You are on page 1of 6

On the development of large-scale structures of a jet normal

to a cross ow
T. T. Lim, T. H. New, and S. C. Luo
Department of Mechanical Engineering, National University of Singapore, 10 Kent Ridge Crescent,
Singapore 119260
Received 25 February 2000; accepted 6 December 2000
It is well known that vortex rings are the dominant ow structures in the near eld of a free jet, and
this has led many researchers to believe that they also occur in a jet in cross ow JICF. Previous
studies have postulated that these vortex rings deform and fold as they convect downstream, which
culminates in the formation of vortex loops at both the upstream and the lee-side of the jet column.
In this paper, we take a fresh look at the vortical structures of JICF in water by releasing dye at
strategic locations around the jet exit. The results show that there is no evidence of ring vortices in
JICF, and the postulation that vortex loops are formed from the folding of the vortex rings does not
reect the actual ow behavior. The presence of a counter-rotating vortex pair CVP at the jet exit
is found to inhibit the formation of the vortex rings. Instead, vortex loops are formed directly from
the deformation of the cylindrical vortex sheet or jet column, without going through the vortex rings,
in a process similar to the buoyant jet and wake structures studied by Perry and Lim J. Fluid Mech.
88, 451 1978. 2001 American Institute of Physics. DOI: 10.1063/1.1347960
INTRODUCTION
The jet in cross ow JICF is relevant to a wide variety
of technological applications, including plume dispersion,
control of missiles, chimney ow and V/STOL aerodynam-
ics. In the past few decades, our knowledge of JICF has been
increasing steadily due to the combined efforts from physical
experiments,
119
numerical simulations,
2027
and theoretical
analyses.
28,29
For an overview of the recent developments in
this area, the reader is referred to the review article by
Margason.
30
It has been recognized that when a jet is discharged nor-
mal to a cross ow, there is a complex interaction between
the jet and the cross ow, leading to the formation of four
vortical structures. They are a the leading-edge or upstream
vortices which form along the jet/cross ow interface above
the jet exit, b the horseshoe vortices which appear upstream
of the jet exit, c the wake vortices which occur downstream
of the jet exit, and d the counter-rotating vortex pair CVP
which appear in the far eld. These vortical structures have
been studied extensively by many researchers. For example,
Kelso et al.,
15
Sykes et al.,
24
and Coelho and Hunt
31
studied
the vortex dynamics of the exiting jet; Moussa et al.,
7
Fric
and Roshko,
14
Kelso et al.,
15
McMachon et al.
32
and Eiff and
Keffer
33
investigated the wake vortex system; Krothapalli
et al.
34
and Kelso and Smits
35
examined the horseshoe vor-
tex system, and Kamotani and Greber,
4
Moussa et al.,
7
Scorer,
36
and Fearn and Weston
37
studied the CVP. Despite
the considerable progress made in these areas, many aspects
of the ow eld remain poorly understood. Among those is
the evolution of the jet shear layer vortices. In an attempt to
interpret the ow physics, Andreopoulos,
9
Kelso et al.,
15
Sykes et al.,
24
and Lim et al.
38
have used vortex rings to
model the ow structures. They postulated that the vortex
rings deform and stretch as they propagate downstream,
leading to the eventual formation of the loop vortices. How-
ever, our recent studies New et al.
39
on a different aspect of
JICF do not show any evidence of vortex rings in the ow,
and that the loop vortices do not really originate from the
vortex rings. This unexpected nding prompted us to make a
detour from our original plan, and to pursue this study with
the aim of resolving the above anomaly.
EXPERIMENTAL SETUP
The experiments were conducted in a recirculating water
channel in the Fluid Mechanics Laboratory at the National
University of Singapore. The channel has a Plexiglas test
section that measures 0.4 m 0.4 m, and is powered by a
variable-speed ac motor driving a centrifugal pump on the
return leg of the ow circuit. The pump, together with a
buttery valve located downstream of the pump exit, pro-
vides a wide range of cross ow velocities in the section.
Before the ow entered the test section, it was passed
through a honeycomb, three ne screens, and a 4:1 contrac-
tion section. This was to ensure that the turbulence level in
the cross ow was kept to a minimum, and the boundary
layer on the test section oor remained laminar.
The jet nozzle comprised of a 14:1 circular contraction
chamber with an exit internal diameter (D) of 32.47 mm,
and was mounted ushed with the main water channel oor.
To create the transverse jet, a gravity-feed technique utilizing
an overhead tank was used. Here, a small constant-speed
pump was used to supply water from the tank downstream of
the test section to the overhead tank located 1.5 m above the
test section oor. The tank had an overow compartment
which allowed the overow to be directed back to the main
water channel just upstream of the honeycomb. The jet ve-
PHYSICS OF FLUIDS VOLUME 13, NUMBER 3 MARCH 2001
770 1070-6631/2001/13(3)/770/6/$18.00 2001 American Institute of Physics
Downloaded 25 Jun 2008 to 216.54.44.227. Redistribution subject to AIP license or copyright; see http://pof.aip.org/pof/copyright.jsp
locity was controlled by a ball-valve located some distance
upstream of the jet exit and was measured by using an elec-
tromagnetic ow meter. Another larger capacity ow meter
was used to measure the cross ow velocity.
To visualize the ow, dye/alcohol mixture with a spe-
cic gravity of approximately one was released slowly onto
the inner surface of the jet nozzle, through a 1 mm wide
circumferential slit located 2.5D upstream of the nozzle exit.
In some runs, uorescein disodium dye was used in conjunc-
tion of a 5 W argon ion laser.
RESULTS AND DISCUSSION
Since the aim of the present study is to resolve the issue
of the vortex structures in JICF, we pay particular attention
to the generation of these vortices near the jet exit. Figure 1
shows a sequence of photographs illustrating how the cylin-
drical shear layer issuing from the nozzle deformed and
rolled up into concentrated vortices. Here, the velocity ratio
(VRV
j
/V

) was approximately 4.6, and the Reynolds


number ReV
j
D/) was about 1600, where V
j
and V

are
the jet and the cross ow velocity, respectively, D is the
nozzle diameter and is the kinematic viscosity of water.
The scale of the pictures can be inferred through the cylinder
diameter (D) of 32.47 mm. These photographs were ob-
tained by releasing dye uniformly around the circumferential
slit mentioned earlier. In all cases, the dye concentration is
kept to a minimum so that the salient features of the ow are
not unduly obscured by the thickness of the dye. In many of
the previous studies, a high concentration of dye or smoke
made the interpretation of the ow difcult, and in some
cases led to misinterpretation of the ow eld, particularly
on the lee-side of the jet. In this study, since the dye was
released at the location where the vorticity was generated, it
is reasonable to assume that the dye marked the vorticity, at
least during the early stages of the ow development. Hence,
the behavior of the dye sheet depicted in Fig. 1 gives a good
indication of the behavior of the vortex sheet. However, one
should be cautioned that since dye generally diffuses much
slower than vorticity, and over a long period of time, the
region occupied by the dye may no longer give a true indi-
cation of the region of vorticity see Lim
40
.
Inspection of Fig. 1 clearly shows that as the cylindrical
vortex sheet or shear layer emerged from the nozzle, it
underwent three distinct folding processes. One process led
to the formation of the CVP on the sides of the jet column,
and the other two resulted in the formation of two rows of
vortices; one at the upstream side identied as A, in Figs. 1
and 2, and the other at the lee-side of the jet identied as
B. The latter two vortices are henceforth referred to as the
FIG. 1. A sequence of photographs
showing how the folding of the cylin-
drical shear layer or vortex sheet
from the jet nozzle leads to the even-
tual formation of the counter-rotating
vortex pair, and the upstream and lee-
side vortices indicated as A and B, re-
spectively. In a, the time has arbi-
trarily been set to 0.00 s. It is
important to note that although the
ow structures look complicated, the
original uid leaving the nozzle re-
mains in its original cylindrical bound-
ary. To avoid overcrowding of the g-
ure, the sign of circulation of the
upstream and the lee-side vortices are
shown only in d as semi-circular ar-
rows.
FIG. 2. Sectional view of the vortex structures in the center plane of a jet
issuing normal to a cross ow. The photograph is obtained by premixing the
jet uid with the uorescein dye, and then illuminated with a narrow sheet
of laser light. Note that the vortices A and B correspond approximately to
those in Fig. 1d. The counter-rotating vortex pair is not visible in the
photograph because it is out of the illumination plane. This picture also
clearly shows that the original uid leaving the nozzle remains in the cylin-
drical boundary.
771 Phys. Fluids, Vol. 13, No. 3, March 2001 Large-scale structures of a jet normal to a cross ow
Downloaded 25 Jun 2008 to 216.54.44.227. Redistribution subject to AIP license or copyright; see http://pof.aip.org/pof/copyright.jsp
upstream vortices and the lee-side vortices, respectively.
These vortices resemble a daisy chain of interlocking
loops, which are similar to the buoyant jet structures ob-
served by Perry and Lim.
41
What is interesting about these
loop vortices is that they were not produced by the bending
or folding of the vortex rings, as some researchers were led
to believe. In fact, there is no evidence of vortex rings in any
of our studies. This observation has also been made by Yuan
et al.
27
in their large eddy simulation of a round jet in cross
ow.
To better understand how these loop vortices evolved,
video images of the ow eld were replayed in slow motion.
They show that as soon as the cylindrical vortex sheet see
also Fig. 1 emerged from the nozzle, it immediately folded
up on its edges to form a CVP. This nding is consistence
with the experimental observation of Kelso et al.,
15
and the
large-eddy simulation of Yuan et al.
27
Furthermore, it was
found that the CVP played a signicant role in preventing
the cylindrical vortex sheet, which contains circular vor-
tex lines, from rolling up into vortex rings. This is in contrast
to the free jet, where the absence of CVP allows the vortex
sheet to roll up axisymmetrically into vortex rings. In the
case of JICF, the vortex sheet can roll up freely only at the
upstream and the lee-side of the jet column since the CVP
suppresses similar roll up at the two sides. This scenario
gives rise to two rows of loop vortices separated by the CVP
at the sides. Close examination shows that this formation
process is similar to the buoyant jet structures observed by
Perry and Lim,
41
although the details are different. During
the folding of the vortex sheet, the uid that originated from
the nozzle remained within the cylindrical boundary, and
there was no indication of holes or openings in the original
vortex sheet. This feature is clearly illustrated in Fig. 2,
where the jet uid, which had been premixed with uores-
FIG. 3. Authors interpretation of the
nally developed vortex structures of
JICF. a The sketch shows how the
arms of both the upstream and the
lee-side vortex loops are merged with
the counter-rotating vortex pair. b
Cross sectional views of a taken at
various streamwise distances. Note
their close resemblance with the laser
cross sections of the jet depicted in
Fig. 5.
FIG. 4. Detailed sketches of the proposed model. a The sketch shows how
the vortex loops give rise to the resultant Section B-B. b Section E-E along
the deected jet centerline in the streamwise direction. This sketch repre-
sents the laser cross section of JICF depicted in Fig. 2.
772 Phys. Fluids, Vol. 13, No. 3, March 2001 Lim, New, and Luo
Downloaded 25 Jun 2008 to 216.54.44.227. Redistribution subject to AIP license or copyright; see http://pof.aip.org/pof/copyright.jsp
cein dye, was illuminated with a thin laser sheet in the jet
center plane. Here, both the upstream and the lee-side vorti-
ces can be clearly seen while the CVP is not visible because
it was outside the laser plane. It is worth noting at this point
that although the sectional views such as the one shown in
Fig. 2 provide valuable information about the internal struc-
tures of the ow eld, one should be mindful that it is dif-
cult to deduce three-dimensional ow structures from study-
ing the sectional view alone. For example, the mushroom-
shape vortices in Fig. 2 may lead one to misinterpret them
as part of the same vortex rings, however, as the broad band
pictures in Fig. 1 clearly show, they are actually part of the
two separate vortex loops.
As the vortex loops convected downstream, their side
arms were swept downstream by the cross ow and sub-
sequently paired up with the CVP see Fig. 1d. During
merging, the side arms were stretched by the CVP, and
this resulted in realignment of the vorticity with that of the
CVP. Although not shown in the photographs, we suspect
that the mutual interaction between the loop vortices far
downstream was partly responsible for causing the vortices
to undergo instability and became turbulent, leaving behind
only the CVP with ne scale motion superimposed on it.
Figure 3 shows our interpretation of the nally developed
ow structure. Note how the CVP is initiated at the sides of
the cylindrical vortex sheet and how the vortex loops are
formed from the folding of the vortex sheet. The break in
the gure indicates that section D-D is taken far away from
nozzle exit. For clarity, the ne-scale structures are not
shown in the gure. In fact, the best way to understand the
nal structure is to make a model using wire gauze tube, a
technique used successfully by Perry and Lim
41
to illustrate
how buoyant wake and jet structures are formed. By continu-
ing to wrap the sides of the wire gauze tube, one can obtain
the CVP. Similarly, by continuing to fold the upstream edge
and the lee-side of the wire gauze tube, one can obtain the
two rows of vortex loops as shown in Fig. 3a. Indeed, a
wire gauze model was built and used as an aid to visualize
and construct the 3D view of the proposed model. Figure
4a shows the 3D view of the proposed jet structures, show-
ing how the upstream and lee side vortex loops may have
interacted to produce Section B-B via their side arms.
And Fig. 4b shows the cross section E-E along the de-
ected jet centerline, showing the same cross sectional view
in Fig. 2.
To demonstrate that the proposed model is consistent
with the experiment, laser cross sections of the actual jet
structures normal to the jet axis see Fig. 5 are compared
with the corresponding cross sections of the model see Fig.
3b, and the agreement between them is good. It is of in-
terest to note that in a recent large-eddy simulation of a
round jet in crossow by Yuan et al.,
27
it was found that the
FIG. 5. Laser cross sections of a jet
taken with the laser plane perpendicu-
lar to the jet axis. The symbol s de-
notes the distance measured from the
oor and along the jet trajectory, and
D is the nozzle diameter. In b, the
side arms of the upstream loop have
already merged with the CVP while
those of the lee-side loop are still dis-
tinct. The latter also merged with the
CVP further downstream as can be
seen in c and d.
773 Phys. Fluids, Vol. 13, No. 3, March 2001 Large-scale structures of a jet normal to a cross ow
Downloaded 25 Jun 2008 to 216.54.44.227. Redistribution subject to AIP license or copyright; see http://pof.aip.org/pof/copyright.jsp
near eld of the jet consists of three different vortical struc-
tures, namely spanwise rollers, hanging vortices, and vertical
streaks. These structures are presented in terms of the iso-
surface of vorticity, as shown in Fig. 5a of their paper.
While it is clear that their spanwise rollers correspond to our
upstream and lee-side vortices, it is not so clear as to how the
hanging vortices and the vertical streaks are formed. The
uncertainty arises because their isosurface of the vorticity
plot does not give a clear picture of how the three vortex
structures are interconnected. But the comment by Yuan
et al.
27
that the hanging vortices undergo vortex breakdown
similar to that reported by Kelso et al.
15
leads us to suspect
that the hanging vortices are likely to be the side arms of the
upstream vortex loops before it merges with the CVP. As for
the vertical streaks, we believe that they are the side arms of
both the upstream and lee-side vortex loop vortices, which
have been stretched by the CVP during merging. The stretch-
ing causes the vorticity of the side arms to realign with that
of the CVP. Overall, our observation agrees well with that of
Yuan et al.
27
An important feature of the model depicted in Figs. 3
and 4 is that the upstream loops can exist without the lee-side
loops. This is in contrast to the model using vortex rings,
which suggests that for every upstream vortex loop, there
must be a corresponding one on the lee-side. However, from
our experimental results and the results of Kelso et al.,
15
single-sided vortex loops are known to exist in JICF when
the velocity ratio is low, typically less than one. This is
clearly shown in Fig. 6. Here, one can see that the vortex
loops are pointed downstream rather than upstream as is
commonly observed in high velocity ratio cases. The orien-
tation of the vortex loops suggests that, at low velocity ratio,
the jet structures have transformed into wake structures
see Perry and Lim
41
. Under this condition, the buoyant
wake model of Perry and Lim
41
is equally applicable here.
Other examples in nature where single-sided loops are
known to occur include the smoke structures issuing from
factory chimney in cross ow, and a cigarette smoke in a
slight draft see Perry and Lim
41
. However, one should be
aware that, for these two cases, the buoyancy of the hot
smoke does play an important role in determining the nal
structures of the ow. Nevertheless, these examples provide
further support to our model that the large-scale structures of
JICF consists essentially of loop vortices, which are not
caused by the folding of the vortex rings. We believe that
vortex rings will not form naturally in JICF unless they
are deliberately red into a cross ow see Sykes et al.,
24
Chang and Vakili,
42
and Lim et al.
38
, or when the velocity
ratio is close to innity i.e., a free jet. Far away from these
conditions, the cross ow will cause the circular vortex lines
surrounding the cylindrical vortex sheet to tilt, and the vortex
sheet to roll up into loop vortices.
CONCLUSION
We have shown that the large-scale structures of jet dis-
charging normally into a cross ow consist of loop vortices
which are not caused by the folding of the ring vortices as
have been previously thought otherwise by many research-
ers. Our proposed model is consistent with the experimental
observations and the large-eddy simulation of Yuan et al.
27
It
provides a more realistic representation of the large-scale
motion of JICF for all velocity ratios higher than 1. For cases
where the velocity ratio is lower than one, the model pro-
posed by Perry and Lim
41
for the buoyant coowing wakes
gives a good representation.
ACKNOWLEDGMENT
The authors would like to thank the National University
of Singapore for the support of this project under RP 950638.
1
J. F. Keffer and W. D. Baines, The round turbulent jet in a crosswind,
J. Fluid Mech. 15, 481 1963.
2
B. D. Pratte and M. Baines, Proles of the round turbulent jet in a
crossow J. Hydraul. Div., Am. Soc. Civ. Eng. 92, 53 1967.
3
N. A. Durando, Vortices induced in a jet by a subsonic crossow,
AIAA J. 9, 325 1971.
4
Y. Kamotani and I. Greber, Experiments on a turbulent jet in a cross
ow, AIAA J. 10, 1425 1972.
5
P. Chassaing, J. George, A. Claria, and F. Sananes, Physical character-
istics of subsonic jets in a cross stream, J. Fluid Mech. 62, 41 1974.
6
G. Bergeles, A. D. Gosman, and B. E. Launder, The near-eld character
of a jet discharged through a wall at 90 degrees to a main stream, Trans.
ASME: J. Heat Trans. 98, 373 1976.
7
Z. M. Moussa, J. W. Trischka, and S. Eskinazi, The near eld in the
mixing of a round jet with a cross-stream, J. Fluid Mech. 80, 49 1977.
8
D. Crabb, D. F. G. Durao, and J. H. Whitelaw, A round jet normal to a
crossow, J. Fluids Eng. 103, 142 1981.
9
J. Andreopoulos, On the structure of jets in a crossow, J. Fluid Mech.
157, 163 1985.
10
N. Rajarantnam and T. Gangadharaiah, Vortex structure of circular jets
in crossow, J. Wind Eng. and Industrial Aero 12, 155 1983.
11
J. Andreopoulos and W. Rodi, Experimental investigation of jets in a
cross ow, J. Fluid Mech. 138, 93 1984.
FIG. 6. Photographs showing the
wake structures from the nozzle at the
velocity ratio of 1. a Side view. b
Plan view taken at a different instance.
Note that the vortex loops are pointed
downstream see New, Ref. 43.
774 Phys. Fluids, Vol. 13, No. 3, March 2001 Lim, New, and Luo
Downloaded 25 Jun 2008 to 216.54.44.227. Redistribution subject to AIP license or copyright; see http://pof.aip.org/pof/copyright.jsp
12
J. E. Broadwell and R. E. Briedenthal, Structure and mixing of a trans-
verse jet in incompressible ow, J. Fluid Mech. 148, 405 1984.
13
J. W. Wu, A. D. Vakili, and F. M. Yu, Investigation of the interacting
ow of nonsymmetric jets in crossow, AIAA J. 26, 940 1988.
14
T. F. Fric and A. Roshko, Vortical structure in the wake of a transverse
jet, J. Fluid Mech. 217, 1 1994.
15
R. M. Kelso, T. T. Lim, and A. E. Perry, An experimental study of round
jet in cross ow, J. Fluid Mech. 306, 111 1996.
16
B. A. Haven and M. Kurosaka, Kidney and anti-kidney vortices in cross
ow jets, J. Fluid Mech. 352, 27 1997.
17
L.-E. Brizzi, E. Foucault, and J.-L. Bousgarbie`s, Vortices organization in
the near eld of a jet issuing normally into a crossow, J. Flow Vis. &
Image Processing 5, 17 1998.
18
S. H. Smith and M. G. Mungal, Mixing, structure and scaling of the jet
in cross ow, J. Fluid Mech. 357, 83 1998.
19
J. N. Blanchard, Y. Brunet, and A. Merlen, Inuence of a counter rotat-
ing vortex pair on the stability of a jet in a cross ow: An experimental
study by ow visualizations, Exp. Fluids 26, 63 1999.
20
S. V. Patankar, D. K. Basu, and S. A. Alpay, Prediction of the three-
dimensional velocity eld of a deected turbulent jet, J. Fluid Eng. 99,
758 1977.
21
A. R. Karagozian, An analytical model for the vorticity associated with
a transverse jet, AIAA J. 24, 429 1986.
22
D. J. Needham, N. Riley, and J. H. B. Smith, A jet in cross-ow, J.
Fluid Mech. 188, 159 1988.
23
D. J. Needham, N. Riley, C. C. Lytton, and J. H. B. Smith, A jet in
cross-ow. Part 2, J. Fluid Mech. 211, 515 1990.
24
R. I. Sykes, W. S. Lewellen, and S. F. Parker, On the vorticity dynamics
of a turbulent jet in a crossow, J. Fluid Mech. 168, 393 1986.
25
R. W. Claus and S. P. Vanka, Multigrid calculations of a jet in cross-
ow, J. Propul. Power 8, 425 1992.
26
M. Rudman, Simulation of the near eld of a jet in a cross ow, Exp.
Therm. Fluid Sci. 12, 134 1996.
27
L. L. Yuan, R. L. Street, and J. H. Ferziger, Large-eddy simulations of a
round jet in crossow, J. Fluid Mech. 379, 71 1999.
28
R. H. Nunn, Vorticity growth and decay in the jet in crossow, AIAA
J. 23, 473 1985.
29
J. H. W. Lee, L. Li, and V. Cheung, Semianalytical self-similar solution
of bent-over jet in cross-ow, J. Eng. Mech. Div., Am. Soc. Civ. Eng.
125, 733 1999.
30
R. J. Margason, Fifty years of jet in cross ow research, AGARD-CP-
534, 1-1141 1993.
31
S. L. V. Coelho and J. C. R. Hunt, The dynamics of the near eld of
strong jets in crossows, J. Fluid Mech. 200, 95 1989.
32
H. M. McMahon, D. D. Hester, and J. G. Palfrey, Vortex shedding from
a turbulent jet in a cross-wind, J. Fluid Mech. 48, 73 1971.
33
O. S. Eiff and J. F. Keffer, On the structures in the near-wake region of
an elevated turbulent jet in a crossow, J. Fluid Mech. 333, 161 1997.
34
A. Krothapali, L. Lourenco, and J. M. Buchlin, Separated ow upstream
of a jet in a crossow, AIAA J. 28, 414 1990.
35
R. M. Kelso and A. J. Smits, Horseshoe vortex systems resulting from
the interaction between a laminar boundary layer and a transverse jet,
Phys. Fluids 7, 153 1995.
36
R. S. Scorer, Natural Aerodynamics Pergamon, New York, 1958.
37
R. Fearn and R. P. Weston, Vorticity associated with a jet in a cross
ow, AIAA J. 12, 1666 1974.
38
T. T. Lim, R. M. Kelso, and A. E. Perry, A visual study of vortex rings
red transversely into a cross-ow, 13th Australasian Fluid Mechanics
Conference, Monash University, Australia, Vol. 2, pp. 961964 1998.
39
T. H. New, T. T. Lim, and S. C. Luo, On the effects of velocity proles
on the topological structures of a jet in cross ow, TSFP-1 Proceedings,
pp. 647652 1999.
40
T. T. Lim, Dye and smoke visualization in Flow Visualization: Tech-
niques and Examples, edited by A. J. Smits and T. T. Lim Imperial
College Press, London, 2000, pp. 4369.
41
A. E. Perry and T. T. Lim, Coherent structures in coowing jets and
wakes, J. Fluid Mech. 88, 451 1978.
42
Y. K. Chang and A. D. Vakili, Dynamics of vortex rings in cross-ow,
Phys. Fluids 7, 1583 1995.
43
T. H. New, A topological study of a jet in cross ow, B. Eng thesis,
National University of Singapore 1998.
775 Phys. Fluids, Vol. 13, No. 3, March 2001 Large-scale structures of a jet normal to a cross ow
Downloaded 25 Jun 2008 to 216.54.44.227. Redistribution subject to AIP license or copyright; see http://pof.aip.org/pof/copyright.jsp

You might also like