You are on page 1of 25

Environ Fluid Mech

DOI 10.1007/s10652-013-9290-7

ORIGINAL ARTICLE

Initial mixing of inclined dense jet in perpendicular


crossflow

Chris C. K. Lai · Joseph H. W. Lee

Received: 24 July 2012 / Accepted: 8 May 2013


© Springer Science+Business Media Dordrecht 2013

Abstract A comprehensive experimental investigation for an inclined (60◦ to vertical)


dense jet in perpendicular crossflow—with a three-dimensional trajectory—is reported. The
detailed tracer concentration field in the vertical cross-section of the bent-over jet is mea-
sured by the laser-induced fluorescence technique for a wide range of jet densimetric Froude
number Fr and ambient to jet velocity ratios Ur . The jet trajectory and dilution determined
from a large number of cross-sectional scalar fields are interpreted by the Lagrangian model
over the entire range of jet-dominated to crossflow-dominated regimes. The mixing during
the ascent phase of the dense jet resembles that of an advected jet or line puff and changes to
a negatively buoyant thermal on descent. It is found that the mixing behavior is governed by a
crossflow Froude number F = Ur Fr . For F < 0.8, the mixing is jet-dominated and governed
by shear entrainment; significant detrainment occurs and the maximum height of rise Z max
is under-predicted as in the case of a dense jet in stagnant fluid. While the jet trajectory in the
horizontal momentum plane is well-predicted, the measurements indicate a greater rise and
slower descent. For F ≥ 0.8 the dense jet becomes significantly bent-over during its ascent
phase; the jet mixing is dominated by vortex entrainment. For F ≥ 2, the detrainment ceases
to have any effect on the jet behavior. The jet trajectory in both the horizontal momentum and
buoyancy planes are well predicted by the model. Despite the under-prediction of terminal
rise, the jet dilution at a large number of cross-sections covering the ascent and descent of
the dense jet are well-predicted. Both the terminal rise and the initial dilution for the inclined
jet in perpendicular crossflow are smaller than those of a corresponding vertical jet. Both
the maximum terminal rise Z max and horizontal lateral penetration Ymax follow a F−1/2
dependence in the crossflow-dominated regime. The initial dilution at terminal rise follows
a S ∼ F1/3 dependence.

C. C. K. Lai
Zachry Department of Civil Engineering, Texas A&M University, College Station, TX, USA
e-mail: chrislck@neo.tamu.edu

J. H. W. Lee (B)
Department of Civil and Environmental Engineering, Hong Kong University of Science
and Technology, Clear Water Bay, Hong Kong, China
e-mail: jhwlee@ust.hk

123
Environ Fluid Mech

Keywords Mixing and transport · Negative buoyant jets · Dense jets · Desalination ·
Vortex entrainment · Jet integral models

1 Introduction

In recent years seawater desalination has been adopted or planned in coastal cities (e.g. Sin-
gapore and Adelaide) as a key component of water supply to cope with population increase
and/or address challenges of water shortage induced by climate change. With advances in
membrane (Reverse Osmosis) technologies, it is expected that desalination will be increas-
ingly competitive in cost compared to alternative water supplies. As a byproduct of desali-
nation, a concentrated brine solution has to be disposed of; the brine has to be returned to
a neighboring water body in a way that minimizes the impact on the benthic community.
The water quality objective is typically set as 5–10 % increase in background salinity (e.g.
EU guidelines) or an upper salinity limit that should not be exceeded [1]. Similar to domes-
tic wastewater discharges, the concentrated brine is typically discharged from a submarine
outfall as inclined high velocity dense jets for the maximization of initial dilution. In some
near shore situations, a surface discharge via an open channel may also be adopted as a
common and cost-effective alternative [2]. Similar dense effluent discharges also arise from
the dissolution of salt caverns for storing oil (e.g. Strategic petroleum reserve, USA), cold
water inflows into lakes, and from Liquified Natural Gas plants. The study of initial mixing of
dense jets is important for the management of coastal water quality. There is a large volume
of literature on the modelling of a dense jet in a stagnant ambient (e.g. [3–9]).
Due to river inputs or tides, a non-zero ambient current exists in the coastal environment
for the great majority of time; predicting the initial mixing in moving water is essential for
environmental risk assessment. Field and mathematical model studies have well demonstrated
that the initial dilution can increase significantly even in the presence of a weak current
(e.g. [10,11]). In addition, wastewater effluent is often discharged horizontally offshore into
an alongshore coastal current to maximize the length of the jet trajectory before the plume
surfaces or reaches the trap level; such a buoyant jet traces out a three-dimensional trajectory.
Whereas detailed tracer concentration measurements in the jet cross-section have been made
for horizontal positively buoyant jets in a perpendicular crossflow, with a 3D trajectory (e.g.
[12,13]), corresponding measurements for dense jets have not been reported.
When the jet is discharged in a direction opposite to that of buoyancy (dense jet), it is
recognized that a buoyant instability exists in the lower half of the jet, leading to the phenom-
enon of “detrainment”—the peeling off of dense fluid from the jet proper. Such a phenomenon
can be clearly observed for jets in stagnant fluid or in a weak crossflow (e.g. [6,9,10]). The
mixing of an inclined dense jet in a crossflow with 3D trajectory was first investigated in a
comprehensive manner by Roberts and Toms [4]. The jet was inclined at 60◦ to vertical and
discharged into a crossflow at different horizontal angles. All the runs were made at a fixed
jet densimetric Froude number and the emphasis was more on counterflow and coflow cases;
only five experiments were conducted for a perpendicular crossflow, with a 3D trajectory. In
the experiments, the visual terminal height of rise was recorded and the minimum dilution at
terminal rise was determined from a concentration transect measured by a suction sampling
technique. Based on the limited measurements and previous work of Tong and Stolzenbach
[14] it was concluded that the difference in dilution between a 60◦ and a vertical (90◦ ) jet
was insignificant. Lindberg [15] performed a comprehensive set of experiments of negatively
buoyant jets with 3D trajectories for a wide range of jet densimetric Froude numbers and
ambient current speeds; the jets were inclined at 30◦ , 45◦ , 60◦ and 90◦ to the vertical and

123
Environ Fluid Mech

discharged into a perpendicular crossflow. The maximum visual terminal height of rise and
horizontal penetration of the jet were measured. There was significant scatter in the data,
which was attributed to the “sweeping away of the descending fluid, and the interaction
between the rising and falling fluid”. In a study of jet transition behavior, Kikkert et al. [16]
used a light attenuation (LA) technique to measure the laterally (or vertically) integrated
concentration profiles of dense jets. Two experiments with dense jets in a crossflow were
reported; the cross-sectional concentration field was not measured and difficulties in measur-
ing the jet trajectory by the LA technique (due to detrainment) were also encountered. More
recently, cross-sectional concentration measurements for a vertical dense jet in a crossflow—
with a 2D trajectory—have been made using a 3D laser induced fluorescence (LIF) technique
[17]. They noted the formation of a vortex pair in the jet descending phase for moderately
strong crossflows (Ur Fr ∼ 0.9) and derived empirical relationships for trajectory and mini-
mum dilution. In all cases considered, the time-averaged jet sections had an elongated lower
half that signifies jet detrainment and is congruent with the earlier investigations based on
a shadowgraph technique [15] and probe-based technique (e.g. [18]). On the other hand,
Palomar et al. [19] performed a comprehensive comparison of the predictions from several
jet integral models (VISJET [11], CorJet [20] and UM3 [21]) with experiments. For a dense
jet in crossflow, the data in Roberts and Toms [4] and Gungor and Roberts [17] were used.
However, the measurements in Roberts and Toms were probe-based, while the experiments
performed in Gungor and Roberts were all in the jet-dominated regime (see next section).
Thus, the model-data comparison in Palomar et al. cannot be regarded as conclusive or com-
plete: (i) first, many important cases in the crossflow-dominated regime are not compared; (ii)
the exact location of impact point is often difficult to define precisely due to the interaction
of jet and bottom density current. This hinders meaningful comparison to be made. In order
to resolve the above issues, it is necessary to measure the cross-sectional concentration field
(and dilution) along the entire jet trajectory (see experiments).
In summary, the earlier measurements on jet dilution of dense jets in a crossflow were usu-
ally made at vertical transects along a visually-determined jet centerline with limited spatial
resolution [4,14,18]. For dense jets with 3D trajectories in particular, it is difficult to obtain
minimum (cross-sectional) jet dilution accurately based on concentration measurements of
limited spatial resolution. As far as we are aware, detailed cross-section tracer concentration
measurements of a horizontal dense jet discharged into a perpendicular crossflow have not
been carried out. The lack of a complete picture of mixing along the jet trajectory seriously
hampers the critical assessment of jet integral models (e.g. [11,20,22]). In addition, it is often
conjectured that the existence of detrainment would lead to an increased jet dilution as in the
case of stagnant ambient [9]. However, in a bent-over jet, vortex entrainment dominates, and
shear entrainment is insignificant. The effect of detrainment for the mixing of dense jets in
a current has not been elucidated in previous studies.
In this paper we report detailed cross-sectional concentration field measurements for an
inclined dense jet into a perpendicular crossflow, with a 3D trajectory. Experiments are carried
out for a dense jet inclined at 60◦ to vertical; a broad range of jet densimetric Froude number
U
Fr and jet to ambient velocity ratios K = Uaj (Fr = 5 − 40, K = 2.4 − 83) are covered. Tracer
distributions at different downstream vertical jet sections (normal to Ua ) are measured using
the LIF technique. We first show by length scales and dimensional analysis that the dense jet
behavior is mainly determined by a crossflow Froude number Ur Fr as a demarcation between
weak and strong crossflows. The theoretical tool used for analysis and interpretation of data—
the Lagrangian jet integral model JETLAG is then briefly outlined. The experimental set up
and procedure are then described. The visual jet trajectory and the cross-sectional scalar

123
Environ Fluid Mech

Fig. 1 An inclined dense jet in a perpendicular crossflow

field are then presented and compared with model predictions. Empirical correlations of the
measured maximum height of rise and lateral horizontal penetration, as well as the terminal
dilution are compared with data of independent investigations.

2 Theory

2.1 Dimensional analysis

Figure 1 shows an inclined turbulent round dense jet discharging into a perpendicular cross-
flow Ua ; without loss of generality the ambient current Ua is assumed to flow in the positive
x-direction. The turbulent jet (diameter D) discharges with velocity U j and density ρo > ρa ,
where ρa is the ambient density. The jet is oriented at a vertical angle θo to the horizontal
plane, and at a horizontal angle φo with the crossflow; φo is defined as the angle between
the positive x-axis and the projection of U j on the x–y plane. In this study, φo = 90◦ in all
experiments.
The jet mixes with the ambient fluid as it rises, but is decelerated by virtue of its negative
buoyancy; a maximum height of rise Zmax is reached before the jet falls over as a negatively
buoyant plume to impinge on the nearby sea bed as a bottom spreading density current. The
concentration field c(x, y, z) of a dense jet depends on the jet discharge concentration Co , the
initial density difference ρo = ρo − ρa , U j , D, θo and φo . In addition, the mixing behavior
is affected by the complex jet impingement on the bottom and the interaction of the jet and
the density current. The jet discharge densimetric  Froude number, a measure of the ratio of
inertia to buoyancy, can be defined as Fr = U j / go D, where g = gravitational acceleration
and go = (ρo /ρa )g = reduced gravitational acceleration.

123
Environ Fluid Mech

The inclined dense jet can be characterized by the jet discharge volume flux Q o =
U j π D 2 /4, kinematic momentum flux Mo = U 2j π D 2 /4, and specific buoyancy flux Bo =
Q o go . For a dense jet in stagnant fluid, a characteristic length (e.g. maximum height of rise)
can be expressed in terms of the source fluxes as:
Z max = f (Q o , Mo , Bo , θo ) (1)
By dimensional analysis it can then be shown that:
Z max
= f (l M /l Q , θo ) (2)
lM
or alternatively as:
Z max
= f (Fr, θo ) (3)
Fr D
3/4 1/2
where l M = Mo /Bo is a momentum length scale formed from the momentum and
1/2
buoyancy fluxes, and l Q = Q o /Mo = (π/4)0.5 D is a source length scale formed from the
volume and momentum fluxes. l M is a measure of the distance within which jet momentum
is more important than buoyancy, and l Q represents the length over which source geometry is
important. It should be noted that l M and l Q can be related to the jet diameter and densimetric
Froude number as l M = (π/4)1/4 Fr D and l Q = (π/4)1/2 D.
Similarly, the dilution flow entrained at a level z can be written as:
Q = f (Q o , Mo , Bo , z, θo ) (4)
By dimensional analysis, the total dilution flow or jet volume flux Q can be written as:
1/2
Q Bo
5/4
= f (z/l M , l M /l Q , θo ) (5)
Mo
As a characteristic dilution (e.g. average or minimum dilution) is proportional to the volume
flux, S ∼ Q/Q o , it can be shown that the characteristic jet dilution can be expressed as:
 
Sm Z max
= f , Fr, θo = f (Fr, θo ) (6)
Fr Fr D
Experiments have shown that for large jet densimetric Froude numbers (Fr ≥ 20), the
dimensionless terminal rise Z max /(Fr D) and initial dilution S/Fr are independent of Fr
(e.g. [5,9]).
The above dimensionless framework can be extended to the case of a crossflow. Consider
the jet with initial momentum Mho = Mo cos θo ∼ Mo in the + y direction. In the presence
of an ambient current (in + x direction), the dense jet mixes with the crossflow by shear and
vortex entrainment, and traces out a three-dimensional jet trajectory while being bent over.
Similar to a positively buoyant jet in an ambient current, a crossflow momentum length scale
1/2 1/2
can be defined as Mho /Ua ∼ Mo /Ua = lm . lm is a measure of the transverse distance in
the momentum (x-y) plane at which the jet momentum induced velocity (v ∼ Mho y −1 ∼
1/2

Mo y −1 ) has decayed to the ambient current value; lm is hence a measure of the horizontal
1/2

transverse distance at which the jet starts to be bent over.


By dimensional analysis, the maximum height of rise can be written as:
Z max = f (Q o , Mo , Bo , Ua , θo , φo ) (7)

123
Environ Fluid Mech

(a)

(b)

Fig. 2 Behavior of inclined dense jets in a perpendicular crossflow for a jet-dominated cases and b crossflow-
dominated cases

or alternatively as (for given θo and φo ):


 
Z max lM lM
= f , (8)
lM l Q lm
Fr DUa
It should be noted that l M /lm ∼ 1/2 can be written as l M /lm = (4/π)1/4 Ur Fr , where
Mo
Ur = Ua /U j is the ambient to jet velocity ratio.
Consider first the limiting case of a dense jet in a weak crossflow, with l M /lm ∼ Ur Fr 1.
Near the source, as the jet is hardly bent over, the behavior is similar to a dense jet in stagnant
fluid, slightly advected in the + x direction (Fig. 2a). Based on previous experiments and
theory with dense jets in stagnant fluid (e.g. [6,9]), it is known that the dense jet is momentum-
dominated before the terminal rise height is reached, with Z max ∼ l M . In other words, the
terminal rise occurs in the so called momentum-dominated near field (MDNF) of the jet [10],
with Z max << lm . In addition, both flow visualization and concentration measurements show
that for θo = 60◦ , there is detrainment and peeling off of jet fluid from the lower half of the
jet due to buoyant instability. For this weak crossflow case, the terminal rise will be reached
when the jet is hardly bent over, and significant detrainment of jet fluid can be expected.

123
Environ Fluid Mech

On the other hand, in a strong crossflow, l M /lm ∼ Ur Fr  1, the dense jet will be
significantly bent over during the jet-momentum dominated regime; the mixing will be dom-
inated by the ambient current during most of the jet’s ascent to the terminal level (Fig. 2b).
Hence the mixing of such a dense jet would be analogous to the behavior of advected line
puffs and advected line thermals in a crossflow (e.g. [11]). The terminal rise occurs in the
so called momentum-dominated far field (MDFF)of the jet [10], with Z max >> lm (Fig.2b).
Ultimately, the mixing is dominated by the buoyancy, and the behavior will be similar to
that of advected line thermals in the so called buoyancy-dominated far field (BDFF) of
the jet. As experiments have shown that the mixing in the MDNF is governed by shear
entrainment (similar to dense jet in stagnant fluid), and that in the bent-over phase (advected
line puffs and thermals) is governed by vortex entrainment, it is expected that the dense jet
mixing in a perpendicular crossflow is governed by l M /lm . The parameter F = Ur Fr can
also be interpreted as a crossflow densimetric Froude number.
By dimensional analysis, the terminal rise of a 60o dense jet in a perpendicular crossflow
can then be written as:
 
Z max lM lM
= f , = f (Ur Fr, Fr ) (9)
Fr D lm l Q
Smin
= f (Ur Fr, Fr ) (10)
Fr

2.2 Numerical modelling

The Lagrangian model JETLAG/VISJET is employed for analysis and interpretation of the
experimental results. JETLAG/VISJET is a well-proven general model that predicts the
near field mixing of an arbitrarily-inclined round buoyant jet in a current [11,22,23]). The
unknown jet trajectory is viewed as a sequential series of plume elements which increase
in mass as a result of shear entrainment (due to the jet discharge) and vortex entrainment
(due to crossflow)—while rising by buoyant acceleration and being sheared over by the
crossflow. The model tracks the evolution of the average properties of a plume element
at each step by conservation of horizontal and vertical momentum, conservation of mass
accounting for entrainment, and conservation of tracer mass/heat. The increase in mass of
the plume element at each step is computed as a function of two components: the shear
entrainment due to the relative velocity of the plume element and the ambient velocity, and
the vortex entrainment due to the ambient crossflow. An entrainment hypothesis is adopted
for the shear entrainment; the entrainment coefficient α varies with the local densimetric
Froude number Fr L as well as the relative velocity in the direction of the jet axis [9,23]. The
entrainment due to the crossflow is modelled using the projected area entrainment hypothesis;
the vortex entrainment is computed by the ambient flow intercepted by the ’windward’ face
of the plume element. The theoretical basis of the formulation has been discussed in [24];
2D/3D turbulence model calculations have shown that for advected line puffs and thermals all
ambient fluid presented to the front of the puff/thermal section gets entrained. The transition
from the jet-dominated (shear entrainment) to the crossflow-dominated (vortex entrainment)
regime is modelled in a general manner by a general entrainment formulation which is
incorporated in VISJET [11].The performance of the model has been extensively validated
against basic data of positively buoyant jets in stagnant fluid or in a stratified crossflow
(e.g. [25]).

123
Environ Fluid Mech

Fig. 3 Experimental setup for a 60◦ dense jet discharged into a perpendicular crossflow Ua

3 Experiments

A series of experiments were carried out in a 1.2m wide flume partitioned from a 5 × 11 m
× 1 m deep shallow water basin. Figure 3 shows a schematic diagram of the experimental
setup. A jet nozzle was mounted at the center of a 1.8 m long Perpex false floor; the horizon-
tal jet orientation with respect to the ambient current, φo , can be controlled with a turntable
mechanism; the vertical jet angle θo can be adjusted (0–90◦ ) via a hinge joint. The jet dis-
charge was supplied from a constant head tank and monitored by a calibrated rotameter. The
crossflow was generated by a re-circulating flow system consisting of a centrifugal pump
and diffuser manifolds for uniform flow distribution at the inflow and outflow ends of the
shallow water basin. The crossflow was monitored by a Kent electromagnetic flow meter. The
uniformity of the ambient current at the measurement sections was checked with Acoustic
Doppler Velocimetry (ADV). For the range of ambient velocity Ua studied (1.34–12.4 cm/s),
the velocities around the jet nozzle was within ±5% from the cross-sectional average
velocity.
Three jet diameters were used: D = 3, 5 and 11 mm; the port height h p was fixed at 5 cm.
All experiments were performed with θo = 60◦ and φo = 90◦ , and were designed such that
the ambient flow was at least ten times the jet entrainment demand, and free from boundary
effects. By the nature of the set up, the jet configuration in the experiments was a mirror
image of that in the definition diagram (Figs. 1, 2).
Table salt (NaCl) was added to fresh tap water as the source of negative buoyancy and the
initial density difference was smaller than 3.3 % in all experiments. Rhodamine-6G fluores-
cence dye was used as a tracer; the initial dye concentration Co was typically in the range
of 0.1–0.6 mg/L. The non-intrusive planar LIF technique was used to measure the tracer
concentration field of the jet at selected vertical planes normal to the ambient crossflow Ua .
A 2 mm thick laser sheet was generated by a 5W Argon-ion laser (Spectra-Physics, Stabilite
2017) and a high frequency rotating mirror. Jet images were captured with an underwater

123
Environ Fluid Mech

Fig. 4 General observations of 60o dense jets in perpendicular crossflows; showing the effects of increasing
Ur Fr (hence Ua ) on jet behavior (Fr ≈ 24, symbols are data points of measured trajectory, solid and dotted
lines are the predicted (JETLAG) centerline trajectory and visual jet boundaries)

CCD camera equipped with a Nikon 35mm lens; the image resolution was 576 × 768 pix-
els and captured jet images were digitized as 8-bit gray scale bitmap files. A Hoya orange
filter was used to filter out scattered laser light and other background light. The camera
was mounted on a movable platform that could traverse in the longitudinal x-direction.
The field of view of the images was 26 × 35 cm2 . Once the camera focus was properly
adjusted, the separation between the camera and the image section was held fixed in all
experiments. This permitted the use of a single length conversion (pixels to cm) factor and
eliminated the dependence of LIF calibration results on the image distance to the camera.

123
Environ Fluid Mech

The linear response between dye concentration and light intensity for the present range of
concentrations had been established. More details of the experimental set up can be found in
Lai[26].
A total of 45 experiments were carried out, covering jet densimetric Froude number in
U
the range of Fr = 5 − 40 and jet to ambient velocity ratio of K = Uaj = 1/Ur = 2.4 − 83.
The jet Reynolds number Re was well above the critical limit of 2000. A summary of the
experimental parameters is given in Table 1.

3.1 Experiment procedure

The source solution with the required Co was first prepared and the ambient current was
set up one hour before measurements. The ambient and source temperatures were mea-
sured by an electronic thermometer and the corresponding densities by a Kyoto Electron-
ics density meter (model DA-500). The glass panel through which the laser sheet passed
was cleaned to remove any air bubbles, dust and grease. The area surrounding the mea-
surement section was darkened; background images were taken. The jet discharge was
then turned on and the dense jet flow allowed to reach a steady state; jet images were
then captured at a sampling frequency of 10 Hz for a period of 60–120 s. The laser sheet
and CCD camera (at fixed separation) were then moved to the next vertical jet section
and the measurements repeated. Source fluid with 10, 20 and 40 dilution was placed
in a Perspex calibration box (60 cm × 33 cm by 4.1 cm thick) and illuminated with the
laser sheet; the corresponding light intensity was determined. The initial dye concentration
Co was obtained from the average of the three measurements. The water in the shallow
water basin was drained completely at the end of 2–3 runs and replaced with fresh tap
water.

4 Results

4.1 General flow observation

The overall character of a dense jet in a perpendicular crossflow depends on the dimensionless
parameter l M /lm or Ur Fr . Figure 4 shows the measured jet trajectory in the horizontal
momentum (x–y) plane as well as the side view of the observed 60◦ dense jet in the buoyancy
(x–z) plane in a representative experiment. The jet densimetric Froude number is Fr ≈ 24
and the crossflow Froude number is Ur Fr ≈ 0.8 − 2.2. The jet trajectory is determined from
the centre of mass of the jet cross-sectional concentration field (see later discussion). The
centerline trajectory and visual boundaries predicted by the JETLAG model are also shown.
It can be seen that for Ur Fr ∼ 1 the jet is significantly bent-over near the terminal rise
with a side stream of detrained fluid moving predominantly in the direction of the cross-
flow Ua . Although the trajectory in the momentum (x–y) plane is well-predicted, the termi-
nal rise is higher than predicted as a result of the detrainment (loss of negative buoyancy)
as in the stagnant ambient case [9]. At Ur Fr ≈ 1.1, it is seen that the maximum termi-
nal rise occurs later than when the jet is bent over. With increase in ambient current, the
mixing behavior of the dense jet appears to be dominated by the ambient crossflow dur-
ing its ascent to the terminal level. It is hence not surprising that JETLAG predictions in
both the momentum and buoyancy planes are in good agreement with observations in this
regime.

123
Environ Fluid Mech

Table 1 Experimental parameters of 60◦ dense jets discharged perpendicularly into a uniform crossflow;
total number of experiments = 45
ρ
Run no. H (mm) D (mm) U j (cm/s) Ua (cm/s) Co (mg/L) ρa (%) Re Fr K = 1/Ur Ur Fr
U
= Uaj

F5-K2a 228 11 30.2 12.40 0.12 3.17 3800 5.2 2.4 2.12
F5-K2c 230 11 30.2 12.29 0.02 3.28 3834 5.1 2.5 2.07
F5-K5a 228 11 30.2 6.20 0.12 3.17 3800 5.2 4.9 1.06
F5-K5c 230 11 30.2 6.15 0.02 3.28 3834 5.1 4.9 1.03
F5-K10a 228 11 30.2 3.22 0.12 2.93 3851 5.4 9.4 0.57
F5-K10b 423 11 30.2 3.10 0.06 3.26 3791 5.1 9.8 0.52
F8-K4a 228 11 49.7 12.40 0.12 3.28 6401 8.4 4.0 2.09
F8-K4b 230 11 49.7 12.29 0.02 3.28 6303 8.4 4.0 2.06
F8-K8a 228 11 49.7 6.20 0.12 3.28 6401 8.4 8.0 1.04
F8-K8b 230 11 49.7 6.15 0.02 3.28 6303 8.4 8.1 1.04
F8-K8c 423 11 49.7 6.19 0.06 3.09 6206 8.6 8.0 1.07
F8-K16b 423 11 49.7 3.10 0.06 3.26 6234 8.4 16.1 0.52
F12-K4a 255 11 49.7 10.88 0.18 1.67 6136 11.7 4.6 2.56
F12-K4b 287 5 39.9 8.94 0.30 2.33 2274 11.8 4.5 2.64
F12-K5a 354 11 35.1 6.51 0.12 0.79 4400 12.0 5.4 2.23
F12-K5d 378 5 39.9 7.46 0.20 2.37 2183 11.7 5.3 2.19
F12-K7b 287 5 39.9 5.66 0.30 2.33 2274 11.8 7.0 1.67
F12-K10a 350 11 49.7 4.78 0.18 1.67 6136 11.7 10.4 1.13
F12-K10b 354 11 35.1 3.40 0.12 0.79 4409 12.0 10.3 1.16
F12-K10d 378 5 39.9 3.88 0.20 2.37 2183 11.7 10.3 1.14
F12-K10e 396 5 47.8 4.36 0.12 2.99 2474 12.5 11.0 1.14
F12-K15a 354 11 35.1 2.37 0.12 0.79 4409 12.0 14.8 0.81
F12-K15c 378 5 39.9 2.77 0.20 2.38 2203 11.7 14.4 0.81
F18-K8 354 5 59.9 7.25 0.20 2.44 3105 17.3 8.3 2.09
F18-K16a 354 5 59.9 3.85 0.20 2.44 3076 17.3 15.5 1.11
F18-K16b 396 5 75.1 4.36 0.12 2.99 3787 19.6 17.2 1.14
F18-K21a 354 5 59.9 2.81 0.20 2.44 3164 17.3 21.3 0.81
F18-K21b 396 5 75.1 3.17 0.12 2.99 3787 19.6 23.7 0.83
F24-K9 270 5 79.8 8.92 0.30 2.28 4620 23.9 8.9 2.67
F24-K11a 253 5 79.8 7.45 0.60 2.33 4498 23.6 10.7 2.20
F24-K11b 378 5 79.8 7.46 0.20 2.37 4377 23.4 10.7 2.19
F24-K11c 396 5 102.4 8.07 0.12 2.99 5163 26.7 12.7 2.10
F24-K14 270 5 79.8 5.62 0.30 2.28 4620 23.9 14.2 1.68
F24-K21a 315 5 79.8 3.82 0.40 2.27 4743 23.9 20.9 1.14
F24-K21b 378 5 79.8 3.88 0.20 2.37 4377 23.4 20.6 1.14
F24-K28b 378 5 79.8 2.77 0.20 2.38 4417 23.4 28.8 0.81
F24-K28c 396 5 102.4 3.17 0.12 2.99 5301 26.7 32.3 0.83
F24-K45 295 5 79.8 1.78 0.20 2.45 4690 23.1 44.8 0.52
F40-K15a 313 3 110.8 7.43 0.40 2.76 3739 38.8 14.9 2.60
F40-K15b 390 3 110.8 7.38 0.20 2.62 3807 39.9 15.0 2.66
F40-K20a 313 3 110.8 5.57 0.40 2.77 3773 38.8 19.9 1.95

123
Environ Fluid Mech

Table 1 continued
ρ
Run no. H (mm) D (mm) U j (cm/s) Ua (cm/s) Co (mg/L) ρa (%) Re Fr K = 1/Ur Ur Fr
U
= Uaj

F40-K20b 390 3 110.8 5.64 0.20 2.62 3815 39.9 19.6 2.03
F40-K30 390 3 110.8 3.76 0.20 2.62 3807 39.9 29.5 1.36
F40-K40 390 3 110.8 2.15 0.20 2.62 3807 39.9 41.2 0.97
F40-K80 390 3 110.8 1.34 0.20 2.62 3807 39.9 82.7 0.48

4.2 Vertical jet cross-sections

Figure 5 shows typical measured tracer concentration field in consecutive vertical cross-
sections of the bent-over dense jet in an experiment (Fr = 8.4, K = 8, Ur Fr = 1.04).
The concentration contour lines are shown at an interval of 0.05Cmax from C/Cmax = 1
down to 0.25, where Cmax is the cross-sectional maximum concentration; the location of the
concentration maximum is indicated in the figure by a cross. For sections before terminal
rise Zmax , x = 2.3 − 9.1D, the double concentration peak characteristic of a jet in crossflow
can be clearly seen. As compared to a vertical discharge, the centerline bisecting the double
peak structure (advected line puff) is initially inclined to the vertical at an angle close to
the initial vertical jet discharge angle of θo = 60◦ . As Zmax is approached, the inclination
angle reduces as the jet aligns with the vertical direction due to the increasing effect of
negative buoyancy. The predicted jet trajectory is in good agreement with observations in
the horizontal momentum plane; as the effect of detrainment and hence the loss of negative
buoyancy cannot be accounted for in the JETLAG model, the terminal rise is under-predicted.
It is also noteworthy that the concentration peaks are not symmetrical, possibly reflecting the
effect of detrainment. The distinct double-peak structure disappears beyond terminal rise;
the scalar field is more circular after Zmax (x = 15.9 − 27.3D).
Figure 6 shows the corresponding concentration contours for a dense jet with the same jet
densimetric Froude number but with the ambient current Ua and hence Ur Fr doubled. The
general kidney-shaped distribution is still observed before Zmax but the double concentration
maxima are replaced by a broader, kidney shaped and smeared region of high concentration
at the center. The bifurcation in scalar field is less apparent in the stronger ambient current; in
general only one distinct concentration peak is observed. These observations are consistent
with the observed cross-sectional concentration field in the vertical cross-section of a bent-
over positively buoyant horizontal heated jet in perpendicular crossflow [13]. It is seen that
for this crossflow-dominated experiment (Ur Fr ≈ 2.1), the predicted jet trajectory is in good
agreement with experimental data in both the momentum and buoyancy plane.
The jet trajectory can be defined by either (i) the location of the concentration maximum
Cmax , (Yc ,Zc ); or (ii) the position (Ycm ,Zcm ) of the centre of mass of the concentration
distribution. When the scalar field is bifurcated, the use of the centre of mass is more suitable;
the two definitions give similar results when there is only one single concentration peak or
when the bifurcation is weak. Detailed study has shown that the location of the centre of mass
is not sensitive to the choice of jet boundary defined by C/Cmax in the range of 0.1–0.4.

4.3 Trajectory and visual jet radius

The jet trajectories defined by the centre of mass (Ycm ,Zcm ) are compared with predictions
for representative experiments with Ur Fr < 1 (Fig. 7); with Ur Fr ≈ 1 (Fig. 8); and with

123
Environ Fluid Mech

(a)

(b)

Fig. 5 a Measured tracer concentration field in the vertical cross-sections of a bent-over 60◦ dense jet (Run
no. F8-K8a, Fr = 8.4, K = 8, Ur Fr = 1.04, l M = 8.7 cm, lm = 7.87 cm); showing the change in the
structure of scalar field before and after terminal rise; and b measured trajectory defined by the center of mass
(Ycm ,Zcm )

Ur Fr > 2 (Fig. 9). It is seen that over the range of Fr ≈ 5 − 40; K = 2.4 − 41 the
same trends of dense jet behavior can be observed. In general, for large Fr the line of
demarcation between jet-dominated and crossflow-dominated regimes seems to be around
Ur Fr ≈ 0.8. This is consistent with the Z max data of Roberts and Toms [4]. In the jet-
dominated regime, detrainment is expected to be significant. It can be seen that while
the jet trajectory in the momentum plane is well-predicted, the terminal rise is underpre-
dicted by about 20 % which is consistent with the measurements of a dense jet in a stagnant
ambient [9].
Figure 10 shows the comparison of the dimensionless measured maximum height of
rise, Z max /Fr D as a function of Ur Fr . Related data of other investigators are also shown

123
Environ Fluid Mech

(a)

(b)

Fig. 6 a Measured tracer concentration field in the vertical cross-sections of a bent-over 60◦ dense jet (Run
no. F8-K4a, Fr = 8.4, K = 4, Ur Fr = 2.09, l M = 8.7cm, lm = 3.92cm); and (b) measured trajectory defined
by the center of mass (Ycm ,Zcm )

for comparison. In general the maximum height of rise measured from the detailed cross-
section concentration contours are consistent with previous data [4,14]. As predicted by
dimensional analysis, Eq. 9 (Sect. 2.1), the data correlates well with Ur Fr ; for a weak
current or near stagnant case, Ur Fr < 0.8, the dimensionless measured height of rise is the
same as that for a stagnant case. Due to detrainment of jet fluid, the maximum height of
rise is under-predicted by about 20–30 % in this jet-dominated regime [9]. However, in the
crossflow-dominated regime, Ur Fr > 1, the dimensionless terminal rise height is seen to
decrease as ∼ (Ur Fr )−1/2 . Both the data and theory show that Z max for a 60◦ jet is less
than than of a vertical dense jet; the best fit of the data of Gungor and Roberts [17] gives
Z max /Fr D = 2.5(Ur Fr )−1/3 for a vertical jet. On the other hand, the best fit of our data gives
Z max /Fr D = 1.79(Ur Fr )−0.48 ; this dependence is also close to that obtained by Lindberg
[15]. It is also seen that the JETLAG predictions reproduce the asymptotic behavior in strong
crossflows (large Ur Fr ). However, the asymptotic behavior is quite different from the −1/3
behavior reported for vertical dense jets in a crossflow [17]. Similarly, the observed maximum
horizontal penetration of the dense jet, Ymax , is compared with predictions in Fig. 11. It
is interesting to note that both the data and model show a similar variation with Ur Fr ,

123
Environ Fluid Mech

Fig. 7 Comparison between measured and predicted jet centerline trajectory for Ur Fr < 1

Ymax /Fr D = 1.4(Ur Fr )−0.48 . The best fit of the data of Lindberg [15] is also comparable
with the present data. It should also be noted that the present data and that of Lindberg [15]
cover a sufficiently wide range of the crossflow parameter, up to Ur Fr ≥ 3. On the other
hand, the experiments of Gungor and Roberts [17] are characterized by 0.23 < Ur Fr < 0.92;
and the majority of the data are within the jet-dominated regime. Similarly the best fit of the
data of Roberts and Toms [4] include five runs with 0.23 < Ur Fr < 1.87; their data are in
good agreement with the best fit of this study.
The visual jet boundary (of a dyed jet) demarcates the turbulent and non-turbulent regions
of the flow; this turbulence property can be characterized by a turbulent intermittency func-
tion γ that expresses the percentage of time a locality is occupied by the jet fluid. For pure
jets, plumes, and dense jets in stagnant ambient, as well as coflow jets, advected line puffs
and thermals, the contour γ = 0.5 corresponds to the visual boundary [9,11]. On the other

123
Environ Fluid Mech

Fig. 8 Comparison between measured and predicted jet centerline trajectory for Ur Fr ≈ 1

123
Environ Fluid Mech

Fig. 9 Comparison between measured and predicted jet centerline trajectory for Ur Fr ≥ 2

123
Environ Fluid Mech

Fig. 10 Measured terminal rise height Zmax

Fig. 11 Measured lateral maximum penetration Ymax at Zmax

hand, theory and LIF experiments have also shown that the 0.25Cmax concentration con-
tour corresponds approximately to the visual jet boundary for the above basic flows. This
correspondence of the visual boundary with a concentration contour greatly facilitates the

123
Environ Fluid Mech

Fig. 12 Comparison between measured and predicted normalized visual jet radius R/D

interpretation of experimental data. As long as the local dye concentration is sufficient to


ensure a good signal-to-noise ratio, this definition of the visual boundary does not depend on
Co ; this condition is satisfied with the chosen range of Co .
Finally, the relationship between the visual jet radius and the top-hat jet radius B adopted
in JETLAG can be established [25,27]. For bent-over jet and plumes in crossflow, both
experiments and turbulence model computations have revealed the existence of added mass
[11,27]. For integral models that neglect added mass, the predicted √jet half-width (B) is larger
than the visual half-width Bv (“top-hat” profile) by a factor of 1 + k, where k = 1 for a
bent-over buoyant jet in crossflow. Because the added mass effect is√not considered √ in the
JETLAG model, the visual half-width can be determined as Bv = B/ 1 + k = B/ 2. This
physical interpretation is supported by extensive experiments.
The comparison for the dimensionless visual jet radius R/D is shown in Fig. 12; R was
found from the jet area bounded by the 0.25Cmax contour in the experiments i.e. it is defined
as the radius of a circle of equivalent area. In general the measured visual jet radius is about
18.5 % larger than the predicted (Bv ), again reflecting the effect of detrainment which leads
to a more elongated jet cross-section.

4.4 Dilution

The measured minimum jet dilution Smin is defined by Co /Cmax (Co = initial dye
concentration) and St is determined from the measured cross-sectional concentration field at
the visually observed location of Z max . Table 2 gives the measured minimum and average
dilution at Z max as well as the centre of mass of the concentration field. The average dilu-
tion is determined from the average concentration in the jet area bounded by the 0.25 Cmax
concentration contour. Figure 13 shows the comparison between measured and predicted
normalized minimum dilution at terminal rise as a function of the crossflow Froude number.
The maximum concentration is obtained from the JETLAG predicted average concentra-
tion using a heuristic formulation for near-far field transition [11] such that in the limit of

123
Environ Fluid Mech

Table 2 Measured minimum and average dilution St at Zmax


St
Run no. Fr K Ur Fr Center of mass St Fr
x/FrD ycm /FrD zcm /FrD min avg

F40-K15a 38.8 14.9 2.60 1.72 0.65 0.76 36.79 63.86 0.95
F24-K9 23.9 9 2.66 1.67 0.38 0.49 24.79 43.07 1.04
F12-K4a 11.7 4.6 2.56 1.71 0.43 0.57 13.39 22.45 1.15
F12-K4b 11.8 4.5 2.62 1.69 0.30 0.59 13.74 23.42 1.16
F40-K20a 38.8 19.9 1.95 1.29 0.72 1.01 32.77 56.47 0.84
F24-K11b 23.4 10.7 2.19 1.71 0.63 0.84 23.09 41.42 0.99
F24-K11c 26.7 12.7 2.10 2.25 0.67 0.88 29.38 56.52 1.10
F18-K8 17.3 8.3 2.10 1.74 0.68 0.73 19.08 34.59 1.10
F12-K5a 12.0 5.4 2.23 1.67 0.48 0.60 12.61 21.81 1.05
F12-K5d 11.7 5.4 2.19 1.74 0.43 0.62 13.52 24.75 1.15
F8-K4a 8.4 4.0 2.08 1.36 0.46 0.60 8.35 14.08 1.00
F5-K2a 5.2 2.4 2.12 1.32 0.46 0.55 5.63 9.75 1.09
F5-K2c 5.1 2.5 2.07 1.34 0.51 0.58 5.73 10.03 1.13
F24-K14 23.9 14.2 1.68 1.26 0.67 0.81 24.01 42.98 1.00
F12-K7b 11.8 7.1 1.66 1.27 0.61 0.72 11.92 22.01 1.01
F40-K30 39.9 29.5 1.35 2.09 0.84 1.25 43.81 79.13 1.10
F40-K40 39.9 41.2 0.97 1.17 0.78 1.46 34.90 64.64 0.87
F24-K21a 23.9 20.9 1.14 0.94 0.74 1.15 19.61 35.16 0.82
F18-K16a 17.3 15.5 1.11 1.16 0.96 1.21 16.71 30.46 0.97
F12-K10a 11.7 10.4 1.12 0.78 0.75 1.03 11.46 18.17 0.98
F12-K10b 12.0 10.3 1.16 0.76 0.69 0.98 11.02 17.79 0.92
F12-K10d 11.7 10.3 1.14 0.85 0.87 1.21 11.30 20.93 0.96
F12-K10e 12.5 11.0 1.14 0.80 0.80 1.04 12.04 22.32 0.96
F8-K8a 8.4 8.0 1.04 1.09 0.80 1.08 7.88 13.19 0.94
F5-K5d 5.2 4.9 1.07 0.87 0.75 0.93 4.76 8.51 0.92
F24-K28b 23.4 28.8 0.81 0.64 0.76 1.32 16.70 29.16 0.71
F24-K28c 26.7 32.3 0.83 0.56 0.69 1.14 17.83 31.16 0.67
F18-K21a 17.3 21.3 0.81 0.58 0.88 1.37 12.66 22.77 0.73
F18-K21b 19.6 23.7 0.83 1.02 0.95 1.44 16.62 27.23 0.85
F12-K15c 11.7 14.4 0.81 0.86 0.87 1.40 8.62 15.22 0.74
F40-K80 39.9 82.7 0.48 0.33 1.09 1.50 22.29 37.93 0.56
F24-K45 23.1 45 0.51 0.87 1.08 1.28 12.42 23.08 0.54
F8-K16b 8.4 16.1 0.52 0.54 1.03 1.27 5.01 9.04 0.60
F5-K10a 5.4 9.4 0.57 0.42 1.04 1.23 4.83 8.20 0.89
F5-K10b 5.1 9.8 0.52 0.45 1.07 1.24 4.14 7.37 0.81

jets/plumes, Cmax /Cavg = 1.7 (the theoretical value for a Gaussian concentration profile,
see e.g. [10]), while an experimentally-determined value of 2.3 (for a horse-shoe shaped
double-peak structure) is used for bent-over buoyant jets [13]. It is interesting to note that in
general, the initial dilution at Zmax is notably smaller for an inclined jet (60o ) as compared
with a vertical jet in a crossflow; in addition, the minimum dilution varies with (Ur Fr )1/3 in
the crossflow-dominated regime. The dilution is well-predicted for Ur Fr = 0.8 − 2.7. This

123
Environ Fluid Mech

Fig. 13 Measured normalized minimum terminal rise dilution St

(Ur Fr )1/3 behavior is supported by the model, and is distinctly different from the (Ur Fr )1/2
dependence as reported by Roberts and Toms [4] or Gungor and Roberts [17] for the vertical
jet. This may be attributed in their inclusion of data of jet-dominated experiments in the best
fit; in fact the dilution data of Roberts and Toms [4] for Ur Fr ≥ 1 are in good agreement
with the present best fit and also the predictions of JETLAG.
The dimensional analysis (Eqs. 9, 10) show that the normalized non-dimensional maxi-
mum height of rise and dilution may also depend on the jet densimetric Froude number Fr
in addition to the crossflow Froude number F; this has been supported by data for a dense jet
in stagnant fluid (e.g. [9]). This Fr dependence seems to be much weaker for a crossflow.
For example, for F ≈ 2 the dilution is rather independent of Fr with St /Fr ≈ 1.0 − 1.1 over
the range of Fr = 5 − 40 (Table 2). For F ∼ 0.8, St /Fr ≈ 0.7 − 0.85 over the range of
Fr ∼ 10 − 30. The Fr dependence is only notable for the jet-dominated regime; for F ∼ 0.5,
the dilution varies from St /Fr ≈ 0.6 for Fr ∼ 10 − 40 to St /Fr ≈ 0.8 − 0.9 for Fr ≈ 5
(Table 2).
Additional insights on the dense jet mixing can be obtained by comparing the predicted
and measured variation of maximum concentration (as inferred from the vertical cross-
section measurements) with distance downstream (Fig. 14). In Fig. 14a the concentration
variation in both the ascent to terminal rise as well as the descent phase is illustrated for
representative jet-dominated experiments. For example, in the experiment with Fr ≈ 8
and Ur Fr ≈ 1 (left most transect in Fig. 14a; corresponding to the visualisation in Fig.
5), the terminal rise is located at x ≈ 10 cm. It is seen that despite the under-prediction
of maximum rise height (due to inability to account for detrainment), the measured dilu-
tion is reasonably predicted by the model in both the ascent and descent phases, until close
to impingement with the bottom boundary. By referring to Fig. 5b, one possible expla-
nation is that although the terminal rise is under-predicted, the mixing gained from the

123
Environ Fluid Mech

(a)

(b)

Fig. 14 Comparison between measured and predicted maximum concentration c/Co = 1/Smin as function
of downstream distance x for a jet-dominated cases; and b crossflow-dominated cases

longer jet trajectory during the descent phase is negated or compensated by the loss of
negative buoyancy, resulting in similar observed dilutions. Fig. 14b) shows similar con-
centration comparisons for the crossflow-dominated cases, Ur Fr > 1; for example the
experiment with Fr ≈ 8 and Ur Fr ≈ 2.1 (left most transect in Fig. 14b; corresponding
to the visualization in Fig. 6). It is seen the predicted concentration is in closer agreement
with measurements due to the relatively smaller influence of detrainment on the overall jet
mixing. The comparison of predicted and measured minimum and average dilution, Smin
and Savg , in a large number of vertical cross-sections is shown in Fig. 15; overall the dense
jet dilution is quite well predicted. In the comparison, data for Ur Fr ≥ 0.8 and from jet
cross-sections without the lower edge touching the bottom boundary are used. The aver-
age dilution Savg is given by Co /Cavg where Cavg is the measured average concentration
in the jet area Av bounded by the 0.25Cmax concentration contour [25]. For a bent-over jet
in crossflow, u ≈ Ua (where u = x-velocity), and the theoretical flux-average dilution is
given by S = Q/Q o = (Ua A j )/Q o = (Ua π B 2 )/Q o , where Q = local jet volume flux;
Q o = source discharge flow; and A j = jet cross-sectional area = Av (B/Bv )2 by simple
scaling.

5 Concluding remarks

A comprehensive experimental investigation on a 60o dense jet in perpendicular crossflow


has been conducted. The detailed tracer concentration field in the cross-section of the bent-
over jet is measured for a wide range of jet densimetric Froude number and jet to ambient
velocity ratios (Fr = 5 − 40, K = 2.4 − 83). The experiments enable a systematic analysis
of the entire range of jet-dominated to crossflow-dominated cases. The mixing behavior is

123
Environ Fluid Mech

(a)

(b)

Fig. 15 Comparison between measured and predicted jet dilution for experiments with Ur Fr > 0.8 a Smin b
Savg

elucidated by employing length scale analysis and a detailed interpretation of scalar field
using the Lagrangian model JETLAG. Both the theory and experiments demonstrate the
following:
1. The crossflow Froude number F = Ur Fr ∼ l M /lm is the appropriate scale in classifying
weak and strong crossflows for a dense jet. It expresses the relative location where the

123
Environ Fluid Mech

flow behaves like a jet and where it behaves like an advected puff. As the terminal rise
scales with l M in a weak crossflow, it can be expected that for Ur Fr << 1 i.e. lm >>
l M the jet is hardly bent-over and behaves like an advected-jet before terminal rise. The
asymptotic regime transition can therefore be expected to occur when Ur Fr ∼ 1 and
from experiments this is determined as 0.8. For F < 0.8, the mixing is jet-dominated
and governed by shear entrainment; significant detrainment occurs and the maximum
height of rise is under-predicted as in the case of a dense jet in stagnant fluid.
2. For F ≥ 0.8 the dense jet becomes significantly bent-over during the its ascent phase;
the jet mixing is dominated by vortex entrainment. For F ≥ 2, the detrainment ceases to
have any effect on the jet behavior. The jet trajectory in both the horizontal momentum
and buoyancy planes are well predicted by the JETLAG model. Both the maximum
terminal rise and horizontal lateral penetration follow a Z max ∼ (F)−1/2 in the crossflow-
dominated regime. The initial dilution at terminal rise follows a S ∼ (F)1/3 dependence.
Both the terminal rise and the initial dilution for the inclined jet in perpendicular crossflow
are smaller than those of a corresponding vertical jet.
3. Despite the under-prediction of terminal rise, the jet dilution at a large number of cross-
sections covering the ascent and descent of the dense jet are well-predicted by the model.
The longer jet trajectory and reduced negative buoyancy due to detrainment appear to
be two compensating factors affecting dilution.
4. As in a positively buoyant jet, the scalar field in a bent-over dense jet in crossflow is
kidney-shaped. Depending on the relative strength of initial jet momentum to ambient
velocity, the structure changes from a plateau of high concentration to the distinct double-
peak distribution. The bifurcation is most apparent before terminal rise and degenerates
into a more circular distribution in the descent phase where the jet turns into a descending
advected line thermal. In the terminology of asymptotic flow regimes [10], as a dense
jet rises to a maximum height by virtue of its momentum, and falls onto the bottom
boundary by virtue of its negative buoyancy, the mixing undergoes stages that resemble
an advected jet (MDNF), advected line puff (bent over jet, MDFF) and advected line
thermal (bent over plume upside down, BDFF).
It should be noted that the dilution on impact with the bottom boundary depends on the
interaction of the impinging jet and the bottom spreading layer. The prediction of the impact
dilution requires a proper treatment of the coupling of near field and the intermediate field,
and will be separately reported.

Acknowledgment This research is supported by a Grant from the Hong Kong Research Grants Council
(HKU 713908E)

References

1. Latteman S, Hopner T (2008) Environmental impact and impact assessment of seawater desalination.
Desalination 220:1–15
2. Abessi O, Saeedi M, Bleninger T, Davidson M (2012) Surface discharge of negatively buoyant effluent
in unstratified stagnant water. J Hydro-Environ Res 6(3):181–193
3. Zeitoun MA, Reid RO, McHihenny WF, Mitchell TM (1970) Model studies of outfall system for desali-
nation plants, Research and Development Progress Rep. 804, Office of saline water, U.S. Dept. of Interior,
Washington, D.C.
4. Roberts PJW, Toms G (1987) Inclined dense jets in flowing current. J Hydraul Eng ASCE 113(3):323–341
5. Roberts PJW, Daviero G (1997) Mixing in inclined dense jets. J Hydraul Eng ASCE 123(8):693–699
6. Kikkert GA, Davidson MJ, Nokes RI (2007) Inclined negatively buoyant discharges. J Hydraul Eng ASCE
133(5):545–554

123
Environ Fluid Mech

7. Shao D, Law AWK (2010) Mixing and boundary interaction of 30deg and 45deg inclined dense jets.
Environ Fluid Mech 10(5):521–553
8. Papakonstantis IG, Christodoulou GC, Papanicolaou PN (2011a) Inclined negatively buoyant jets 1:
geometrical characteristics. J Hyd Res 49(1):3–12
9. Lai CCK, Lee JHW (2012) Mixing of inclined dense jets in stationary ambient. J Hydro-Environ Res
IAHR-APD 6(1):9–28
10. Fischer HB, List EJ, Koh RCY, Imberger J, Brooks NH (1979) Mixing in inland and coastal waters.
Academic Press, New York
11. Lee JHW, Chu VH (2003) Turbulent jets and plumes—a Lagrangian approach. Kluwer Academic Pub-
lisher, The Netherlands
12. Ayoub GM (1973) Test results on buoyant jets injected horizontally in a cross flowing stream. Water Air
Soil Pollut 2(4):409–426
13. Cheung V (1991) Mixing of a round buoyant jet in a current. PhD Thesis, The University of Hong Kong,
p 202
14. Tong SS, Stolzenbach KD (1979) Submerged discharges of dense effluent, Rept. No. 243, Ralph M.
Parsons Lab., MIT, Cambridge, MA.
15. Lindberg WR (1994) Experiments on negatively buoyant jets, with and without crossflow. NATO advanced
research workshop on “Recent advances in jets and plumes” NATO ASI Series E. Kluwer Academic,
Viana do Castelo, Portugal 255:131–145
16. Kikkert GA, Davidson MJ, Nokes RI (2010) Buoyant jets with three-dimensional trajectories. J. Hydraul
Res 48(3):292–301
17. Gungor E, Roberts PJW (2009) Experimental studies on vertical dense jets in a flowing current. J Hydraul
Eng ASCE 135(11):935–948
18. Pincince AB, List EJ (1973) Disposal of brine into an estuary. J Water Pollut Control Fed 45(11):2335–
2344
19. Palomar P, Lara JL, Losada IJ (2012) Near field brine discharge modeling part 2: validation of commercial
tools. Desalination 290:28–42
20. Jirka GH (2004) Integral model for turbulent buoyant jets in unbounded stratified flows. Part 1. The single
round jets. Environ Fluid Mech 4:1–56
21. Frick WE (2004) Visual plume mixing zone modelling software. Environ Model Softw 19:645–654
22. Lee JHW, Cheung V (1990) Generalized Lagrangian model for buoyant jets in current. J Environ Eng
ASCE 116(6):1085–1106
23. Lee JHW, Cheung V, Wang WP, Cheung SKB (2000) Lagrangian modeling and visualization of rosette
outfall plumes. Proc. Hydroinformatics 2000, University of Iowa, July 23–27, 2000 (CDROM).
24. Lee JHW (2012) The hunter rouse lecture—mixing of multiple buoyant jets. J Hydraul Eng ASCE
138(12):1008–1021
25. Lai ACH, Yu D, Lee JHW (2011) Mixing of a rosette jet group in a crossflow. J Hydraul Eng ASCE
137(8):787–803
26. Lai CCK (2009) Mixing of inclined dense jets. MPhil Thesis, The University of Hong Kong, p 132
27. Chu PCK (1996) Mixing of advected line puffs. PhD Thesis, The University of Hong Kong, p 222

123

You might also like