X.

Li
Research Assistant
Department of Mechanical Engineering,
Clemson University,
Modeling of Heat Transfer in a
Clemson, SC 29634-0921
e-mail: [email protected] Mist/Steam Impinging Jet
The addition of mist to a flow of steam or gas offers enhanced cooling for many appli-
J. L. Gaddis cations, including cooling of gas turbine blades. The enhancement mechanisms include
Professor
effects of mixing of mist with the gas phase and effects of evaporation of the droplets. An
Mem. ASME
Department of Mechanical Engineering,
impinging mist flow is attractive for study because the impact velocity is relatively high
Clemson University,
and predictable. Water droplets, less than 15 ␮m diameter and at concentrations below 10
Clemson, SC 29634-0921
percent, are considered. The heat transfer is assumed to be the superposition of three
e-mail: [email protected]
components: heat flow to the steam, heat flow to the dispersed mist, and heat flow to the
impinging droplets. The latter is modeled as heat flow to a spherical cap for a time
dependent on the droplet size, surface tension, impact velocity and surface temperature.
The model is used to interpret experimental results for steam invested with water mist in
T. Wang a confined slot jet. The model results follow the experimental data closely.
Professor
关DOI: 10.1115/1.1409262兴
Mem. ASME
Energy Conversion and Conservation Center,
Keywords: Augumentation, Droplet, Evaporation, Heat Transfer, Impingement,
University of New Orleans, Modeling
New Orleans, Louisiana 70148-2220
e-mail: [email protected]

Introduction velocity fluctuations. In the wall-jet region the effect was slight.
The Nusselt number was found to increase by a factor of 2.7 for
The addition of mist to a flow of steam or gas offers enhanced
mass flow ratios 共solid/gas兲 of 0.8.
cooling for many applications, including cooling of gas turbine Guo et al. 关2兴 studied the mist/steam flow and heat transfer in a
blades 关1–3兴. The mechanisms of heat transfer enhancement in- straight tube under highly superheated wall temperatures. It was
clude effects of mist momentum on the gas phase and effects of found that the heat transfer performance of steam could be sig-
evaporation of the droplets, both directly and via the gas. In- nificantly improved by adding mist into the main flow. An average
creased specific heat and lower bulk temperature are also typical enhancement of 100 percent with the highest local heat transfer
features of a mist flow. In a mist/steam jet impingement flow, the enhancement of 200 percent was achieved with less than 5 percent
interaction of the droplets and the target wall becomes pronounced mist. In an experimental study with a horizontal 180 deg tube
because of the relatively high impact velocity and well defined bend Guo et al. 关3兴 found both the outer wall and the inner wall of
because the velocity is relatively predictable. the test section exhibited a significant and similar heat transfer
While single-phase jet impingement cooling has been studied enhancement. The overall cooling enhancement of the mist/steam
extensively, few studies have been found on mist jet impingement. flow increased as the main steam flow increased, but decreased as
Goodyer and Waterston 关1兴 considered mist/air impingement for the wall heat flux increased.
turbine blade cooling at surface temperatures above 600°C. They To explore the mechanism of mist heat transfer, interaction of
suggested that the heat transfer was dominated by partial contact droplets with the wall has been studied extensively. Wachters
between the droplets and the target surface, during which the et al. 关7兴 considered the impact of droplets about 60 ␮m impact-
droplets vaporized at least partially. A vapor cushion and the elas- ing a heated surface in the range of 5 m/s. Impinging droplets
tic deformation of the droplets were responsible for rejecting the could only maintain the spheroidal state with relatively high sur-
droplets. Addition of 6 percent water was found to improve the face temperatures. The required temperature depended on thermal
stagnation point heat transfer by 100 percent, diminishing away properties and roughness of the surface as well as the Weber num-
from the stagnation point. Droplet size was found to have little ber of the droplets. In the spheroidal state very low rates of heat
effect for 30 ␮ m⬍d 32⬍200 ␮ m. flow were observed.
Takagi and Ogasawara 关4兴 studied mist/air heat and mass trans- To obtain fundamental information concerning the heat transfer
fer in a vertical rectangular tube heated on one side. They identi- processes in spray cooling, Pederson 关8兴 studied the dynamic be-
fied wet-type heat transfer at relatively low temperatures and post- havior and heat transfer characteristics of individual water drop-
dryout type at higher temperatures. In the wet region the heat lets impinging upon a heated surface. The droplet diameters
transfer coefficient increased with increased heat flux. In the post- ranged from 200 to 400 ␮m, and the approach velocities ranged
dryout region the heat transfer coefficient increased with droplet from 2 to 8 m/s. The wall temperature ranged from saturation
concentration and flow velocity and with decreased droplet size. temperature to 1000°C. Photographs of the impingement process
Mastanaiah and Ganic 关5兴 confirmed that the heat transfer coeffi- showed that even the small droplets studied broke up upon im-
cient decreased with increased wall temperature. pingement at moderate approach velocities. The heat transfer data
Yoshida et al. 关6兴 focused on the effect on turbulent structure showed that approach velocity was the dominant variable affect-
with a suspension of 50 ␮m glass beads. In the impinging jet ing droplet heat transfer and that surface temperature had little
region, the gas velocity was found to decrease due to the rebound effect on heat transfer in the non-wetting regime. The droplet
of beads, accompanied by an increase in the normal direction deformation and break-up behavior for droplets 200 ␮m in diam-
eter did not appear significantly different from that for larger drop-
Contributed by the Heat Transfer Division for publication in the JOURNAL OF
lets. He also found that, for any given parameters in the non-
HEAT TRANSFER. Manuscript received by the Heat Transfer Division June 7, 2000; wetting regime, a minimum velocity could exist below which the
revision received April 23, 2001. Associate Editor: V. P. Carey. droplets deformed consistently without break-up.

1086 Õ Vol. 123, DECEMBER 2001 Copyright © 2001 by ASME Transactions of the ASME

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Fig. 1 Schematic diagram of test section

Chandra and Avedisian 关9,10兴 presented photographs of heptane heat flux of Joulean heating in the wall divided by the wall to
droplets impacting a heated surface. The relatively large 共⬎1 mm兲 saturation temperature difference, the heat transfer coefficient of
droplets at We⫽43 showed sensitivity to the surface temperature. the second panel is produced. Panel three shows the enhancement,
At low temperature the droplets spread and evaporated while at defined as the ratio of heat transfer coefficients with and without
higher temperature nucleate boiling was evident. Above the mist at the same Reynolds number. The cooling effect is signifi-
Leidenfrost temperature the droplets rebounded without any evi-
dence of wetting.
Buyevich and Mankevich 关11,12兴 modeled the impacted par-
ticles as liquid discs separated by a vapor layer whose thickness is
that of the wall roughness. The liquid mass flux was assumed
small enough to prevent formation of a liquid film on the heated
surface. Based on the energy conservation of the droplet as well as
the flow and heat conduction of the vapor interlayer between the
droplet and wall, a critical impact velocity was identified to deter-
mine whether a droplet rebounds or is captured. Depending on
their approach velocity, the impinging droplets are either reflected
almost elastically or captured by the heated surface and com-
pletely vaporized within a sufficiently short time. They applied the
model to dilute mist impingement with reported agreement with
experiment.
Fujimoto and Hatta 关13兴 studied deformation and rebound of a
water droplet on a high-temperature wall. For Weber numbers of
10 to 60, they computed the distortions of the droplet as it flat-
tened, contracted, and rebounded. They used a simple heat trans-
fer model to confirm that surface tension dominates vapor produc-
tion in the rebounding process. Hatta et al. 关14兴 gave correlations
of contact time and contact area of the droplet with Weber num-
ber.
Li et al. 关15兴 presented an experimental study for 1.1 bar steam
invested with water mist in a confined slot jet. Figure 1 is a sche-
matic of the test article having a slot of width 7.5 mm located in
a flat injection plate. The jet impacted a target wall of length 250
mm spaced 22.5 mm from the injection plate. The flow section
had a width of 100 mm and Pyrex walls allowed vision of the
heated surface. The droplet velocity and size distribution was ob-
tained by a phase Doppler particle analyzer 共PDPA兲. The experi-
mental results are typified by Fig. 2. In the first panel the depres- Fig. 2 A typical heat transfer result of mistÕsteam jet impinge-
sion of temperature caused by mist is shown. Using the measured ment „q ⬙ Ä7.54 kWÕm2, ReÄ14000, and m l Õ m s ÄÈ1.5 percent…

Journal of Heat Transfer DECEMBER 2001, Vol. 123 Õ 1087

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 cant near the stagnation point and decreases to a negligible
amount at 6 jet widths downstream. Up to 200 percent heat trans-
fer enhancement at the stagnation point was achieved by injecting
only ⬃1.5 percent by mass of mist. Direct observation through the
Pyrex wall showed a dry heated surface in the experiment condi-
tions, though no observations were made capable of disclosing the
behavior of individual droplets in brief contact. The purpose of
this communication is to model the processes of the experiment
and trends with heat flux, mist concentration, and vapor velocity,
based on Li et al. 关15兴.

Basic Assumptions and Model

Fig. 3 Modeling of heat transfer from wall to droplet
In mist/steam jet impingement, the droplets will not only influ-
ence the flow and temperature fields of the steam but also may
interact directly with the target wall. In the experiment 关15兴, water
droplets, less than 15 ␮m diameter and at concentrations below 5 The commercial code, FLUENT 关17兴, a solver for the complete
percent, impacted a heated surface with wall superheat below Navier Stokes equations using finite volume schemes, has been
60°C at a velocity up to 12 m/s. To model the heat transfer of the used to predict the trajectory of droplets including the determina-
mist/steam impinging jet under these conditions, the following tion of the impact velocity onto the heated surface. Complete
assumptions and approximations are made in this study: details are included in Li 关18兴 and only salient features are in-
cluded here. The domain of Fig. 1 supplied with appropriate
• The wall is sufficiently heated to prevent accumulation of boundary conditions in the entrance and exit regions was subdi-
liquid. vided to yield grid-independent results. The turbulent flow was
• The interaction between droplets is ignored, since the average modeled in several ways, with the k-␧ model found to yield sub-
spacing between droplets is large. stantial agreement with the heat transfer results in single-phase
• Because the droplet is small, no breakup is considered. steam flow. This computational model was combined with the
• The droplet is at the saturation temperature before entering dispersed-phase option of the program wherein droplets seeded in
the thermal boundary layer. the entrance region of the flow were tracked and allowed both to
• The droplet has a less important effect on the velocity bound- affect the vapor flow and to evaporate in transit through the su-
ary layer than on the thermal boundary layer. perheated layer. The droplets in the flow react with the fluid ac-
Under these assumptions, the heat transfer of mist/steam jet cording to drag on a sphere at the slip velocity between the droplet
impingement is divided into three different parts: heat transfer and the fluid, usually very near the low velocity Stokes Flow
from the target wall to the steam flow, heat transfer from the target asymptote.
wall to droplets and heat transfer between the steam and droplets. Direct Contact Heat Transfer. According to Buyevich and
No radiative heat transfer is considered since the wall temperature Mankevich 关11兴 共B&M model兲, the droplet will depart from the
is not very high in the current study and it is estimated to be less wall if the impact velocity is below a critical velocity, and stick if
than 2 percent of the total heat transfer. above. The critical velocity given by the B&M model is only
Heat Transfer From the Target Wall to the Steam. Heat about 0.6 m/s for d⫽10 ␮ m, ⌬⫽0.5 ␮ m and T w ⫺T sat⫽30°C.
transfer due to the steam is modeled as heat convection of a This means that for the conditions of the current study most of the
single-phase steam flow. Because of the disturbance by droplets droplets will stick to the wall. According to the B&M model a
on the boundary layer, this portion is subject to modification of sticking droplet will stay on the wall until evaporated completely.
the heat transfer coefficient of steam-only jet impingement flow. A If most of the particles stick to the wall and evaporate completely
detailed analysis of this effect must involve the effect of droplets the enhancement of heat transfer will be much higher than ob-
on the flow field and the turbulence characteristics. The heat trans- served. Therefore the B&M model is found to be inadequate for
fer enhancement through the effect of droplets on the flow has this study.
been assumed to be of secondary importance. Experimental study The actual interaction between the droplets and wall is very
by Yoshida et al. 关6兴 found 170 percent enhancement by adding 80 complicated; it includes a continuous deformation of the droplet
percent by mass glass beads of diameter 50 ␮m to the airflow. and is affected by droplet size and surface conditions. In this
Considering the effect of the particles includes boundary layer study, the heat transfer from wall to droplet is modeled simply by
disturbance as well as other cooling effects, the enhancement of transient heat conduction to a spherical cap with a contact angle of
the single-phase heat transfer due to droplets on the flow is pro- 60 deg based on Gould 关19兴 and Neumann et al. 关20兴. The corre-
jected to be less than 4 percent with a mist mass ratio of 2 percent. sponding height and base diameter of the cap are 0.464 and 1.608
times the original droplet diameter, respectively. Figure 3 shows
Heat Transfer From the Target Wall to Droplets. Although the basic model ( ␦ ⫽0.464d). The configuration of the flattened
many studies have been conducted on the interaction of the drop- droplet is assumed fixed until conditions for rebound are estab-
let with the bounding wall, few of these studies can be used to lished.
model the heat transfer from the target wall to droplets in the Quasi-steady heat flow to a droplet has been considered by
present study because of the different ranges of droplet size and many authors including Sadhal and Martin 关21兴 and Sadhal and
flow parameters. Unlike spray cooling, where the droplet momen- Plesset 关22兴. Under some conditions exact solutions may be ob-
tum is supplied by a device, small mist droplets may not be able to tained. In the current work there is a need to include the transient
hit the wall because of the drag force in the present study. Based warming of the droplet, as brief contact is anticipated. Since it is
on trajectory analysis, it is believed that larger droplets will hit the difficult to obtain an analytical solution, this problem is solved
wall if the approach velocity is high enough. Though neglected in numerically by using FLUENT 关17兴, with a non-uniform grid 共r,
trajectory analysis herein, the droplets are subject to the lift y兲 of 50⫻50. Assuming a small fraction of the droplet evaporates
‘‘force’’ of Ganic and Rosenhow 关16兴 due to the momentum im- before rebounding, a fixed-geometry 共no allowance for the de-
balance of asymmetric evaporation. A droplet in a temperature crease of mass in evaporation兲 transient solution is sought with a
gradient near a heated wall is heated faster on the wall side. The uniform initial temperature of T sat , the cap surface maintained at
difference in evaporation rate results in a lifting effect estimated to T sat , and the base 共wall兲 suddenly raised to T w . Figure 4共a兲
be of minor consequence for the conditions of this study. shows the non-dimensional results for the total base heat flow Q

1088 Õ Vol. 123, DECEMBER 2001 Transactions of the ASME

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 form in response to the superheat until it reaches a sufficient pres-
sure to repel the droplet. We reason that the pressure must over-
come surface deformation 共expressed through ␴ /d兲 and supply an
exit velocity 共pressure expressed through We ␴ /d兲 proportional to
the entering velocity of impact. The temperature required to pro-
vide this pressure is that associated by the slope of the liquid-
vapor saturation curve, wherein the required pressure is translated
to a required superheat. Finally the superheat is linearly related to
the product of wall superheat and the residence time. Expressed
non-dimensionally, there results
␣tr
⫽
c dT
d 2 T w ⫺T sat d P
冏冋
sat
␴
d
共 1⫹We/8兲 . 册 (2)

Fig. 4 Heat transfer process between droplet and wall by di- Here the constant, c, which depends on the geometry selected for
rect conduction „Q is the heat conduction from the target wall the heat conduction model, is found by trial and error to be about
to the droplet…: „a… total wall heat; and „b… superheat of droplet.
4.8⫻10⫺2 to agree with the experiment. The effect of impact
velocity is not strong if the Weber number is small. This residence
time is actually an effective value because it simplifies the defor-
in terms of ␣ t/d 2 . During contact the droplet is superheated in the mation process of the droplet on the wall. The residence time of a
amount given in Fig. 4共b兲. The heat entering the base and not 10 ␮m droplet with a Weber number of 1 and a temperature dif-
residing in the droplet as superheat is conducted to the surface and ference of 30°C is 1.7 ␮s. This model is expected to fail at high
is evaporated. There is no reference to the heat of vaporization wall temperature where the residence time goes to zero. In this
because this quantity is not converted to a mass flow. The surface case, however, it is believed that the droplet will still contact the
of the liquid maintained at the saturation temperature implies that wall for at least the lower bound established by the deformation
the evaporative heat flux is included in the computation. For a process.
temperature difference (T w ⫺T sat) of 30°C, the heat conduction in
Heat Transfer Between the Droplet and Steam. Heat trans-
1.2 ␮s evaporates 5 percent of a 5 ␮m droplet. Because the frac-
fer between the droplets and steam can be modeled by considering
tion of droplet evaporated is small, the assumption of constant
droplets as a distributed heat sink. The droplets evaporate into the
domain size and shape yields a fast, yet reasonable result.
superheated steam inside the thermal boundary layer and act to
Residence Time on Target Surface. Once a droplet hits the quench the boundary layer. Based on the superposition concept
wall, whether it rebounds from the wall depends on the wall tem- the temperature of mist/steam flow is divided into two parts, T
perature and impact velocity. The heat conduction model above ⫽T 1 ⫹T 2 . T 1 (x,y) is the temperature of steam-only flow and
cannot give the essential condition for rebounding. To complete T 2 (y) is the temperature depression caused by the mist.
this model, the residence time of the droplet on the wall must be The two-dimensional energy equation with a distributed heat
determined. It is conceivable that the droplets may wet the surface sink is given as
and stick on the heated wall until a vapor layer forms from nucle-
⳵T ⳵T ⳵ 2T ⳵ 2T
ation at the base. Upon formation of this layer the droplet would ␳ c pu ⫹ ␳ c pv ⫽k s 2 ⫹k s 2 ⫺k s ␤ 2 共 T⫺T sat兲 . (3)
return to its spheroidal shape and depart. A concept in pool boiling ⳵x ⳵y ⳵x ⳵y
has a waiting time during which the region near the wall becomes The last term is a heat sink per unit volume to a distributed sur-
superheated to the point where nucleation becomes spontaneous. face at temperature T sat . The coefficient, ␤, is equal to
Based on nucleation in a small cavity on the heated surface, Mikic
(12c mist␳ s d 10 / ␳ 1 d 330) 0.5 and c mist is the mist concentration. k s ␤ 2 is
and Rohsenow 关23兴 studied the waiting time and provided the
following simple estimate: the hA of the droplets per unit volume with hd/k s ⫽2 and
k s ␤ 2 (T⫺T sat) is the heat sink per volume. hd/k s ⫽2 is chosen for
t w⫽
1
再 共 T w ⫺T sat兲 r c
␲ ␣ T w ⫺T sat共 1⫹2 ␴ / ␳ g H f g r c 兲 冎 2
. (1)
slip ReⰆ1 for most droplets in the current study.
The equation for T 1 can be written as
Here r c is the radius of the nucleation cavity. This equation ⳵T1 ⳵T1 ⳵T1 ⳵ 2T 1
gives a waiting time of about 11 ␮s with r c ⫽2 ␮ m and T w ␳ c pu ⫹ ␳ c pv ⫽k s 2 ⫹k s 2 . (4)
⳵x ⳵y ⳵x ⳵y
⫺T sat⫽30°C. The principal attractive feature of this concept is
that the waiting time decreases slightly as the wall temperature For the current study, the boundary conditions for T 1 include
increases. Because this waiting time depends strongly on the value ⳵ T 1 / ⳵ x⫽0 at x⫽0 and x⫽L/2 and for y
of r c that is difficult to determine, this model cannot be applied T 1 ⫽T w at y⫽0 (5a)
confidently for the present study. Besides, this model does not
account for the effects of the droplet size and impact velocity. T 1 ⫽T sat at y→⬁. (5b)
Although the impact velocity was considered, the scale of the
residence time given by Hatta et al. 关14兴 did not include any wall Solution for Eq. 共4兲 subject to 共5兲 together with flow descriptions
temperature effect. The reason may be that their experiment was will produce a result for pure steam. In this work no solution is
conducted at a very high wall temperature 共above the Leidenfrost presented; rather the result is known from experiment to produce
temperature兲. If the wall temperature is low, the free-slip boundary h 0 (x)⫽q ⬙ /(T w ⫺T sat). In 关15兴 the experimental result is shown to
condition used in their study cannot be used any more. This basis agree substantially with other investigations. In lieu of an analyti-
for time scale will give a constant cooling enhancement for all cal solution, the following near-wall temperature distribution is
wall temperatures, which is not the case from experiments. The assumed.
Hatta model is expected to be valid as the temperature rises; it T 1 ⫽ 共 T w ⫺T sat兲 e ⫺yh 0 /k s ⫹T sat , (6)
should from a lower bound for the contact time.
For our selected model it is assumed that the droplet will de- where h 0 is the heat transfer coefficient obtained from experimen-
form into the lens shape of Fig. 3 and remain on the wall momen- tal study. T w and h 0 depend on x.
tarily without wetting gaining superheat according to the transient Considering T 2 is a function of y only, the equation for T 2 can
process of heat conduction discussed already. A vapor layer will be simplified as

Journal of Heat Transfer DECEMBER 2001, Vol. 123 Õ 1089

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 d 2T 2 dT 2 Table 1 Experimental cases
ks ⫺ ␳ c pv ⫺k s ␤ 2 共 T 1 ⫹T 2 ⫺T sat兲 ⫽0. (7)
dy 2 dy
The boundary conditions for Eq. 共7兲 are
T 2 ⫽0 at y⫽0 and y→⬁. (8)
This equation is first solved without considering the second term
and the result for ␤ ⫽h 0 /k s can be given as
␤ 2 共 T w ⫺T sat兲 ⫺ ␤ y ⫺yh /k
T 2⫽ 共e ⫺e 0 s 兲 . (9)
␤ 2 ⫺ 共 h 0 /k s 兲 2
Therefore, the heat transfer augmentation due to mist, h 2 , defined
as ⫺k s (dT 2 /dy) 兩 y⫽0 /(T w ⫺T sat), can be given by
Table 2 Results of the model
h 2 共 ␤ k s /h 0 兲 2
⫽ . (10)
h 0 ␤ k s /h 0 ⫹1
Assume that h 0 ⫽100 W/m2K, c mist⫽2 percent and d 10 /d 330
⫽1010 m⫺2, a value of h 2 /h 0 ⫽0.069 is obtained. A value for c mist
of 10 percent will give a value of 0.269 for h 2 /h 0 .
The effect of the second term, ␳ s c p v dT 2 /dy can be evaluated
using Eq. 共9兲. The velocity, v , leaving the boundary layer is esti-
mated to be only about 0.25 mm/s by integrating the vapor gen-
erated from droplet evaporation. For the conditions of the ex-
ample, this results in a value of the second term about 1 percent of
the sink term and is neglected. The solution for T 2 given by Eq.
共9兲 is accepted as an approximation.
The liquid concentration near the target wall might be different case shown in Fig. 2. The predicted results are given in Table 2.
from the average concentration because the droplets cross the The input to the analytical model includes h 0 , T w , T sat , m l /m s ,
streamlines. However, migration of the droplets away from the as well as the droplet size distribution by PDPA. In Table 2, q 1⬙
wall occurs due to the lift forces and turbulent dispersion makes ⫽h 0 (T w ⫺T sat) is the single-phase heat transfer from wall to
the mist concentration more uniform and close to the average steam, q 2⬙ using Eq. 共10兲 is the quenching effect of the mist; and
value. Therefore, the quenching effect of the mist is estimated q 3⬙ is the direct heat conduction during the contact time of Eq. 共2兲
with the average concentration. Surveys by PDPA support this from wall to droplet. It can be seen that the predicted results and
assumption. the experimental data have good agreement, especially when con-
sidering the experimental uncertainty. The relative size of the vari-
Model Validation ous contributions is shown clearly in Table 2 and q ⬙exp/q⬙1 is the
The average heat transfer within x/b⬍1 is considered. As an heat transfer enhancement ratio, h mist /h 0 . The q 3⬙ component
example, a distribution of droplet size from the experimental dominates q 2⬙ . Both q 2⬙ and q 3⬙ become important in proportion to
study is given in Fig. 5共a兲 at the jet exit for Re⫽14,000 and mist concentration.
m l /m s ⫽1.5 percent. This size distribution, obtained by PDPA
measurement entering the test section, gives the average diameters
of d 10⫽4.7 ␮ m and d 30⫽6.4 ␮ m. By using FLUENT 关17兴, the Prediction of Parametric Effects
droplet distribution impacting the wall is given in Fig. 5共b兲. For The general aim of the prediction is to determine the heat trans-
the case cited, it is predicted that droplets less than 5 ␮m will not fer due to droplet injection, given wall temperature, Reynolds
impact the wall, which means there is no direct heat conduction number, liquid concentration and droplet distribution. Firstly, the
from the wall to droplets. The heat transfer to small droplets is analytical model discussed above requires the droplet size distri-
mainly through the steam. Though not shown there is divergence bution. Secondly, the impinging velocity and deposition rate on
of the pathlines resulting in diminished droplet flux at the stagna- the heated surface must be known. These can be evaluated respec-
tion point. The impact velocity varies with droplet size and injec- tively by empirical equations or obtained by numerical simulation.
tion location and for the case cited it ranges up to 12 m/s. Thirdly, determine the heat removal from the target wall directly
Table 1 lists five different cases to be predicted. Case 1 is the by the droplets. Lastly, add the heat transfer by the two other
components and obtain the total heat transfer.
The current analytical model can successfully predict the effect
of various parameters observed in the experiment. When the wall
temperature increases, the heat transfer from wall to steam and
from steam to droplets will increase proportionally with the tem-
perature difference. However, the heat transfer due to the direct
conduction from wall to droplets will change little 共compare cases
1 and 5兲 because the residence time becomes short. Therefore, the
ratio of heat transfer coefficients will decrease, which has been
observed in experimental studies. Figure 6 shows the predicted
result of the wall temperature effect, given the mist concentration
and impact velocity for 10 ␮m droplets and a single-phase heat
transfer coefficient of 150 W/m2-K. As shown in this figure, the
droplet impact velocity is an important variable affecting droplet
heat transfer, a trend in agreement with the experiment by Peder-
Fig. 5 Droplet distribution and number at jet exit and on target son 关9兴. Figure 7 shows the comparison of the model result and
wall: „a… at jet exit; and „b… impacting on target wall „ x Õ b Ë1…. the experimental data. Here the droplet size distribution measured

1090 Õ Vol. 123, DECEMBER 2001 Transactions of the ASME

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 depends on the jet velocity as well as the droplet size. Small
droplets always have a small impact velocity if they have enough
momentum to reach the wall.
Given wall temperature and mist concentration, if the jet veloc-
ity increases, the heat transfer from wall to steam will increase but
the heat transfer from steam to droplet will decrease due to the
thinner boundary layer 共see Eq. 共10兲兲. The heat transfer from wall
to droplet will increase as more droplets hit the wall at higher
impact velocity. As a result, the overall heat transfer enhancement
increases when the jet velocity increases. This tendency is verified
by experiment.

Conclusions
Fig. 6 Predicted effect of the wall temperature on mistÕsteam A model for mist/steam jet cooling has been developed and
heat transfer at different mist concentrations and droplet im- presented which considers the total heat flow to be comprised of
pact velocities
three components. A single-phase-like heat flow and a boundary
layer quenching effect account for heat flow leaving the surface
through the steam. To this is added a heat flow occurring in brief
contacts with impacting droplets.
Heat conduction from the wall to droplets is found to be the
dominant enhancement mechanism. The quenching effect of drop-
lets in the steam flow becomes important when the mist concen-
tration is high. The heat transfer to small droplets is mainly
through the steam while larger droplets hit and cool the heated
wall by direct heat conduction.
Because the enhancement increases at lower wall temperature,
the contact time for direct conduction varies inversely with wall
superheat. A contact time correlation is proposed which, with a
simple conduction model, accounts for the observed heat transfer
within the experimental uncertainty. The model depends on size
distribution, impact velocity and density for droplets, requiring a
dispersed-phase trajectory model.
All mechanisms of cooling are proportional to mist concentra-
tion. The effect of vapor velocity is mildly positive on the en-
hancement. The effect of droplet size has both positive and nega-
tive components and the model has implied predictions but these
are not known from experiment.
Fig. 7 Comparison of the predicted result by the model and
experimental data
Acknowledgments
The authors would like to thank Graver Separations 共Wilming-
from experiment is used; impact density and velocity have been ton, DE兲 for donating the steam filters for the experiment. We also
predicted by FLUENT’s dispersed flow feature. The agreement is want to thank Mee Industries Inc. 共El Monte, CA兲 for donating the
substantial. pressure atomizers and the high pressure pump. We appreciate the
Figure 8 shows the predicted trend of enhancement with mist help from Dr. T. Guo in setting up the test facility. This research
concentration, given the wall temperature, droplet size and impact was sponsored by the U.S. Department of Energy under the con-
velocity. The same single-phase heat transfer coefficient as in Fig. tract DOE/AGTSR 95-01-SR-034, and was managed by Dr. N.
6 is used. The enhancement of heat transfer is proportional to the Holcombe at the Federal Energy Technology Center and by Dr.
mist concentration. With a fixed value of impact velocity, smaller Larry Golan at the South Carolina Institute for Energy Studies.
droplets provide greater enhancement. However, this result cannot
be used simplistically because the impact velocity of a droplet Nomenclature
A ⫽ area 共m2兲
b ⫽ jet width 共7.5 mm兲
c ⫽ mass concentration
cp ⫽ specific heat capacity 共J/kg-K兲
d ⫽ diameter of droplet 共␮m兲
d 10 ⫽ arithmetic mean diameter 共␮m兲
d 30 ⫽ volume mean diameter 共␮m兲
d 32 ⫽ Sauter mean diameter 共␮m兲
Hfg ⫽ latent heat 共J/kg兲
h ⫽ heat transfer coefficient⫽q ⬙ /(T w ⫺T sat 共W/m2-K兲
h mist ⫽ heat transfer coefficient of mist 共W/m2-K兲
h0 ⫽ steam-alone heat transfer coefficient 共W/m2-K兲
k ⫽ heat conductivity 共W/m-K兲
m ⫽ mass flow rate 共kg/s兲
P ⫽ pressure 共N/m2兲
Fig. 8 Predicted effect of the mist concentration on mistÕ
Q ⫽ heat conduction⫽ 兰 0⬘ q ⬙ Adt(J)
steam heat transfer at different wall temperatures and droplet q⬙ ⫽ heat flux 共W/m2兲
diameter Re ⫽ Reynolds number ( ␳ s v j 2b/ ␮ s )

Journal of Heat Transfer DECEMBER 2001, Vol. 123 Õ 1091

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 r ⫽ coordinate in the radial direction 共m兲 Heat Transfer of A Two-Dimensional Impinging Jet with Gas-Solid Suspen-
T ⫽ temperature 共K兲 sions,’’ NHTC, 2, pp. 461– 467.
关7兴 Wachters, L. H. J., Smulders, L., Vermeulen, J. R., and Kleiweg, H. C., 1966,
t ⫽ time 共s兲 ‘‘The Heat Transfer from A Hot Wall to Impinging Mist Droplets in The
tr ⫽ residence time 共s兲 Spheroidal State,’’ Chem. Eng. Sci., 2, pp. 1231–1238.
y, v ⫽ velocity components in x, y directions 共m/s兲 关8兴 Pederson, C. O., 1970, ‘‘An Experimental Study of the Dynamic Behavior and
vj ⫽ jet velocity 共m/s兲 Heat Transfer Characteristics of Water Droplet Impinging upon a Heated Sur-
face,’’ Int. J. Heat Mass Transf., 13, pp. 369–381.
We ⫽ Weber number ( ␳ v 2 d/ ␴ ) 关9兴 Chandra, S., and Avedisian, C. T., 1992, ‘‘Observations of Droplet Impinge-
x ⫽ coordinate along the target wall 共m兲 ment on a Ceramic Porous Surface,’’ Int. J. Heat Mass Transf., 35, No. 10, pp.
y ⫽ coordinate perpendicular to the target wall 共m兲 2377–2388.
␣ ⫽ thermal diffusivity 共m2/s兲 关10兴 Chandra, S., and Avedisian, C. T., 1991, ‘‘On the Collision of a Droplet With
␤ ⫽ variable defined in Eq. 共3兲 a Solid Surface,’’ Proc. R. Soc. London, Ser. A, 432, pp. 13– 41.
关11兴 Buyevich, Yu. A., and Mankevich, V. N., 1995, ‘‘Interaction of Dilute Mist
⌬ ⫽ thickness of vapor layer 共m兲 Flow with a Hot Body,’’ Int. J. Heat Mass Transf., 38, pp. 731–744.
␦ ⫽ height of spherical cap 共m兲 关12兴 Buyevich, Yu. A., and Mankevich, V. N., 1996, ‘‘Cooling of a Superheated
␮ ⫽ dynamic viscosity 共kg/m-s兲 Surface with a Jet Mist Flow,’’ Int. J. Heat Mass Transf., 39, pp. 2353–2362.
␳ ⫽ density 共kg/m3兲 关13兴 Fujimoto, H., and Hatta, N., 1996, ‘‘Deformation and Rebounding Processes
␴ ⫽ surface tension 共N/m兲 of a Water Droplet Impinging on a Flat Surface above Leidenfrost Tempera-
ture,’’ ASME J. Fluids Eng., 118, pp. 142–149.
Subscripts 关14兴 Hatta, N., Fujimoto, H., Kinoshita, K., and Takuda, H., 1997, ‘‘Experimental
Study of Deformation Mechanism of a Water Droplet Impinging on Hot Me-
l ⫽ liquid phase tallic Surfaces above Leidenfrost Temperature,’’ ASME J. Fluids Eng., 119, pp.
s ⫽ steam 692–199.
sat ⫽ saturated 关15兴 Li, X., Gaddis, J. L., and Wang, T., 2001, ‘‘Mist/Steam Heat Transfer in Con-
w ⫽ wall fined Slot Jet Impingement,’’ ASME J. Turbomach., 123, No. 1, pp. 161–167.
关16兴 Ganic, E. N., and Rosenhow, W. M., 1977, ‘‘Dispersed Flow Heat Transfer,’’
Int. J. Heat Mass Transf., 20, pp. 885– 866.
References 关17兴 Fluent, 1997, FLUENT 4.4 User’s Guide, I–III, Fluent Inc.
关18兴 Li, X., 1999, ‘‘Cooling by a Mist/Steam Jet,’’ Ph.D. dissertation, Dept. of
关1兴 Goodyer, M. J., and Waterston, R. M., 1973, ‘‘Mist-Cooled Turbines,’’ Conf. of
Mechanical Engineering, Clemson University, SC.
Heat and Fluid Flow in Steam and Gas Turbine Plant, Proc. of Institution of
Mechanical Engineers, pp. 166 –174. 关19兴 Gould, R. F., 1964, Contact Angle, Wettability and Adhesion, American Chem-
关2兴 Guo, T., Wang, T., and Gaddis, J. L., 2000, ‘‘Mist/Steam Cooling in a Heated istry Society, DC.
Horizontal Tube Part I: Experimental System and Part II: Results and Model- 关20兴 Neumann, A. W., and Spelt, J. K., 1996, Applied Surface Thermodynamics,
ing,’’ ASME J. Turbomach., 122, pp. 360–374. Marcel Dekker, Inc., NY.
关3兴 Guo, T., Wang, T., and Gaddis, J. L., 2000, ‘‘Mist/Steam Cooling in a 180- 关21兴 Sadhal, S. S., and Martin, W. W., 1977, ‘‘Heat Transfer Through Drop Con-
Degree Tube,’’ ASME J. Heat Transfer, 122, No. 4, pp. 749–756. densate Using Differential Inequalities,’’ Int. J. Heat Mass Transf., 20, No. 12,
关4兴 Takagi, T., and Ogasawara, M., 1974, ‘‘Some Characteristics of Heat and Mass pp. 1401–1407.
Transfer in Binary Mist Flow,’’ Proc. of 5th Int. Heat Transfer Conf., Tokyo, 关22兴 Sadhal, S. S., and Plesset, M. S., 1979, ‘‘Effect of Solid Properties and Contact
Japan Society of Mechanical Engineers, pp. 350–354. Angle in Dropwise Condensation and Evaporation,’’ ASME J. Heat Transfer,
关5兴 Mastanaiah, K., and Ganic, E. N., 1981, ‘‘Heat Transfer in Two-Component 101, No. 1, pp. 48 –54.
Dispersed Flow,’’ ASME J. Heat Transfer, 103, pp. 300–306. 关23兴 Mikic, B. B., and Rohsenow, W. M., 1969, ‘‘Bubble Growth Rates in Non-
关6兴 Yoshida, H., Suenaga, K., and Echigo, R., 1988, ‘‘Turbulence Structure and Uniform Temperature Field,’’ Prog. Heat Mass Transfer, 2, pp. 283–292.

1092 Õ Vol. 123, DECEMBER 2001 Transactions of the ASME

Downloaded From: http://heattransfer.asmedigitalcollection.asme.org/ on 01/29/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use