The
boiling of highly subcooled liquid flow (mainly water) is actively and widely
used in the cooling systems of apparatuses and installations that require
removal of extremely high heat flux densities (q). The record
q
values recorded in [1] under conditions of high mass velocities, subcooling,
and pressures are higher than 200 MW/m2. An additional
attractiveness of this cooling method is given by the fact that the boiling
process itself is concentrated in a thin surface layer. As a result, at q far
from critical, the hydraulic resistance of the flow differs slightly from that
in the case of single-phase flow [2].
A
large number of works have been devoted to studying the boiling of a flow of a
strongly subcooled liquid, but their main goals were to study the integral characteristics
of the process (critical heat flux density and heat transfer coefficients).
Less attention has been paid to the statistical characteristics of vapor
bubbles (distribution over diameters, evolution of sizes and shapes, average
lifetimes, etc.) [3]. However, for the development of the theory and models of
the process, it is quite important to have a volume of reliable experimental
data on such characteristics. In classical works [4, 5], along with modern [6],
the main method for obtaining statistical data was video filming of bubble
evolution in the direction perpendicular to the boiling surface. This method
did not allow tracking the evolution of the bubble shape, information about
which can be obtained by visualizing the bubble in profile.
In [7–9],
the characteristics of vapor bubbles were studied using video filming in
profile during boiling of subcooled water. Heating of the boiling surface of a
large area (compared to the size of the bubble) was carried out by electric
current. The works [7, 8] were devoted to the study of separation diameters of
bubbles. In [9], the evolution of the shape and size of vapor bubbles was
studied. The studies described in these articles were carried out in the range
of low
q
corresponding to the regime of "isolated" bubbles
during boiling of subcooled water. In experiments [7, 8], the values of the
heat flux density were up to 200 kW/m2, and in [9] up to 500 kW/m2.
When going to higher
q, it is impossible to trace the evolution of an
individual vapor bubble with the help of video recording in profile due to the
electrical heating of a large area surface due to the shading of each other by
the bubbles. This problem can be solved by using localized heating of a surface
of a small area (comparable to the size of a bubble) to obtain and study a
single bubble. In the present work, surface heating was localized using a laser
beam. Water with a high degree of subcooling Δ
tsub
>
40°C was used as a heat carrier. The experiments were carried out in the range
of heat flux densities corresponding to the developed bubble boiling regime (q
= 1.0–1.6 MW/m2).
To
obtain detailed information about the change in the size and shape of bubbles
at different stages of its evolution (growth, constant size, collapse), a high
speed of video recording is required. For example, the duration of the growth
stage to the maximum size is usually 100–150 µs. In order for at least 5 frames
to fit into this period of time, a shooting speed with a frequency of at least
50 kHz (20 μs between frames) is required. The average bubble lifetime in
the studied range of parameters was 500–900 μs. Consequently, each bubble
at a shooting frequency of 50 kHz was recorded for at least 25 frames. To
obtain statistics, information about the evolution of several bubbles is
needed, which requires the processing of a large number of frames. In order to
automate the processing of video frames of single bubbles, a program was
written in the PyCharm integrated development environment with the Python 3.10
interpreter, the processing results of which showed high efficiency. The
developed program will allow processing large arrays of experimental data at
acceptable time and labor costs.
The
experimental setup (fig. 1) consisted of a closed circuit with water, the
temperature of which was varied using two heat exchangers (heater and cooler).
The flow rate and liquid temperature were measured by a turbine flow meter and
resistance thermometers, respectively.
The
length of the test section of the setup was 70 mm, its cross section was made
rectangular 21 x 5 mm. Glass windows were installed on two sides and front
walls of the test section to enable video recording of the boiling process in
different directions. The heating plate, the thickness of which was 0.1 mm, was
made of Kh18N9T steel. The plate was mounted on a single wall of the test
section that did not have a glass viewing window. Its diameter was 13 mm. Laser
radiation was used to heat the outer surface of the plate.
To
obtain a focused laser beam on an area 1–2 mm in diameter (a Jenoptik laser
diode with a power of up to 100 W was used as its source), a system of two
lenses was used. The first lens converted the beams emerging from the laser
fiber into a parallel beam, and the second lens (position
4
in fig. 1)
focused it on the outer surface of the heating plate. To increase the degree of
absorption of laser radiation, a thin layer of a special aerosol based on
graphite was deposited on the surface of the plate. At the corresponding
wavelength of the laser diode used (808 ± 3 nm), the emissivity of such a
coating was measured, which was ~0.9.
Using
a high-speed video camera (position
2
in fig. 1), the evolution of
single vapor bubbles on the heating plate was visualized through the side
windows. The video recording frequency was 50 kHz with an exposure time of up
to 3 μs. To measure the temperature of the outer surface of the plate, a
thermal imaging camera (position
3
in fig. 1) was used, from which the
temperature of the boiling surface was then recalculated.
Fig. 1. General
view of the
experimental setup,
measuring and
auxiliary equipment. The numbers indicate:
1 – test section;
2 – camera for high-speed video recording;
3 – thermal imaging camera;
4 – lens for focusing laser radiation;
5 – backlights.
To
automate the processing of frames obtained as a result of high-speed video recording,
a program was written in the PyCharm integrated development environment with a
Python 3.10 interpreter using a specialized library of computer vision and
image processing algorithms OpenCV.
The
source data was a video in .avi format obtained as a result of the experiment,
which was cut into frames using the first module of the program. Frames were
saved in a special directory. Then, using the second module of the program,
written on the basis of the J. Canny algorithm [10], a black-and-white version
of each frame was created, where the boundaries of objects were highlighted
with white contours. Fig. 2 shows frames from a video recording of a single
bubble, and fig. 3 their version after processing.
The
black-and-white version of the frames obtained this way made it possible to
accurately determine the size of the bubbles. The bubble diameter is indicated
as
d
in fig. 2a and fig. 3a. To determine the size of the bubbles using
the obtained black-and-white versions of the frames, the algorithm of the third
module of the program was used, in which a loop was run over the matrix of
pixels and their maximum and minimum values were found along the ordinate axis.
By subtracting the minimum value from the maximum value, the bubble diameter was
determined in pixels, which was then recalculated in µm using the known frame
scale.
After
obtaining the size of the bubble on each frame, the program allowed plotting
the change in diameter over time. Fig. 4 shows a comparison of graphs of
changes in the size of the bubble (which video frames are shown in fig. 2)
versus time, obtained when processing each frame «manually» and using the
developed program. Points are highlighted in red on the graphs corresponding to
the sizes of bubbles on the corresponding frames before (fig. 2) and after
processing in the program (fig. 3). According to fig. 4, it can be seen that
the processing of frames, carried out «manually» and with the help of the
program, showed a fairly high coincidence of the results.
Fig. 2. Frames of video recording of the
evolution of a single vapor bubble:
Δtsub = 42
°C;
ρw = 0
kg/(m2·s);
q = 1.4
MW/m2.
Time from
bubble initiation: (a) – 60
μs;
(b) – 200
μs;
(c)
– 340 µs; (d) – 460
μs.
Frame size 0.8 x 1.0 mm. The exposure time is 3 µs.
Fig. 3. Frames of
the evolution of a single vapor bubble after processing in the developed
program:
Δ
tsub
= 42 °C;
ρw = 0
kg/(m2·s);
q
= 1.4
MW/m2.
Time from
bubble initiation: (a) – 60
μs;
(b) – 200
μs;
(c)
– 340 µs; (d) – 460
μs.
Frame size 0.8 x 1.0 mm.
(a)
(b)
Fig. 4. Graph of
the change in the diameter of a single vapor bubble over time
(Δtsub = 42 °C;
ρw = 0
kg/(m2·s);
q = 1.4
MW/m2):
(a) –
obtained by «manually» processing; (b) – obtained by processing in the
developed program.
Using
the program developed in the PyCharm integrated development environment with
the Python 3.10 interpreter, a method for automated processing of the results
of high-speed video recording of the evolution of a single vapor bubble during
boiling of a subcooled liquid is implemented. The developed method
significantly reduces the time and labor costs for processing the experimental
material and makes it possible to include a significantly larger array of
experimental data in the analysis.
This
work was supported by the Ministry of Science and Higher Education of the
Russian Federation (State Assignment ¹ 075-01129-23-00).
1.
Mudawar I., Bowers M. B. Ultra-high critical heat flux (CHF) for
subcooled water flow boiling — I: CHF data and parametric effects for small
diameter tubes // Intern. J. Heat and Mass Transfer. 1999. V. 42. ¹ 8. P.
1405–1428.
2. Zeigarnik Yu.A., Kirillova
I.V., Klimov A.I., Smirnova E.G. Some measurements of hydraulic resistance in
boiling heated to bellow the saturation temperature // High Temperature. 1983.
V. 21. ¹ 2. P. 234–239.
3.
Vasil’ev
N.V., Zeigarnik Yu.A., Khodakov K.A. Boiling in Forced Convection of Subcooled
Liquid as a Method for Removing High Heat Fluxes (Review): Part 1.
Characteristics, Mechanism, and Model of the Process, Heat Transfer, and
Hydraulic Resistance // Thermal Engineering.
2022. V. 69. ¹ 4. P. 235–251.
4.
Gunther F.C. Photographic study of surface boiling heat transfer
to water with forced convection // Transactions of the ASME, J. Heat Transfer. 1951.
V. 73.
Ð. 115–123.
5. Treshchev G.G. Experimental
study of the mechanism of heat transfer in the surface boiling of water //
Thermal Engineering. 1957. ¹ 5. P. 44–48.
6. Vasil’ev N.V., Zeigarnik Yu.A.,
Khodakov K.A., Maslakova I.V. Boiling of liquid subcooled to saturation temperature
in channels as a method for removal of ultimate heat fluxes // Thermal Engineering.
2019. V. 66. P. 350–360.
7. Sugrue R., Buongiorno J.,
McKrell T. An experimental study of bubble departure diameter in subcooled flow
boiling including the effects of orientation angle, subcooling, mass flux, heat
flux, and pressure // Nuclear Engineering and Design. 2014. V. 279. P. 182–188.
8. Xu J.J., Chen B.D., Xie
T.Z. Experimental and theoretical analysis of bubble departure behavior in
narrow rectangular channel // Progress in Nuclear Energy. 2014. V. 77. P. 1–10.
9. Cao Y., Kawara Z., Yokomine
T., Kunugi T. Visualization study on bubble dynamical behavior in subcooled
flow boiling under various subcooling degree and flowrates // Intern. J. Heat
and Mass Transfer. 2016. V. 93. P. 839–852.
10. Canny J. F. A computational
approach to edge detection // IEEE Trans. Pattern Analysis and Machine
Intelligence. 1986. V. 8. P. 678–698.
