by user

Category: Documents





Nuclear Engineering and Design 240 (2010) 2215–2224
Contents lists available at ScienceDirect
Nuclear Engineering and Design
journal homepage: www.elsevier.com/locate/nucengdes
PIV measurements of turbulent jet and pool mixing produced by a steam jet
discharge in a subcooled water pool
Yeon Jun Choo, Chul-Hwa Song ∗
Korea Atomic Energy Research Institute, Daeduk-daero 1045, Yuseong-gu, Daejeon 305-353, Republic of Korea
a r t i c l e
i n f o
Article history:
Received 5 February 2009
Received in revised form
14 September 2009
Accepted 15 September 2009
a b s t r a c t
This experimental research is on the fluid-dynamic features produced by a steam injection into a subcooled water pool. The relevant phenomena could often be encountered in water cooled nuclear power
plants. Two major topics, a turbulent jet and the internal circulation produced by a steam injection, were
investigated separately using a particle image velocimetry (PIV) as a non-intrusive optical measurement
technique. Physical domains of both experiments have a two-dimensional axi-symmetric geometry of
which the boundary and initial conditions can be readily and well defined. The turbulent jet experiments
with the upward discharging configuration provide the parametric values for quantitatively describing a
turbulent jet such as the self-similar velocity profile, central velocity decay, spreading rate, etc. And in the
internal circulation experiments with the downward discharging configuration, typical flow patterns in
a whole pool region are measured in detail, which reveals both the local and macroscopic characteristics
of the mixing behavior in a pool. This quantitative data on the condensing jet-induced mixing behavior
in a pool could be utilized as benchmarking for a CFD simulation of relevant phenomena.
© 2009 Elsevier B.V. All rights reserved.
1. Introduction
Direct Contact Condensation (DCC) between steam and water
may be regarded as an inevitably occurring phenomenon in water
cooled nuclear power plants. In the design of the APR 1400, an
in-containment refueling water storage tank (IRWST) is installed
to play the role of a primary heat sink and the collector of the
steam discharged from the reactor coolant system (RCS) during a
feed-and-bleed operation and a rapid depressurization of the RCS
(Song et al., 2007). In this discharge process, a heat and mass transfer through the DCC between steam and subcooled water occurs
very effectively. Different scales of experiments to investigate the
thermo-hydraulic characteristics in this DCC process have been
conducted (e.g., Weimer et al., 1973; Kim et al., 1997; Cho et al.,
1998; Song et al., 1998, 2007). These studies, however, have mostly
focused on steam jet shapes, condensation regimes and a condensational heat transfer. In addition, most of these works have been
interested in the internal characteristics of a steam jet.
General characteristics of free shear single phase jets are
explained well in many textbooks (e.g., Abramovich, 1963; Pope,
2000) and other open literatures. However, information on the
steam jet-induced turbulent jet and the resultant internal circulation in a confined pool cannot be found in previous researches.
∗ Corresponding author at: Thermal Hydraulics Safety Research Division.
Tel.: +82 42 868 8876; fax: +82 42 868 8362.
E-mail address: [email protected] (C.-H. Song).
0029-5493/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
The turbulent jet produced by a steam injection could be one of
the important topics because the effectiveness of a thermal mixing in a pool and the safety concerns relevant to the IRWST design
are closely related to the overall features of this steam-induced
turbulent jet. Kim and Youn (2008) measured the velocity distribution of a condensing jet discharged by a single hole sparger using a
point-wise Pitot tube and compared the measured velocity profiles
with theoretical models. Even though their results have provided
information about the mixing characteristics of a steam injection,
their usefulness is rather limited due to adopting a local measuring
technique. Recently Van Wissen et al. (2005) presented the characteristics of a turbulent jet discharged from a ring-shaped orifice
using PIV measurements and presented detailed information on the
jet. However, their results are only valid for their specific type of
There are few efforts to simulate the overall mixing phenomena caused by a large steam discharge in a pool using a CFD tool
such as Gamble et al. (2001) and Kang and Song (2008). In spite
of the convenience and usefulness of the CFD analysis approach, it
is strongly recommended to validate the computed results against
reliable experimental data. However, there is a lack of proper experimental data for a steam jet-induced turbulent jet and the resultant
mixing phenomena in a pool.
Systematic efforts via experimental and numerical studies on
steam jet-induced pool mixing phenomena were recently made
by KAERI’s thermal-hydraulic research group (Song et al., 2007).
Local measurements of the pool temperature distribution in a
large quench tank were compared with the CFD analysis results
Y.J. Choo, C.-H. Song / Nuclear Engineering and Design 240 (2010) 2215–2224
Fig. 1. JICO experimental facility.
which were determined based on the so-called condensation region
model, which provides the inlet boundary conditions to the whole
calculation domain, as a simplified analytical model to describe a
steam jet region. This local information was insufficient to be compared directly between experimental and numerical results and,
consequently, more precise information on the steam jet-induced
turbulent jet and the resultant global pool mixing is needed to
reasonably describe the macroscopic behavior in a pool.
Based on the above background, the quantitative and qualitative experimental data adequate to validate the results of a
numerical simulation of the relevant problems are presented in
this study. Numerical simulation, which directly compare with
these experimental results, is on-going in parallel with this study.
This paper, dealing with only experimental part, is divided into
two kinds of topics: the steam jet-induced turbulent jet and the
overall mixing produced by a steam jet discharge in a pool, and
Fig. 2. Steam nozzle (all unit is in millimeter).
Y.J. Choo, C.-H. Song / Nuclear Engineering and Design 240 (2010) 2215–2224
Table 1
Test matrix for the turbulent jet and pool mixing experiments.
Test I.D.
Mass flux (kg/m2 s)
Steam temperature (◦ C)
Nozzle pressure (kPa)
Pool temperature (◦ C)
Turbulent jet
Pool mixing
provides a set of benchmarking data which will be useful in
validating the results of a numerical simulation of the relevant
2. Experimental facility
The experimental facility, JICO (Jet Injection and Condensation),
used in this work is shown schematically in Fig. 1. The JICO was
designed so as to realize various configurations of the jets and
pool mixing phenomena, such as a condensing steam jet, single
or two-phase jet, plunging jet and a pool mixing driven by these
jets. The transparent side wall, made of acrylic material, is optically
accessible so that non-intrusive optical measurement techniques
can be easily applicable. Moreover, the capability of having an
alternative choice of an upward or downward injection of a jet
allows us to perform a variety of experimental simulations in this
The JICO facility consists of a test section and an electric boiler
to supply saturated steam. The test section consists of two open
tanks, an inner cylinder (an inner diameter of 0.78 m and a height
of 2 m) as a mixing pool of our interest and an outer square tank
(a square cross section of 1 m × 1 m and a height of 1.8 m), and a
replaceable injection nozzle. Two overflow lines (Ø = 50 mm) are
installed to maintain a constant water level in the inner cylinder at a height of 1825 mm. The outer square tank is installed to
eliminate an optical distortion and also to minimize the heat loss
through the inner cylinder wall. Three thermocouples are installed
inside the inner cylinder to measure the pool water temperature.
Electric boiler consists of a feed water storage tank with a capacity of about 0.5 m3 and an immersion electrical heater of 12 kW,
and it can provide a maximum 0.023 kg/s of saturated steam under
an operating pressure of a maximum 10 bar. Before the water
in the storage tank is fed into the boiler, it is pre-heated for a
In the case of the turbulent jet experiment, the steam discharging nozzle is installed at the bottom of the inner cylinder for a
vertical upward injection of the steam. On the other hand, in the
pool mixing experiment, the nozzle is installed at the upper part of
the pool with a downward-facing injection direction. The detailed
dimensions and shape of the nozzle are shown in Fig. 2. In order to
avoid a complexity in the simulation of a discharging nozzle, the
steam nozzle has a single hole with a sharp edge and a straight
flow channel. In order to minimize the pressure drop and to ensure
a developed velocity profile at the nozzle exit, a gradually converged shape of a flow channel is adopted in the nozzle. In addition,
the thermal insulation tube enveloped on the outer surface of the
nozzle makes it possible to minimize the heat loss though the submerged part of the nozzle wall. Steam temperature and pressure to
be discharged are measured at an upstream part of the discharging
nozzle exit.
3. Experimental methods
3.1. Turbulent jet experiment
From a practical viewpoint of validating a CFD analysis tool
against rather complicated thermal-hydraulic phenomena, it is recommended to choose a simple domain of interest. The present
subject, therefore, is on the assumption of a two-dimensional
axi-symmetric flow spatially and a steady-state flow temporally.
Fig. 3a schematically shows the geometrical conditions of the target domain in the case of the turbulent jet experiment. The nozzle
is installed vertically at the bottom plate of the tank so that the jet
axis is in-line with the axi-symmetric line of the target domain. For
the sake of convenience, the wall condition can be assumed to be
adiabatic due to the adoption of the water-filled outer tank; the
Fig. 3. Geometrical description of the experiment (turbulent jet and pool mixing).
Y.J. Choo, C.-H. Song / Nuclear Engineering and Design 240 (2010) 2215–2224
Fig. 4. Schematic representation of the PIV measurement: (a) turbulent jet experiment, (b) pool mixing experiment.
heat loss through both the wall and free surface of the pool was
actually estimated to be negligible during the test.
Free surface in the pool is constantly maintained to be a
1825 mm height by an overflow hole. The radius of the inner tank
from the jet axis to the wall boundary is 390 mm, i.e. the ratio of the
inner cylinder diameter to the nozzle exit diameter (D/d) is 156, so
that we can assume it to be an unconfined jetting condition. The
initial velocity profile at the nozzle exit can be regarded as fully
developed by ensuring the nozzle length is L/d = 20. Since the supplied steam mass flux is small compared with the total amount of
the pool water and the disturbance by the flow through an overflow hole at the free surface is weak, so the influence of the free
surface movement on the nature of the turbulent jet itself could
be neglected. The test matrix for the turbulent jet experiments is
shown in Table 1.
3.2. Pool mixing experiment
During the process of a blowdown from the RCS, steam or highenergy liquid is discharged into the IRWST pool and consequently a
strong thermal-hydraulic mixing process is generated in the pool. In
a real pool, the radial injection through multi-holed spargers makes
the flow characteristics asymmetric and transient in general; it
means that it is difficult to simulate or measure a real situation in a
practical manner. Moreover, even for some possible cases, the real
complex situation would increase the uncertainties of the measurements and numerical simulation. For these reasons, pool mixing
experiments in this work were carried out in simple flow configurations just like in the turbulent jet experiments. Differences of
the pool mixing experiments from the turbulent jet experiments
lie only in the jet injection direction and the level of the free surface. A vertical-downward injection is more advantageous when
compared with an upward one for inducing a strong internal circulating flow in a pool just like in an impinging jet case. The level
of free surface is 850 mm in height from the bottom plate and the
steam discharge nozzle is immersed at about 60 mm below the free
surface, as shown in Fig. 3b. Visual observation shows that the free
surface is quiescent in this condition. The test matrix for the pool
mixing experiments is also shown in Table 1.
4. Measurement techniques and the uncertainties
4.1. PIV technique
In the present work, the PIV technique is used to measure the
velocity field of a turbulent jet and a pool mixing driven by a
steam injection. In recent years, the PIV measurement technique
has shown very promising results in fluid flow researches and has
been used very extensively for velocity field measurements in particular due to its non-intrusive capability. A typical PIV system
consisting of two functions, i.e., image capture and image analysis,
was used. In the image capture system, the light source is a doublehead Nd–YAG laser (Continuum) operated at a frequency of 10 Hz
and a power of 200 mJ per pulse at = 532 nm. MegaPlus ES1.0
PIV camera (1018 × 1008 pixels) operating in a triggered double
exposure mode is coupled with 60 mm Nikon microlens. With this
system, two paired-instantaneous particle images are stored in a
synchronized PIV processor (PIV 2100, Dantec Inc.) and transferred
to a PC. Small non-condensable bubbles (typically 10–100 ␮m in
size) could be incidentally used as tracer particles. In the condensation process of a steam jet, the production of non-condensable
gas bubbles is almost unavoidable due to a possible existence of
dissolved gas in the feed water in the boiler and also due to the
interaction of the pool water with air at the free surface. Nevertheless, it was thought that these bubbles are small enough to follow a
jet flow with a high momentum. However, since local zones with a
small momentum may exist in a pool where even the small bubbles
cannot follow the liquid motion, this non-condensable gas bubble could not be used as a tracer in the pool mixing experiment.
Therefore, fluorescent solid tracer particles (Dantec Inc. FPP-RhB35, 20–50 ␮m) are dispersed into the pool in the case of the pool
mixing experiments, unlike the turbulent jet experiment. The signals scattered from the fluorescent particles can be imaged through
a narrow band-pass filter attached to a lens.
Y.J. Choo, C.-H. Song / Nuclear Engineering and Design 240 (2010) 2215–2224
Fig. 5. Typical PIV images: (a) turbulent jet experiment, (b) pool mixing experiment.
In the process of the image analysis, we have used the Adoptive
Cross-Correlation (ACC) algorithm, which would be rather useful in the case of measuring the velocity for the free shear layer,
where a larger velocity gradient inside an observation area may
exist and possibly introduce an erroneous calculation and/or underestimation. For this reason, two steps of an ACC are adopted in this
work. The initial interrogation widow is 64 × 64 pixels and then the
refinement process is conducted at 32 × 32 pixels over the 2 steps.
The measuring strategies of both experiments are schematically shown in Fig. 4. The measuring field-of-view in the PIV
camera for the turbulent jet and pool mixing experiments are
100 mm × 100 mm and 400 mm × 400 mm, respectively. In the turbulent jet experiment, one camera scans the target area along the
axial direction of the jet from the nozzle exit to a far-field region
(z/d ≈ 0–100) over 6 intervals. In the pool mixing experiment, two
horizontally mounted cameras simultaneously capture the particles image over the whole test section.
In the region of a steam jet and the dense non-condensable gases
at the near field of the nozzle exit (over 70–90 mm from the nozzle
exit), i.e., at the lower right part of region-1 in Fig. 4a and the upper
left part in Fig. 4b, a measurement of the velocity is difficult or its
result would be unreliable even in the possible cases. Fig. 5 shows
some typical images captured in both experiments. As shown in the
image of region-1, a separate particle’s signal cannot be observed.
Along the downstream, the moderate density of the bubble scattering allows for an efficient vector calculation. In the case of the pool
mixing experiment, two cameras cover the entire internal region
of the test section as shown in Fig. 5b.
4.2. Error analysis
Uncertainties of the velocity measurements by PIV and the other
experimental parameters are presented below by estimating them
according to the ISO GUM Guide.
4.2.1. Temperature
Three K-type thermocouples (WATLOW Inc.) are employed for
the pool temperature at three vertical places and one K-type thermocouple for the steam temperature at the nozzle exit. The type-B
uncertainties of each error source are estimated as follows: 0.8 ◦ C
for the K-type thermocouple; 1.1 ◦ C for the compensation cable;
1.0 ◦ C for the digital multi-meter. The overall uncertainty of the
temperature measurement was estimated to be ±1.910 ◦ C at a 95%
confidence level (k-factor 1.96).
Y.J. Choo, C.-H. Song / Nuclear Engineering and Design 240 (2010) 2215–2224
Fig. 6. Typical instantaneous and average velocity vectors and contours of their magnitudes.
4.2.2. Mass flux
Steam mass flux injected by a nozzle is measured by a Vortex type flowmeter (OVAL, EX DELTA) which has an accuracy of
a ±1% reading value. Including the uncertainties of the digital
multi-meter (0.0035% of a reading and 0.0005% of a span) and
the voltage converter (0.1% of reading), the overall uncertainty of
the steam mass flux is ±3.196 kg/m2 s at a 95% confidence level
(k-factor 1.96).
4.2.3. Pressure
Uncertainty of the nozzle exit pressure measured by the pressure senor (Rosemount 3051S, 0.15% of span) was estimated to be
±3.127 kPa at a 95% confidence level (k-factor 1.96).
4.2.4. Velocity
The bias error of the instantaneous velocity vector measured
by the PIV technique is comprised of two major error sources due
to the timing interval and the peak displacement detected during
the correlation calculation. Typically, a FFT-based correlation algorithm has an accuracy of about 0.25–0.1 pixels, whereas the time
interval error can be considered to be negligible. Another source
of an uncertainty that can occur in the experiment is one induced
by distorted images which result from a local temperature gradient, that is, a non-uniform refractive index. Such an undesirable
effect, however, is not so dominant in our experiment. Steam temperature decreases rapidly due to a thermal mixing with a highly
subcooled water. Actually, except for the region downstream of
the steam discharge nozzle, the maximum temperature difference
is only within a few degrees, so that this effect on the uncertainty
could be insignificant. Consequently the uncertainty of the instantaneous velocity measurement is conservatively estimated to be
1% of the absolute value with a consideration of the other uncertainty sources such as the scaling factor between the image and
real dimensions, an optical distortion of the observation window,
etc. Finally, if an additional random error by averaging the instantaneous velocities (300–800 samples) is considered together with
two sigma (2), the total uncertainty of the velocity measurement could be estimated approximately as 1.7–2.5% of an absolute
5. Results and discussion
The present experimental study is focused on understanding
the hydraulic characteristics of a turbulent jet and the pool mixing
driven by a steam injection. The results of these two topics will be
discussed hereunder.
5.1. Turbulent jet driven by a condensing steam jet
Analytical and experimental approaches for the mean quantities of free shear flows have been established comparatively well
as far as a single phase jet, i.e., air-to-air and liquid-to-liquid is considered. It is evident that further researches on a steam jet-driven
turbulent jet with a condensation process, as in this research unlike
a single phase jet, is required. Therefore, it is expected that the
results described in this section could be used in many relevant
researches, which include the validation of a CFD simulation.
5.1.1. Instantaneous and mean velocity field
The typical contours of the instantaneous velocity vectors and
their magnitude between z/d = 36 and 55, i.e., in region-2 in Fig. 4a
are presented in Fig. 6a. The inherent oscillating manner of the jet
is clearly shown, reflecting the feature of an entrainment of the
jet and its interaction with the surrounding water. Evidently as
well, the velocity in the central region is high. On the other hand,
the mean velocity (Fig. 6b) shows clear profiles. In practice, some
physical assumptions for a free shear layer (jet, wake and simple
shear layer) allow for the turbulent flow to be transformed into a
mathematically analyzable form. Also the theoretical description
of the mean flow characteristics of the single phase turbulent jet
may be applicable to a steam jet-driven turbulent jet. Consequently,
understanding of the characteristics of the mean flow in a turbulent jet could be very valuable for describing a turbulent jet. So our
experimental results will be analyzed in this manner.
Fig. 7 shows the entire vector field which was measured and
integrated by means of scanning the target region over 6 intervals.
This is, in appearance, similar to the case of a single phase turbulent
jet. Actual measured region is up to z/d = 100 in all the cases of the
turbulent jet experiments. It is clearly observed that the velocities
at the center lines are decayed due to a turbulent dissipation and
Y.J. Choo, C.-H. Song / Nuclear Engineering and Design 240 (2010) 2215–2224
Fig. 7. Overall velocity pattern: (a) M300T60, (b) M650T60.
the velocity profiles are spreading to the extent of the rate at which
the shear layer grows laterally with the axial distance.
5.1.2. Self-similarity
The concept of a self-similarity, or self-preservation, implies that
the flow has reached a dynamic equilibrium or asymptotic state
in which the mean and higher-order moments evolve together
(Townsend, 1976). Typically, self-similar quantities are represented by some scaled dependent variables and show a universal
aspect over a fully developed region. Two general characteristic
scales are the mean centerline maximum velocity, UC (z) = U(z,0),
and a jet’s half width (r1/2 ) in a round shear jet. In addition, since
a jet’s half width is dependent on (z − z0 ), a dimensionless scale,
r/(z − z0 ), can be an alternate cross-stream scale. Consequently the
cross-stream similarity variable can be taken to be either
z − z0
along the dimensionless scale, . Generally, a Gaussian profile can
be fitted accurately with experimental data:
= exp(−K2 )
where K is a constant to represent the shape of a mean velocity
profile and it is determined from experimental data. The profile of
U/UC in the case of G = 450 kg/m2 s and Tp = 60 ◦ C is typically shown
in Fig. 9. The coincidence of all the measurements onto a single
fitting curve reflects that the jet has a good self-similarity feature.
where z0 is a jet’s virtual origin. Measurements of single-phase jets
indicate that in the far-field the centerline velocity UC is inversely
proportional to the axial distance:
UC =
z − z0
with a coefficient C determined experimentally. The development
of the mean center-line velocity as a function of the distance from
a nozzle is shown in Fig. 8. From these results, a coefficient C and
virtual origin z0 can be determined by fitting the measured data
to Eq. (2). Other self-similar feature of a turbulent jet is the mean
axial velocity profile normalized by the center-line velocity, U/UC ,
Fig. 8. Centerline velocity as a function of the distance to the steam nozzle.
Y.J. Choo, C.-H. Song / Nuclear Engineering and Design 240 (2010) 2215–2224
Table 2
Fitted coefficients of the self-similar turbulent flow for all the test matrices (95% confidence interval within parentheses).
Test I.D.
Decay coefficient (C)
Virtual pole (z0 )
Velocity profile constant (K)
Spreading rate (S)
Other literature
457.1 (446.3–467.8)
460.2 (454.9–465.5)
586.7 (578.2–595.3)
580.8 (573.9–587.7)
587.2 (583.1–591.4)
596.5 (587.0–606.0)
716.1 (706.6–725.6)
720.1 (711.0–729.2)
−2.6D (−3.7 to −1.6)
−3.1D (−3.6 to −2.5)
−1.7D (−2.4 to −1.0)
−0.5D (−1.0 to 0.0)
−0.8D (−1.1 to −0.4)
−2.5D (−3.2 to −1.7)
−1.0D (−1.6 to −0.4)
−0.6D (−1.1 to −0.0)
Dependency (G)
Dependency (Tp )
Increase with Tp
−0.5D to −3D
Decrease with increasing G
76.1 and 75.2 Boersma et al. (1998) and
Panchapakesan and Lumley (1993)
Not clear
Not clear
0.094 Van Wissen et al. (2005) and Hussein et
al. (1994)
Not clear
Not clear
The spreading rate S, which is another coefficient to describe the
shape of a jet, is defined as follows:
r1/2 (z)
z − z0
where r1/2 (z) is the width at which a jet’s axial velocity is half
of the maximum axial velocity UC (z) in the corresponding cross
section, z. This quantity reflects the rate at which the shear layer
grows laterally with the axial distance. Furthermore a straightforward transformation using the relation between K and S gives the
following explanation:
ln2 1/2
All the fitted coefficients mentioned above together with those
of a single phase jet and their dependency on G and Tp are summarized in Table 2. The center-line velocity decay rate C is a key
feature of the steam jet-induced turbulent jet since a jet’s momentum induced by a steam injection can affect the integrity of the
submerged structures in a pool. The center-line decay coefficient
C is strongly dependant on G; it increases with G. The center-line
decay rate shows the tendency of an increase with Tp even though
it is relatively not clear when compared to G.
In contrast to most of the previous results found in open literature on single phase non-condensing jets, the virtual origin z0
for this condensing jet is always negative. There is no doubt that
the virtual origin is mostly decided by the initial condition of the
jet. In the present case of a condensing jet, a sudden phase change
from steam to water occurs in the initial region of the jet and the
Fig. 9. Self-similar profile of the mean axial velocity along the radial direction.
steam jet near the exit of the discharge nozzle has a peculiar shape
depending on the mass flux and the pool temperature, which is
quite different from the case of a non-condensing jet. Quite naturally, the strong driving force and phase change could influence the
near-field condition of the jet and the difference in the virtual origin
of a condensing jet from a single phase jet results from the inherent
nature of a steam jet-driven turbulent jet. The dependency of z0 on
G shows the tendency of a decrease with an increasing G due to the
increase of the height of the steam jet, whereas it shows no clear
evidence for a dependency on Tp .
Both quantities of the velocity profile constant K and the spreading rate S are correlated with each other by Eq. (5). These quantities
were estimated with experimental data at a height of 200 mm
or above where a jet is fully developed and they show a universal feature. The value of the K and S constants were found
to be between 72.4–86.0 and 0.090–0.098, respectively, and they
are almost similar to those of a single phase non-condensing
jet, which is about 76.1 and 0.097, respectively. These results
are consistent with the previous observations not only for a
condensing jet by Van Wissen et al. (2005) but also for a noncondensing jet by other researchers. The dependency of K and S
on G and Tp , however, did not show a clear tendency in the present
5.2. Internal circulation driven by a steam injection
The turbulent jet, which is strongly induced by a steam injection,
entrains the surrounding water and creates an internal circulation
pattern within a pool; it governs the content of a pool mixing. Fig. 10
demonstrates that a coherent circulating flow pattern is created by
the steam injection. As expected, a stronger internal driving flow
can be observed with an increasing mass flux. Key features of the
internal flow pattern in the pool mixing are the location of the center of the recirculation and the existence of a secondary flow. For all
the tests, the eye of a strong recirculation was placed at the bottom
right-hand corner.
For illustrating the typical features of the experimental observation, three local regions were selected and the apparent differences
between G = 300 kg/m2 s and 650 kg/m2 s are shown in Fig. 11. The
region-A is a central part of the pool’s bottom where a jet bounces
to form a weak impinging jet. Due to a stronger re-circulating
current (the region-B), the main current of the jet in the case of
G = 300 kg/m2 s is narrower than that in the case of G = 650 kg/m2 s
(region-A). With an increasing mass flux, the eye of the recirculation was moved upward by the relatively strong entrainment and
a bouncing flow (region-B). Moreover, an apparent secondary flow
(region-C of G = 650 kg/m2 s) appears for the case of a mass flux at
650 kg/m2 s. It is caused by the stronger bouncing flow along the
right wall which tows the comparatively still water at the upper
right region along with it.
Y.J. Choo, C.-H. Song / Nuclear Engineering and Design 240 (2010) 2215–2224
Fig. 10. Internal circulation pattern and velocity field for the mass fluxes of 300, 450 and 650 kg/m2 s with a pool water temperature of 45 ◦ C.
Fig. 11. Local flow patterns and their comparison between G = 300 kg/m2 s and 650 kg/m2 s.
6. Conclusions
Experimental study on the turbulent jet and mixing pattern produced by a steam injection in a pool was conducted separately
using the PIV measurement technique. Self-similar features of the
turbulent jet in an axi-symmetric test domain were measured in
detail and the constants to describe the shape of the jet were
presented for all the test matrices. The pool mixing pattern for
an injected steam mass flux was measured over the whole test
domain. The experiment results for both viewpoints reveal the
detailed velocity structures quantitatively and could be used as
the benchmarking data for the validation of a CFD simulation of
Y.J. Choo, C.-H. Song / Nuclear Engineering and Design 240 (2010) 2215–2224
the relevant phenomena that appear in the nuclear reactor safety
This work was supported by Nuclear R&D Program of the
Korea Science and Engineering Foundation (KOSEF) grant funded
by the Korean government (MEST) (grant code: M2070204000308M0204-00310).
Abramovich, G.N., 1963. The Theory of Turbulent Jets. M.I.T. Press.
Boersma, B.J., Brethouwer, G., Nieuwstadt, F.T.M., 1998. A numerical investigation
on the effect of the inflow conditions on the self-similar region of a round jet.
Phys. Fluids 10 (4), 899–909.
Cho, S., Song, C.H., Park, C.K., Yang, S.K., Chung, M.K., 1998. Experimental study
on dynamic pressure pulse in direct contact condensation of steam discharging into subcooled water. In: The 1st Korea–Japan Symposium on
Nuclear Thermal Hydraulics and Safety (NTHAS98), Pusan, Korea, pp. 291–
Gamble, R.E., Nguyen, T.T., Shiralkar, B.S., Peterson, P.F., Greif, R., Tabata, H., 2001.
Pressure suppression pool mixing in passive advanced BWR plants. Nucl. Eng.
Des. 204, 321–336.
Hussein, H.J., Capp, S.P., George, W.K., 1994. Velocity measurements in a highReynolds-number, momentum-conserving, axisymmetric, turbulent jet. J. Fluid
Mech. 258, 31–75.
Kang, H.S., Song, C.H., 2008. CFD analysis for thermal mixing in a subcooled water
tank under a high steam mass flux discharge condition. Nucl. Eng. Des. 238,
Kim, Y.S., Youn, Y.J., 2008. Experimental study of turbulent jet induced by steam jet
condensation through a hole in a water tank. Int. Commun. Heat Mass Transfer
35, 21–29.
Kim, Y.S., Chung, M.K., Park, J.W., Chun, M.H., 1997. An experimental investigation
of direct condensation of steam jet in subcooled water. J Korea Nucl. Soc. 29,
Panchapakesan, N.R., Lumley, J.L., 1993. Turbulence measurements in axisymmetric
jets of air and helium. Part 1. Air jet. J. Fluid Mech. 246, 197–223.
Pope, S.B., 2000. Turbulent Flows. Cambridge University Press, New York.
Song, C.-H., Cho, S., Kim, H.Y., Bae, Y.Y., Chung, M.K., 1998. Characterization of direct
contact condensation of steam jets discharging into a subcooled water. In: IAEA
TCM Meeting on Advanced Water Cooled Reactor, PSI, Villigen.
Song, C.H., Baek, W.P., Park, J.K., 2007. Thermal-hydraulic tests and analyses for the
APR1400’s development and licensing. Nucl. Eng. Technol. 39, 299–312.
Townsend, A.A., 1976. The Structure of Turbulent Shear Flow, 2nd ed. Cambridge
University Press, Cambridge.
Van Wissen, R.J.E., Schreel, K.R.A.M., Van Der Geld, C.W.M., 2005. Particle image
velocimetry measurements of a steam-driven confined turbulent water jet. J.
Fluid Mech. 530, 353–368.
Weimer, J.C., Feath, G.M., Olson, D.R., 1973. Penetration of vapor jets submerged in
subcooled liquids. AIChE J. 19, 552–558.
Fly UP