010

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 03/27/2014 Terms of Use: http://asme.org/terms

t = 9.9 s
1.0E+04 Tsub = 0 C Tsub = 5oC o Tsub = 7.5 C
o

5.0E+03 q [w/m ]
2

0.0E+00

-5.0E+03

-1.0E+04 0 10 20 30 40 50 60 70 Angular Position, degrees 80 90

t = 76.0 s
1.0E+04
o Tsub = 0 C Tsub = 5oC Tsub = 7.5oC

5.0E+03 q [w/m2]

0.0E+00

-5.0E+03 0 10 20 30 40 50 60 70 Angular Position, degrees 80 90

the liquid. Next the numerical results for the cases where the gravity level is 104 of earth-normal gravity are presented. Here, the computational domain is chosen to be R / l0 , Y / l0 = 0.32, 0.64 for the subcooled boiling cases. Figure 20 shows that for water at saturation temperature, a bubble is predicted to detach from the wall at a size of 205 mm for a wall superheat of 8 C, a contact angle of 38 deg, and a system pressure of 1.013 105 Pa. In this study, considering the initial thermal boundary thickness and the eventual bubble size, different initial bubble sizes 50 mm for saturated and 8 mm for subcooled case and computational domains 250 mm radius for saturated and 80 mm for subcooled case are used for saturated and subcooled cases. Under subcooled boiling conditions, bubble-growth history similar to that noted earlier for ge and 0.01ge is observed and bubble continues to remain attached to the heater. Figure 21 shows the heat ux along the bubble interface at two different times. For both t = 9.9 s and t = 76.0 s, the effect of the liquid subcooling is evident. Also consistent with earlier results for 0.01ge, the rate of heat transfer is even smaller. Figure 22 gives the ow eld and isotherms for 5 C subcooling. In comparison to earth-normal gravity, a blunt plume appears above the bubble beyond 23.7 s. Upon close inspection, it is observed that the bubble size does not remain absolutely constant and the velocity vectors close to the interface indicate that the bubble oscillates in both horizontal and vertical directions. It is also noted that with time, the plume and wiggle in isotherms near the base are enhanced. It is expected that with reduced viscous forces, the inertia forces and convection terms play a relatively more important role, in view of the increased Re. As a result, the isotherms near the wall begin to appear in a more distorted pattern.

Conclusions

Fig. 21 Heat ux as a function of location along bubble interface for wall superheat= 8 C, contact angle= 38 deg, pressure = 1.013 105 Pa, and g / ge = 0.0001

volume growth rate, is 2.8 W, whereas it is 3.8 W for the saturated case. Therefore, this ratio Qmax,Tsub=1C / Qmax,sat is 0.74. In contrast, at g / ge = 1.0, Qmax,Tsub=1C / Qmax,sat is 0.85. This observation conrms that the subcooling affects the rate of heat transfer under microgravity more than it does under earth-normal gravity. The reason for this is that interfacial area for the heat transfer is larger in microgravity. Figure 16 presents the interfacial heat ux along the bubble interface at t = 0.7 and 9.9 s. The heat transfer from microlayer is not included once again. The trend is similar to that shown in Fig. 9 for g / ge = 1.0. However, it must be noted that evaporation and condensation heat uxes under microgravity conditions are much smaller in magnitude than those for earth-normal gravity. This difference is caused by a thicker thermal layer around the bubble over a longer time scale. A thicker thermal boundary layer results in a smaller temperature gradient, which in turn reduces the heat ux. Figures 17 and 18 show the ow eld and isotherms for liquid subcoolings of 0 C and 1 C, respectively, at g / ge = 0.01. Comparing the temperature elds for the two cases, it can be observed that a vortex is caused by condensate owing down on the liquid side for 1 C subcooling. At the early stage such as t = 0.01 s in Fig. 19 for 5 C subcooling, the larger velocity vectors indicate the rapid bubble growth. Subsequently, condensation begins and the area over which condensation occurs increases rapidly. Finally when the condensation rate is larger than the evaporation rate, the bubble begins to shrink. The condensate ows downward along the bubble interface resulting in the counterclockwise vortex in Journal of Heat Transfer

A numerical procedure, coupling the level-set function with the moving mesh method, has been employed to simulate subcooled nucleate boiling under various gravity levels. The effect of subcooling on bubble size is more pronounced under microgravity than under earth-normal gravity. At g / ge = 1.0, a vortex is generated at higher subcoolings in the liquid side resulting in a saddle point in the temperature prole, but not at low subcooling. Under microgravity conditions, the saddle point in temperature prole develops even for low liquid subcoolings. For a given subcooling, the rate of heat transfer around the bubble decreases with reduction in gravity. This is due to stretching of length and time scales.

Acknowledgment
This work received support from NASA under the Microgravity Physics Program Glenn Research Center, Contract No. NNX09AN65G.

Nomenclature
A cp g g H h hev h fg k l0 M m Nu p q R dispersion constant specic heat at constant pressure gravity vector gravity step function grid spacing for the macro region evaporative heat transfer coefcient latent heat of evaporation thermal conductivity characteristic length molecular weight mass ux vector Nusselt number, ql0 / klTw pressure heat ux radius of computational domain or bubble NOVEMBER 2010, Vol. 132 / 111501-13

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 03/27/2014 Terms of Use: http://asme.org/terms

t = 0.016s
0.64 0.64

t = 9.9s

0.48

0.48

y / l0

0.32

y / l0

0.32

0.16 95.13 C 0 -0.32 -0.16 0 r / l0 0.16 0.32

0.16 95.13oC 0 -0.32 -0.16 0 r / l0 0.16 0.32

t = 51.2s
0.64 0.64

t = 106.1s

0.48

0.48

y / l0

0.32

y / l0

0.32

0.16

95.13oC -0.16 0 r / l0 0.16 0.32

0.16

95.13 C

0 -0.32

0 -0.32

-0.16

0 r / l0

0.16

0.32

Fig. 22 Temperature distribution and velocity eld for wall superheat= 8 C, liquid subcooling= 5 C, contact angle= 38 deg, pressure= 1.013 105 Pa, and g / ge = 0.0001 temperature increment between isotherms is 2.57 C

radius of dry region beneath a bubble radial location of the interface at y = h / 2 universal gas constant Reynolds number, lu0l0 / l radial coordinate temperature time characteristic time temperature difference, Tw Tl velocity vector, u , v interfacial velocity vector characteristic velocity rate of vapor volume production from the V micro microlayer Vmicro vapor-side control volume Y height of computational domain y vertical coordinate Greek

Ro R1 R Re r T t t0 T u uint u0

Subscripts

0 c d e int l sat sub v w

initial condition critical departure earth-normal interface liquid saturation subcooling vapor wall partial differentiation with respect to partial differentiation with respect to

Superscripts T transpose

References
1 Gunther, F. G., and Kreith, F., 1954, Photographic Study of Bubble Formation in Heat Transfer to Subcooled Water, California Institute of Technology Progress Report No. 4-120. 2 Ramanujapu, N., 1999, Study of Growth Rate, Departure Frequency and Shape of a Single Bubble During Saturated and Subcooled Nuclear Boiling, Ph.D. prospectus, University of California, Los Angeles, Los Angeles, CA. 3 Ramanujapu, N., and Dhir, V. K., 1999, Dynamics of Contact Angle During Growth and Detachment of a Vapor Bubble at a Single Nucleation Site, Proceedings of the Fifth ASME/JSME Joint Thermal Engineering Conference, San Diego, CA. 4 Singh, S., 1999, Effect of Gravity and Liquid Subcooling on Bubble Dynamics, MS thesis, University of California, Los Angeles, Los Angeles, CA. 5 Siegel, R., and Keshock, E. G., 1965, Effects of Reduced Gravity on Nucleate Boiling Bubble Dynamics in Saturated Water, AIChE J., 104, pp. 509517. 6 Straub, J., Zell, M., and Vogel, B., 1990, Pool Boiling in Reduced Gravity Field, Proceedings of the Ninth International Heat Transfer Conference, pp. 91112. 7 Oka, T., Abe, Y., Tanaka, K., Mori, Y. H., and Nagashima, A., 1992, Observational Study of Pool Under Microgravity, JSME Int. J., Ser. II, 352, pp. 280286. 8 Qiu, D., and Dhir, V. K., 2002, Single-Bubble Dynamics During Pool Under

T o t

thermal diffusivity coefcient of thermal expansion liquid lm thickness nonevaporating liquid lm thickness thermal layer thickness curvilinear coordinate dimensionless temperature interfacial curvature dynamic viscosity kinematic viscosity curvilinear coordinate density surface tension level-set function contact angle

111501-14 / Vol. 132, NOVEMBER 2010

Transactions of the ASME

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 03/27/2014 Terms of Use: http://asme.org/terms

Low Gravity Conditions, J. Thermophys. Heat Transfer, 163, pp. 336345. 9 Son, G., Dhir, V. K., and Ramanujapu, N., 1999, Dynamics and Heat Transfer Associated With a Single Bubble During Nucleate Boiling on a Horizontal Surface, ASME J. Heat Transfer, 121, pp. 623631. 10 Stephan, P., Fuchs, T., Wagner, E., and Schweizer, N., 2009, Transient Local Heat Fluxes During the Entire Vapor Bubble Life Time,Proceedings of the ECI International Conference on Boiling Heat Transfer, Florianpolis, SC, Brazil. 11 Lay, J. H., and Dhir, V. K., 1995, Shape of a Vapor Stem During Nucleate Boiling of Saturated Liquids, ASME J. Heat Transfer, 117, pp. 394401. 12 Wayner, P. C., Jr., 1992, Evaporation and Stress in the Contact Line Region, Proceedings of the Engineering Foundation Conference on Pool and External Flow Boiling, Santa Barbara, CA, pp. 251256.

13 Wu, J., Dhir, V. K., and Qian, J., 2007, Numerical Simulation of Subcooled Nucleate Boiling by Coupling Level-Set Method With Moving Mesh Method, Numer. Heat Transfer, Part B, 51, pp. 535563. 14 Kays, W. M., and Crawford, M. E., 1980, Convective Heat and Mass Transfer, McGraw-Hill, New York, p. 328. 15 Ryskin, G., and Leal, L. G., 1984, Numerical Solution of Free-Boundary Problems in Fluid Mechanics, Part I, J. Fluid Mech., 148, pp. 117. 16 Son, G., 2003, Efcient Implementation of a Coupled Level-Set and Volumeof-Fluid Method for Three-Dimensional Incompressible Two-Phase Flows, Numer. Heat Transfer, Part B, 43, pp. 549565. 17 Son, G., 2001, A Numerical Method for Bubble Motion With Phase Change, Numer. Heat Transfer, Part B, 39, pp. 509523.

Journal of Heat Transfer

NOVEMBER 2010, Vol. 132 / 111501-15

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 03/27/2014 Terms of Use: http://asme.org/terms