Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
To cite this article: H. M. Al-Ahmadi & S. C. Yao (2008): SPRAY COOLING OF HIGH TEMPERATURE METALS USING HIGH MASS
FLUX INDUSTRIAL NOZZLES, Experimental Heat Transfer: A Journal of Thermal Energy Generation, Transport, Storage, and
Conversion, 21:1, 38-54
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to
anyone is expressly forbidden.
The publisher does not give any warranty express or implied or make any representation that the contents
will be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses should
be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims,
proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly in
connection with or arising out of the use of this material.
Experimental Heat Transfer, 21:38–54, 2008
Copyright © Taylor & Francis Group, LLC
ISSN: 0891-6152 print/1521-0480 online
DOI: 10.1080/08916150701647827
Systematic experiments were conducted for the spray cooling of high-temperature stainless
steel using three different types of industrial sprays: full cone and flat hydraulic nozzles
and a flat air-mist nozzle. In the present study, a wide range of mass flux (1.5–30 kg/m2 sec)
is covered, which has never been thoroughly investigated before. Orientations with respect
to gravity and spray angle were also explored.
The data have been analyzed and correlated. The results of this study will be impor-
tant to the product quality control in industries by providing a good estimation of heat flux
at different mass flux, spray types, and surface temperatures, especially for spray cooling
of stainless steel and other metals with similar thermal properties.
Keywords spray cooling, high temperature cooling, spray heat transfer, cooling hot metals,
nozzles
INTRODUCTION
Water spray cooling, which removes heat from a hot subject by spraying water
droplets onto the surfaces, has been used extensively in numerous types of applications,
such as metallurgy, microelectronics, nuclear safety, and aerospace engineering. Water
sprays are used very effectively for heat extraction from solidification in continuous
casting processes. Due to the very different thermal requirements of different processes,
a large number of spray nozzles, including both hydraulic and air-mist, are used under
a variety of conditions. Hydraulic spray nozzles can be classified into several different
categories based on their spray pattern, such as hollow cone, full cone, and flat sprays.
Air-mist spray nozzles, which generally produce droplets of smaller size and higher
velocity, can be classified based on design configuration as internal or external mixing,
and both are capable of producing conical and flat patterns.
No data were found about heat transfer from flat air-mist spray nozzles in the open
literature and the only study about the hydraulic flat nozzle was for low spray-mass
fluxes. Relevant previous investigations can be divided into two categories: individual
droplet cooling and spray cooling of hot surfaces. Various experimental studies have
been conducted for individual droplet cooling. Dependence of droplet dynamics and heat
transfer on droplet Weber number, surface temperature, liquid subcooled temperature,
38
SPRAY COOLING OF METALS AT HIGH TEMPERATURE 39
NOMENCLATURE
G mass density (kg/m2 s) qmin-r the radiative component of heat flux at
Ts hot surface temperature the Leidenfrost point
Tled the Leidenfrost temperature qmin-c the convective component of heat flux
T1 cooling fluid temperature at the Leidenfrost point
T temperature (Kı ) t time (sec.)
q heat flux (kW/m2 ) " emissivity, D 0.97 in this case
qmin total heat flux at the Leidenfrost Stefan-Boltzmann constant 5.670*10 8
point (W/m2 K4 )
Downloaded by [Texas State University - San Marcos] at 19:31 28 April 2013
and surface properties have been discussed in detail. While the reported information is
valuable for a basic understanding, the results describe only single-droplet heat transfer
rate at very low liquid-mass flux.
On the other hand, some researchers have studied the laboratory-produced, mono-
dispersed spray cooling of hot surfaces, which is next in complexity to a single droplet
impinging on a hot surface. Yao and others [2, 3] studied heat transfer of mono-size
sprays for spray density in the range G D 0:3–2 kg/m2 sec. Maximum heat flux of 2
106 W/m2 was reported at the surface temperature of a copper test piece in the range
of 140ı C–160ıC and the Leidenfrost temperature around 250ıC. Air influence was also
discussed in that study. Furthermore, in [1], results of low mass-flux mono-size sprays
with very large drop size were obtained. A parametric comparative study was carried out
to correlate his work with previous studies. Correlations for a wide range of spray heat
transfer are presented.
Most of the industrial spray nozzles are poly-dispersed, with a wide range of droplet
sizes and velocities. This, added to the complex droplet-surface interactions due to large
number of impacting drops, makes the performance of industrial spray difficult to predict
based on the single drop or mono-dispersed spray results.
There are limited heat transfer studies in the open literature about industrial spray
nozzles [4–10]. Most of the authors used commercial hydraulic full-cone nozzles or flat
nozzles at a lower mass flux. Other types of nozzles, such as an air-atomized fan, were
not covered, and for most of them, the mass fluxes are in the low range. For any given
nozzle, the measured heat transfer rate is a function of wall and fluid temperatures, local
spray mass-flux, droplet velocity and size, and nature and finishing of the cooled surface.
Correlating the results of different sprays creates wide differences among these results.
Gaugler [4] studied a copper rod under an industrial full cone nozzle spray. The
following relation for heat transfer rate in the film-boiling region was suggested:
at
250 < Tsurf < 450ı C, and 0:7 < G < 3:7 kg/m2 s:
In two separate studies by Hoogendoorn [5] and Mizikar [6] for hydraulic full cone
nozzles, mass flux was up to 20 kg/m2s. Both used a stainless steel test sample. Horizontal
and vertical downward sprays were included in the first study, while two angles of attack
but only vertical downward spray were covered in the second. It was concluded that
40 H. M. AL-AHMADI AND S. C. YAO
the heat transfer coefficient is a linear function of water mass flux. Furthermore, the
horizontal spray has a lower maximum heat flux, and the angle of spray has a negligible
effect on heat transfer rate.
Extensive review of water spray cooling up to the late 1970s has been provided by
Bolle and Moreau [7]. In the same article, the authors reported spray cooling for stainless
steel using three hydraulic nozzles for mass flux below 7 kg/m2 s. The following corre-
lation was suggested for heat transfer coefficient in film boiling for surface temperature
900 K < Tsurf < 1200 K:
at
1 < G < 7(kg/m2 s/:
The effects of droplet velocity and induced transverse airflow were also discussed.
In a series of articles [8–9], experimental studies of both copper and aluminum test
pieces, which have high thermal conductivities, were reported. The spray densities varied
from 0.6–9.9 kg/m2 s. Results and correlations for heat flux and temperatures at both
wetting and non wetting regimes were presented.
The objectives of the present study are to:
Figure 1. Experimental setup. 1—Unitrust frame, 2—Furnace, 3—Rail, 4—Cooling box, 5—Nozzle, 6—Test
plate, 7—Flexible pipe, 8—Pressure gauge, 9—Flowmeter, 10—Filter, 11—To Computer, 12—Water tank,
13—Valves, 14—Pump.
with 11 tubes arranged in line. The tips of the tubes are tapered to prevent droplet
bouncing, and the collected water of each tube goes to individual collection bottles.
Before any heat transfer test, the pump is switched on and the regulator valves are
adjusted for the desired water pressure, and the air pressure is adjusted if the air-mist
nozzle is used. These valve positions are saved then, only another on/off valve is used
later on. The test plate is heated in the furnace to a temperature of 900ıC–1,100ıC.
Then the plate is moved from the furnace to an exact location under the spray and
the on/off valve is opened, this starts the spray cooling and the data are recorded. The
data are sampled at a rate of 10 or 20 readings per second until the plate temperatures
fall below 100ıC. The total sampling durations were three to five minutes.
A total of 38 tests, classified into 8 groups, have been performed, as shown in Table
2. Two orientations were tested for the first full cone nozzle: vertical down spray with
horizontal test disk position, and horizontal spray with vertical test disk position. For
center tests, the sprays impact the plate at a 90ı angle. For side tests, a 50.8 mm (2 in)
spray-center to disk-center lateral distance was maintained, resulting in a nominal 72ı
spray angle of impact. All air-mist nozzle tests were performed at an air pressure of
310 kPa (45 psi) except for two cases, where the air pressures of 207 and 550 kPa are
applied to reveal the effect of air pressure on heat transfer.
Data Reduction
In the data reduction, data smoothing was applied. Averaged results were then used
for heat transfer calculations. Two-step heat transfer solutions are needed to obtain heat
flux and temperature at the test plate surface. As shown in Figure 2, direct heat conduction
takes place between T1 and T2 thermocouples locations. The direct conduction solution
gives the heat flux at plane T1 . Iterative implicit method was used to solve for this
Downloaded by [Texas State University - San Marcos] at 19:31 28 April 2013
temperature field. The inversed conduction problem was solved with the Weber method
[11] in the region between plan T1 and the plate surface where spray impinges. The linear
transient conduction is reformulated as:
@2 T @T @2 T
˛ D C
; (3)
@x 2 @t @t 2
where
is a small non negative constant, equals 0.01, to make Eq. (3) as a perturbation of
the conventional diffusive conduction equation. In inversed heat conduction, the surface
heat flux at time t depends on the interior temperatures at times both before and after t.
For this reason, the numerical solution was obtained for all time steps at a given spatial
node before any temperature values were computed at the next spatial node.
.n/ .n/ .n/
.Ti C1 2Ti C Ti 1/
D
.x/2
" .nC1/ .n 1/
#
Ti Ti
.nC1/ .n/ .n 1/
.1=˛/ C .T 2Ti C Ti / (4)
2 T t 2 i
Figure 3. Local spray density along the A-A axis for 1st nozzle.
Figure 4. Local spray density along the center B-B axis for the 2nd nozzle.
however, the normalized local distribution pattern does not change. At a fixed location,
the mass flux could vary significantly when the nozzle pressure changes. Uncertainty of
the mass flux measurements are estimated as within 10%.
Boiling Curves
The spray boiling heat transfer curve is similar to that of pool boiling heat transfer
and has the distinctly boiling regimes. There is a significant increase in the film boiling
heat transfer of the spray as compared to pool boiling because when the forced impinging
SPRAY COOLING OF METALS AT HIGH TEMPERATURE 45
Downloaded by [Texas State University - San Marcos] at 19:31 28 April 2013
Figure 5. Surface heat flux versus surface temperature for four different tests.
flow pushes the droplets closer to the hot surface and thereby increases the heat flux in
the film boiling regime significantly. However, this is not the case for the critical heat
flux.
Figure 5 presents the surface flux versus surface temperature for four different
tests. Table 3 shows these test conditions. At low mass flux, in tests #15 and #45,
the heat flux curve experiences a minimum heat flux at the Leidenfrost point. This
is not the case for higher mass flux, tests #253 and #306. Different explanations are
possible for this phenomenon. One might be because the duration of the cooling time
is shorter for high mass flux that there is not enough time for heat flux to develop
into a typical boiling curve. High heat flux causes the temperature of the plate to drop
quickly. Another explanation, reported in the literature [10], is that at high mass flux, the
cooling curve does not experience a minimum point due to the absence of stable film
boiling. Instead, the cooling curve goes through a gradual increase, followed by a rapid
increase at the Leidenfrost point. It is believed that at high mass flux the boiling curve
is difficult to obtain through the transient cooling of the test plate that has low thermal
conductivity. Metal with higher thermal conductivity may achieve better results. In all
Spray density
# Nozzle G (kg/m2 s) Orientation Location
the reported studies, the spray heat flux from copper, which has a much higher thermal
conductivity, is always much higher than the heat flux from steel under similar spray
conditions.
Film Boiling
First of all, free cooling, i.e., room cooling by natural convection and radiation
without spray, was conducted to find plate surface emissivity. The calculations show that
emissivity is around unity. The surface is polished after each test to maintain the similar
surface condition and emissivity. After subtracting radiation component of the surface
Downloaded by [Texas State University - San Marcos] at 19:31 28 April 2013
heat flux, it was found that the heat flux due to water spray impaction increases linearly
with plate surface temperature.
Heat flux at film boiling is correlated with the minimum heat flux and surface
temperature. The radiative and convective heat fluxes are assumed to be separable. The
data show a linear dependency of convective boiling heat flux on temperature difference
ratio as follows:
Ts Tled
q.Ts ; qmin / D qmin-c 1 C 0:5 C qrad .Ts / (5)
Tled T1
where
" D 0:97
the plate causes sub-cooled transition boiling to start at a higher surface temperature.
The Leidenfrost temperature is found to be lower for horizontal spray tests than for
vertical downward sprays.
Critical Temperature
Figure 8 shows the temperature at the maximum heat flux for all tests. Significantly
less effects are observed from the angle of attack (side tests) and orientation, and there
is a general trend for the critical temperature to increase with increasing mass flux.
48 H. M. AL-AHMADI AND S. C. YAO
Downloaded by [Texas State University - San Marcos] at 19:31 28 April 2013
Correlation Equations
Through the experiments, errors can be induced from measurements and data reduc-
tion. Uncertainties can also be revealed from repeating tests. Among the various errors
of measurements, the dominant error comes from the uncertainty of the thermocouple
locations with respect to the surface where the spray impinges. The effect of the un-
certainty on thermocouple locations is amplified through the inversed conduction data
reduction. This uncertainty contributes about 10% error of the results. The numerical
50 H. M. AL-AHMADI AND S. C. YAO
Downloaded by [Texas State University - San Marcos] at 19:31 28 April 2013
Figure 10. Surface heat flux at surface temperature of 200ı C versus local spray density.
Figure 12. Heat flux at the Leidenfrost point from results of previous studies of stainless steel test pieces.
Figure 13. Maximum heat flux from results of previous studies of stainless steel test pieces.
CONCLUSION
1. In general, the Leidenfrost temperature, minimum film boiling heat flux, critical
temperature, and maximum heat flux are strongly dependent on local spray mass
flux. Other spray parameters such as droplet size, velocity, and the type of the
nozzle play a less significant role.
2. Correlation for the Leidenfrost temperature, minimum heat flux, critical tem-
perature, maximum heat flux, and heat flux at 200ıC surface temperature with
local spray mass flux were obtained. Clear trends are observed for most of these
parameters with spray mass flux.
SPRAY COOLING OF METALS AT HIGH TEMPERATURE 53
Downloaded by [Texas State University - San Marcos] at 19:31 28 April 2013
Figure 14. The Leidenfrost temperature from results of previous studies of stainless steel test pieces.
3. At low spray mass fluxes, the minimum heat flux of side tests have a higher value
than center tests. It is not likely due to the angle of impingement effect, but rather
from the effects of sub-cooled liquid film overflow from the center of the plate.
The results of Mizikar’s study [6] support this conclusion.
4. Spray film boiling heat transfer was found to increase linearly with surface
temperature, in addition to the radiation heat transfer. Correlations are presented.
5. Heat flux showed weak dependence on spray density for the wetting region of
the cooling curve.
6. By changing air pressure for the third nozzle (air-mist flat) while keeping water
pressure constant at 413.6 kPa (60 psi), it was demonstrated that the air pressure
effect is not as strong as water mass flux. Yao [2, 3] has concluded that too.
7. The results of this study are in substantial agreement with results from the previous
studies of stainless steel by Hoogendoorn [5], Mizikar [6], and Bolle and Moureau
[7] for the partially overlapped range of spray density.
8. It is believed that at higher spray density, the effect of droplet size or orientation
on heat transfer is minimal. The droplet velocity, on the other hand, may have
important effect.
ACKNOWLEDGMENT
The financial and technical supports of Ken Kasperski and Kristy Tanner at Spraying
Systems Co. are greatly appreciated.
REFERENCES
1. T. Cox, Heat Transfer Experiments and General Correlation for Sprays of Very large Droplets
Impinging on a Heated Surface, Doctoral thesis, Carnegie Mellon University, 1998.
2. K. J. Choi and S. C. Yao, Heat Transfer Mechanisms of Horizontally Impacting Spays, Int. J.
Heat Mass Transf., vol. 30, no. 2, pp. 1291–1296, 1987.
54 H. M. AL-AHMADI AND S. C. YAO
3. S. Deb and S. C. Yao, Analysis on Film Boiling Heat Transfer of Impacting Sprays, Int. J.
Heat and Mass Transf., vol. 32, no. 11, pp. 2099–2112, 1989.
4. R. E. Gaugler, Experimental Investigation of spray Cooling of High Temperatutre Surfaces,
Doctoral thesis, Carnegie Mellon University, 1966.
5. C. J. Hoogendoorn and R. D. Hord, Leidenfrost Temperature and Heat Transfer Coefficients
for Water Sprays Impinging on a Hot Surface, Proceedings of the Fifth Int. Heat Transf. Conf.,
vol. 4, pp. 135–138, 1974.
6. E. Mizikar, Spray Cooling Investigation for Continuous Casting of Billets and Blooms, Iron
and Steel Engineer, pp. 53–70, June, 1970.
7. L. Bolle and J. C. Moureau, Spray Cooling of Hot Surfaces, Multiphase Science and Technol-
ogy, Hemisphere, 1982.
8. I. Mudawar and W. S. Valentine, Determination of Local Quench Curve for Spray–cooled
Downloaded by [Texas State University - San Marcos] at 19:31 28 April 2013