J . Fluid Mech. (1978), 001. 84, part 4, p p .

641-656 64 1
Printed in Great Britain

Similarity in the initial region of annular jets:

three configurations
By N. W. M. K O AND W. T.C H A N
Department of Mechanical Engineering, University of Hong Kong

(Received 12 May 1977)

This paper describes part of a detailed study of the initial region of three annular
jets. The configurations are the basic one, without any bullet in the centre, and those
with a conical and an elliptical bullet. From the mean velocity and turbulence
intensity measurements the initial region can be divided into the initial merging, the
intermediate and the fully merged zones. Within these three zones similarity of both
the mean velocity and the turbulence intensity profiles has been found. The similarity
curves are compared with those for a single jet.

1. Introduction
The initial region of an annular air jet discharging into stationary air has been
only scantily investigated. The early work of Miller & Comings (1960) concerned the
flow structure of dual plane jets issuing into stationary air. Their measurements
showed the subatmospheric pressure region due to entrainment in the centre. This
region was also responsible for the convergence of the dual jets towards the central
axis. Associated with the subatmospheric pressure region was the recirculating vortex
which recycled air in this region. The measurements of an annular jet by Chigier &
Beer (1964) were part of an investigation into the flow structure of coaxial or con-
centric jets. Their mean velocity and static pressure measurements also established
the presence of a subatmospheric region and a vortex in the centre of the annular
jet.
The recent investigation of Ayukawa & Shakouchi (1976) involved a two-dimen-
sional jet attaching to a wall, which also formed a subatmospheric pressure region
and the associated vortex. The work mainly concerned the periodic pressure fluctua-
tions within the subatmospheric pressure region.
Though the above-mentioned work included mean velocity and turbulence in-
tensity measurements, an investigation of similarity of these profiles has not been
attempted. I n this respect, the purpose of the present investigation was to obtain
information about the flow characteristics within the initial region of three annular
jets: a basic annular jet, a conical jet and an ellipsoidal jet. This paper presents the
first part of the investigation, which concerns the mean velocity, turbulence in-
tensity and their similarity. The latter parts will concern the spectral and correlation
measurements of both the turbulence and the pressure fluctuations, which are
associated with the recirculating vortex in the centre and the vortices in the outer
mixing region.
From the similarity curves of the mean velocity and turbulence intensity, the
initial region of the annular jets can be separated into the initial merging, the
22 F L M 84
642 N . W . M . KO and W . T.Chun

D,, D , -t-----t---- --t

#@ Axis of annular jet Y
I
A

I
I Initial
’
I
merging zone I
Intermediate zone
I
Fuliy merged zone
4

FIGURE
1. Schematic profile of initial region of annular jet.

intermediate and the fully merged zones. An attempt will be made to compare the
similarity curves obtained from this investigation with t,he available results for a
single jet.

2. Apparatus
The experiments were carried out within annular jets of three configurations.
A diagram of the arrangement of the nozzles of the three configurations is shown in
figure 1. Basically, a t the nozzle exit the circular annular nozzle has a n outer dia-
meter Do of 6.2 cm and an inner diameter Di of 2-8 em. The area contraction ratio of
bhe nozzles was 15 : 1. The first configuration was the basic ‘ annular ’ nozzle, without
any protrusion or bullet at the nozzle exit. With this type of configuration an internal
recirculating vortex of toroidal form was found (Chigier & Beer 1964). The other two
configurations included a bullet-like protrusion. The intention was t o eliminate the
internal recirulating vortex found in the basic annular configuration. At the exit the
diameter of the bullet was the same as the inner diameter Diof the annular jet so
that no discontinuity of flow occurred. The profile of the bullet in the second configura-
tion was ‘ conical ’, while in the third it was ‘ellipsoidal ’. The length of both the conical
and the ellipsoidal bullet was 1.5 D,,.This length was adopted so that the internal
recirculating vortex was estimated. For simplification the jet with the basic con-
figuration is simply termed an ‘annular’ jet while t,he other two are termed ‘conical’
and ‘ ellipsoidal ’.
The nozzle was supplied with dry, oil-free, high-pressure air from an air reservoir.
The nozzle has a Burgess-type silencer as its settling chamber. The noise from the
upstream control valve was mostly absorbed by a resonator-type silencer before it
arrived a t the nozzle exit. Further absorption of the inherent noise was achieved by
passing the air from the silencer through a layer of 2 mm polyurethane material. In
this way, besides having a lower inherited background level, t,he flow was smoothed,
 Siirtilcirity in crn,lular j e t s G43
resulting in a lower local turbulence intensity a t the exit. I n the velocity range of this
investigation the inherited overall background pressure level a t the exit was 99 dB.
The background turbulence intensity level was 0.22 % of the mean exit velocity.
The inclusion of the conical or ellipsoidal bullet did not significantly affect the back-
ground conditions a t bhe exit.
The hot-wire anemometer used was of constant-temperature type with a linearized
output (Davies & Davies 1966). Only a single wire was used in the present work.
This wire, consisting of tungsten, had a diameter of 5 x m and a length of 2 mm.
Its operating resistance was about 15 Q.
The domain of investigation was mostly confined t o the first seven outer diameters
Do downstream of the nozzle exit. However, the small dimensions of and the compli-
cated flow pattern in the inner mixing region made it difficult to investigate this
region accurately. Thus the main effort was expended on the outer mixing region.
Nevertheless, limited measurements have been obtained in the inner mixing region.
Because of the difficulty encountered in this mixing region, no attempt was made to
correct the single-wire results for the deviation in the flow direction from axial. The
mean exit velocity in the investigation was 50 ms-l. Though another set of measure-
ments, for a n exit velocity of 30 ms-l, has also been obtained for comparison, it will
not be presented.

3. Mean velocity
Domain
The mean velocity profiles for an annular jet have been obtained by Chigier & Beer
(1964). They simply separated the annular jet into the merging zone and the com-
bined jet. This division did not fully describe the flow structures in the initial region.
Thus a division similar to that for coaxial jets (Kwan & KO 1976) is adopted. The
annular jets are divided into the initial merging zone, the intermediate zone and the
fully merged zone (figure 1). The zone which is nearest to the nozzle exit and ends
roughly a t the plane where the potential core disappears is called the initial merging
zone. It is within this zone that the internal recirculating vortex is found in the annular
potential core for the annular jet and the bullet is found for the conical and ellipsoidal
jet. As in the case of single and coaxial jets, the mean velocity within the annular
potential core is the same as the jet exit velocity Do.
Immediately downstream of the initial merging zone is the intermediate zone. It is
within this zone that mixing of the flows from the annular potential core and from the
outer mixing region occurs. It is also within the intermediate zone that the axis of
the annular potential core and the high mean velocity associated with the potential
core merge a t the central axis of the nozzle. The extent of this intermediate zone,
as will be discussed later, is about three outer diameters Do.
The fully merged zone is the one downstream of the intermediate zone. Within this
zone complete merging of the flows from the initial merging zone has occurred. The
fully merged flow behaves like a combined jet and its characteristics are similar to
those of a single jet.
Though the above delineation of the three zones is approximate, it will be supported
by good similarity of the mean velocity and turbulence intensity profiles observed in
thesc zones.
22-2
644 N . W . M.KO and W . T.Chan
I .5

FIGURES 2 (a,b). For legend see next page.

 Similarity in unnula, jets 645

0 3 4 5 6 7
-4D,
FIGURE
2. Mean velocity contours. (a)Annular jet. ( b ) Conical jet. ( e ) Ellipsoidal jet.

Mean velocity contours

- -
The contours of the local mean velocity ratio U/Uo of the annular, conical and ellip-
soidal jets are shown in figures 2 ( a ) ,( b ) and (c) respectively. As shown in figure 2 (a),
the entrainment in the central region of the annular jet is responsible for the genera-
tion of the internal recirculating vortex of toroidal form (Chigier & Beer 1964). This
recirculating vortex is mainly found within the first half-diamet,er downstream of the
nozzle exit. Because of this vortex, the associated subatmospheric pressure and the
radial pressure forces draw the potential core towards the central axis. Besides the
_ - figure 2 (a) also shows that the contours corresponding to mean
above displacement,
velocity
- - ratios U/Uo >, 0.6 are also displaced towards the central axis. The contour
U/Uo = 0.5, however, is roughly parallel to the central axis and is not really affected
by
- -the subatmospheric region associated with the vortex. For the contours with
U/Uo < 0.4 outward displacement is found and is mainly influenced by the entrain-
ment region outside. _ _
Like the results for the annular jet, the contours for U/Uo < 0.5 for the conical and
ellipsoidal jets, as shown in figures 2 ( b ) and (c), are mainly influenced by the - outer
-
entrainment region. Similarly, the effect of the inner entrainment is felt for U/Uo 2
0.6, in spite of the presence of the bullets.
Comparison of the mean_ - velocity contours for the conical and ellipsoidal jets shows
that the contours for U/Uo < 0.5 are basically identical and have the same angle of
spread. Even though the vortex in the centre is eliminated by the bullet, a certain
amount of entrainment is still present, resulting in the displacement of the annular
potential core towards - -the central axis. However, no appreciable difference in the
contours for 0.80 2 U / U o 2 0.6 is found. The noticeable difference between the two
646 N . W . M . KO and W .T.Chan

0
0.5 1 .O 13 2.0 2.5
Diameter ratio, D J D ,
50 ms-l ; x , 30 ms-l ;
3. Variation of reattachment position with diameter ratio. 0,
FIGURE
50 ms-l ; v, Chigier & Beer ; 4 , Miller & Comings; 0,
a, Ayukawa & Shakouchi.
_ _
jets is in the contours for U I U , 2 0.90. This is due to the shapes of the two bullets
used: the conical bullet results in slightly higher degree of entrainment than the
ellipsoidal one. This may be the reason for the bigger displacement of the contours
towards the central axis for the conical jet than for the ellipsoidal jet.
Comparison_between
- the annular jet and the other two jets shows that, although
the contours U / U , = 0.50 of all the jets are at the radial position y/Do = 0.5, the
spread of the outer mixing region of the annular jet (figure 2a) is slightly
_ _ bigger than
that of the other two. As would be expected, the contours with U/Uo > 0.60 of the
annular jet are displaced more towards the central axis. It is due to the presence of
the subatmospheric pressure region in the centre. --
Also from figure 2, the position at which the annular potential core ( U/Uo = 0.995)
of the three jets ends is found at ./Do 2: 1.7, y/Do 2: 0.3. It is independent of the
conditions in the centre of the jet. Even though the termination position is the same,
the shape of the potential core is affected by the conditions in the centre. This leads
one to expect the shape of the potential core of the annular jet to be noticeably
smaller than that of the other two.
As for the potential core, figure 2 further indicates that the end of the axis of the
annular jet (figure l ) , or the point of reattachment, is also independent of the
conditions inside the jet. For the three jets shown in figure 2 the axial position of
reattachement is ./Do 2: 4. This means that, irrespective of the presence of either
the internal vortex or the bullet, the high velocity flow which is inherited from the
annular potenbial core merges at this axial position ./Do 2: 4.
 Similarity in annulnr j e t s 647

- 1.0 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1.0
Non-dimensional radial distance, q t =(J-)'~.J/(~'~.~
-yo.,)

FIGURE 4. Similarity of mean velocity ratio in initial merging zone. Annular, x/D,: 0, 0.6;
A,0.75; V , 1.0; 0,2.0. Conical, x / D , : 0 , 0.5; A, 0.75; v , 1.0; 4 , 1-5; g,2.0. Ellipsoidal,
x/D,: d ,0.5; $, 1.0; 4 ,1.5; d , 1-75; d ,2.0.

This axial position x/Do N 4 for reattachment has also been observed at the lower
exit velocity of 30 ms-1. In other words, the exit velocity does not affect this position.
The position of reattachement of an annular jet has also been given by Chigier &
Beer (1964). At a jet exit velocity of 36 ms-1 reattachment was found at an axial
position of 2.06 Do. The corresponding position for two-dimensional dual jets with
an exit velocity of 22 ms-1 was found at 1.47 Do by Miller & Comings (1960). For the
two-dimensional jet attaching to a wall of Ayukawa & Shakouchi (1976) the axial
position of reattachment was 0.94 Do, for an exit velocity of 30 ms-l. In the results
of Miller & Comings (1960) and Ayukawa & Shakouchi (1976) slight extrapolation
was adopted to obtain the merging distance a t the central axis.
Together with the annular-jet results obtained with the nozzle configuration of
KO & Kwan (1976), the reattachment distance is correlated and plotted in figure 3.
Extremely good correlation is found with the nozzle diameter ratio Do/Di.This good
correlation is interesting because the results of Miller & Comings (1960) and Ayukawa
6 Shakouchi (1976) are for two-dimensional jets. Further, the latter work was for
a jet attaching to a wall.
The trend of the reattachment distance with the diameter ratio Do/Di shown in
figure 3 suggests that for bigger diameter ratios reattachment occurs further down-
stream. This phenomenon may be due to the larger momentum associated with a jet
of bigger diameter ratio. This reasoning is supported by the work of Miller & Comings
(1960) and later of Chigier & Beer (1964), which stated that the total momentum flux
was conserved. I n the case of the initial region of the annular jet the momentum due
to the fluctuating axial velocity was small compared with that due to the mean axial
velocity and that due to the mean static pressure. Chigier & Beer (1964) have further
648 N . W . M . KO and W . T . Chan

- 0.2 -0.1 0 0.1 0.2

Non-dimensional radial distance, 7: =Cv- +D,,)/.r

FIGURE
5. Similarity of mean velocity ratio in initial merging zone. Symbols same as in figure 4.

observed that the pressure term was not as dominant as the mean velocity term. It
was estimated that the pressure term reached a maximum value of I 5 &: of the total
input momentum. According to this, the higher input momentum associated with
the higher Do/Di would result in a mean velocity momentum flux higher than that
due t o the pressure. On balance, this higher mean velocity momentum would be
responsible for the shift of the reattachment further downstream.

similarity
Similarity of the local mean velocity profiles in the different zones of the three jets is
discussed in this subsection. For the initial merging zone the similarity curves for the
outer mixing region of the annular, conical and ellipsoidal-jets
- are shown in figure 4.
I n this figure the non-dimensional mean velocity ratio U/Uo is plotted against the
non-dimensional radial distance ?ti = (y - yo.5)/(yo.9 - where yo.9,yo.5 and yol
are the radial positions where the local mean velocity is equal to 0.9, 0.5 and 0.1 of
the jet exit velocity no.This non-dimensional radial distance was suggested by
Abramovich (1963, p. 9) for coaxial-jet results. For their coaxial-jet results KO &
Kwan (1976) also found that ?ti was a better non-dimensional radial distance for
the primary mixing region than (y - go,)/..
Using the non-dimensional distance tii, very good similarity of the .local mean
velocity profiles for axial distances 0-5 < ./Do < 2 is found for all three jets
(figure 4). For ./Do = 3 only slight deviation from the similarity curve is found.
Further than three outer diameters downstream progressively greater deviation is
observed. The single-jet result is also shown in figure 4.Very good agreement of the
 Similarity zn annular jets 649

-0 2 -0 1 0 01 07
Non-dimensional radial distance. 7 , = 0 - t. D,)

FIUURE
6. Similarity of mean velocity ratio in intermediate and fully merged zone. Annular,
x/D,: 0 , 3; 0, 4;0 ,5 ; 0 , 6 ; 0 , 7. Conical, z/D,: +, 3 ; 8 , 4; 0 ,5; 1 , 6 ; 4,7. Ellipsoidal,
xID,: 6 , 3 ;(f, 4; d , 5 ; d ,6; 6, 7.

similarity curve for the single jet with those for the three jets of this investigation is
found.
Besides the correlation with the non-dimensional radial distance qi, similarity of
the mean velocity profiles in the outer mixing region with 7; = (y - ;tDo)/x has also
been investigated. This non-dimensional radial distance was suggested for the single
jet by Davies, Fieher & Barratt (1963) and other workers. Within the same range of
axial distances 0.5 < x/D, < 2, very good similarity is found for the velocity pro-
files of each jet measured here (figure 5 ) . However, when a comparison is made with
the single jet, differences are found.
The greatest deviation of the similarity curve from that for a single jet is found for
the annular jet. The curve is displaced further towards the central axis. For the
conical jet, however, less displacement towards the central axis is found while the
displacement for the ellipsoidal jet is the least of the three. This difference in the
deviations of the similarity curve within the initial merging zone illustrates clearly
the effect of the entrainment in the centre on the mean velocity. As would be ex-
pected, the presence of the subatmospheric static pressure region and the presence of
the two different bullets in the centre cause the difference in the deviation.
- I-n the intermediate and fully merged zone similarity of the mean velocity ratio
UIU, with the non-dimensional distance qe = ( y - t o , ) /is~found (figure 6), ombeing
the maximum mean velocity in the x,y plane and 0, the equivalent diameter of the
fully merged jet. The equivalent jet is derived from the idea of Eldred et aE. (1971)
that the fully merged jet can be represented by an equivalent jet of identical thrust
having a diameter 0, and an equivalent exit velocity o,.
In the present case 0, is
equal to i7$,,,.
650 N . W . M . KO a d W . T.Chan

0 0.5 1.o 1.5 2.0

Non-dimensional radial distance,
FIGURE
7. Similarity of mean velocity ratio in fully merged zone. Symbols same
as in figure 6.

As shown by figure 6, good similarity of the three jets over axial distances
3 < x/D, < 7 has been found. In addition, good agreement with the single-jet
results has been found. However, more detailed comparison in figure 6 shows better
agreement for ve < 0 and slight deviation from the single-jet results for ve 2 0. For
7, >, 0 the results for the annular jet agree slightly better than those for the conical
and ellipsoidal jets.
- -based on qe has been shown to be good, in the fully merged zone
Though similarity
the similarity of UIU, with the non-dimensional radial distance y/y, has been found
to be better in all three jets (figure 7). Here yc is the radial distance at which the locd
mean velocity is equal to 0.5 of gm. This better similarity is observed only for
x/D, 2 5. Furthermore, very good agreement with the single-jet results in the fully
developed region has also been found.
This good similarity in the fully merged zone suggests that all three jets are self-
preserving from an axial distance of five outer diameters downstream of the nozzle
exit. The good agreement with the single-jet results in the same region suggests that
the combined jets behave exactly like a single jet.
The results discussed above were for an exit velocity of 50 ms-l. Results for an
exit velocity of 30 ms-l, not shown in this paper, have also been obtained. Detailed
comparison between the two sets of results indicates the same similarity of the mean
velocity ratios in the initial merging zone, the intermediate zone and the fully merged
zone. Because the similarity is the same, good agreement with the single-jet results
is found. This independence of the similarity of the exit velocity is the same pheno-
menon as is observed in single jets (Davies et al. 1963; and others).
 S i n i i l a d y in annular j e t s 651

I I I

8. Distribution of turbulence intensity level of annular jet. % / D o :x , 0 ; 4,

FIGURE 1.5; + , 3.5.
Other symbols same a s in figures 4 and 6.

4. Turbulence intensity
Distribution
The axial distribution of the turbulence intensity profiles of the annular jet is shown
in figure 8. Because the profiles of the conical and ellipsoidal jets are basically the
same, except in the inner mixing region, they are not shown in the present paper.
Fundamentally, within the outer mixing region one peak in the turbulence in tensity
profiles is found for all three jets. Over the whole axial distance considered, 0 <./Do
< 7, the peak in the outer mixing region is located at the radial position y/Do N 0-5.
This radial position of the peak is roughly the same as that for a single jet (Davies
et al. 1963; Mollo-Christensen, Kolpin & Martuccelli 1964; and others). However, as
will be discussed later, there is a slight displacement of the radial position of the peak
turbulence intensity towards the central axis.
The position of the peak turbulence intensity in the inner mixing region of the
annular jet is different from those for the conical and ellipsoidal jets. For the annular
jet the peak becomes displaced away from the central axis as the axial distance is
increased. At the axial position ./Do 2: 3, i.e. in the intermediate zone, the peak
turbulence intensity is masked by the intensity level of the outer mixing region.
652 N . W . M . KO and W.T.Chan

0.02 -

I I I I I I I

Axial distance. .v/D,,

9. Axial distribution of maximum turbylence intensity level. 0,

FIQURE annular jet; 0 ,conical
jet; d , ellipsoidal jet; x , single jet.

For the conical and ellipsoidal jets the peak intensity tends to shift towards the
central axis at larger distances downstream. The axial position where the pet&
turbulence of the inner mixing region is masked by the local turbulence of the outer
mixing region is at x/Do 'v 3.5 for the conical jet and ./Do 'v 4 for the ellipsoidal jet.
The axial distribution of the peak turbulence intensity of the outer mixing region
is shown in figure 9. Within the axial distance considered, 1 < ./Do < 7, the peak
turbulence intensity of the annular jet is slightly higher than those of the conical
and ellipsoidal jets, while the results for the latter two jets are the same. The biggeet
difference is observed in the intermediate zone 2 < x/D, < 5 . This suggests that the
effect of the subatmospheric pressure region on the peak turbulence intensity iS
mostly felt in this zone. I n the fully merged zone the three jets have the same peak
intensity.
The single-jet results of Laurence (1956) and KO& Davies (1971) are also shown in
figure 9. Within the initial merging zone comparison shows that the peak intensity
of the single jet is slightly higher than those of the present three jets. In the inter-
mediate zone, however, the single-jet results agree with those for the annular jet.
I n t'he fully merged zone the results for all the jets are the same.
 Similarity in annuiar jets 653

Initial merging
zone, outer Intermediate Fully merged Single
mixing region zone zone jet
Non-dimensional XIDO XlDo w o 4Do
axial distance
Non-dimensional (Y- ~ ~ . d l ( y-Y,~,.,.)~ (Y- We)h (Y- Pe)b (Y- W l x
radial distance
Mean velocity ratio
-
u/uo upm a/um
- -
u/uo
Turbulence (U*)* P o ((3
vo)mx P)w o (ua) w o (US) w o
intensity
TABLE1. Summary of non-dimensionalparameters of annular jets.

FIGURE10. Similarity of turbulence intensity in initial merging zone.

Symbols same as in figures 4 and 8.

Similarity
Similarity of the turbulence intensity profiles in different zones of the three jets is
described in this subsection. The non-dimensional parameters which have been found
for the similarity of the mean velocity profiles will be used to correlate with the local
turbulence (table 1). The similarity curve for the initial merging zone of the three jets
is shown in figure 10.The turbulence intensity ratio, i.e. the ratio of the local intensity
to the maximum intensity in that particular plane, is correlated with the non-
dimensional radial distance T~ = ( y - y,,J/(yWe- yo.1).From this figure good similarity
of the results within the initial merging region y / D o < 2 is found for all three jete.
Deviation starts to occur a t ./Do 2 3.
Careful comparison of the results for the conical and ellipsoidal jets shows that
their similarity curves are the same. When this curve is compared with that for the
annular jet, the similarity curve of the latter tends to be displaced slightly towards
654 N . W . M . KO and W . T.Chan

-0.5 -0.4 -0.3 -0.2 -0.1 0

Non-dimensional radial distance. p,'=Ct'- iD,,)/.r
FIGURE
11. Similarity of turbulence intensity level in initial merging zone.
Symbols same as in figures 4 and 8.

0.14

1 , I I

-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5
Non-dimensional radial distance, p,=Ct*- $D,)/.Y

FIGURE12. Similarity of turbulence intensity level in intermediate and fully merged zone.
Annular: 0 , %/Do= 2.5. Conical: 0 , ./Do = 2.5. Ellipsoidal: 4,./Do = 2.5. Other symbols
sa.me as in figure 6.
 Sinzilarity in annular jets 655
the central axis. This shift, though very slight, indicates the effect of the entrainment
and the presence of the subatmospheric pressure region in the centre.
As has been shown by the mean velocity, the similarity of the turbulence intensity
@)+/Do of the three jets with the non-dimensional radial distance 7; = (y - &Do)/x
is different from the single-jet results (figure 11). The peak of the intensity profiles
of the three jets is found at 7; N 0 and is slightly more displaced towards the central
axis than for the single jet. Among the three jets of this investigation the annular
jet has the biggest displacement and the ellipsoidal jet the smallest of the three. This
agrees with the results for the mean velocity shown in figure 5. However, the dis-
placement of the t,urbulence intensity similarity curve is not as large as that of the
mean velocity.
I n the intermediate and fully merged zones 2 - 5 < x/Do < 7, similarity of the
turbulence intensity profiles has also been found (figure 12). Also as in the case of the
mean velocity, the turbulence intensity @)*/go correlates well with the equivalent
non-dimensional radial distance ve. For the annular jet, however, the similarity for
ye < 0 is not as good as that for ye > 0. The reason may be the effect of the sub-
atmospheric pressure region on the mixing. This poorer similarity for 7e < 0 is also
found in the results for the conical and ellipsoidal jets, though to a lesser degree.
The similarity curves of the conical and ellipsoidal jets are basically the same. They
are slightly different from and are slightly more displaced towards the central axis
than that for the annular jet. The similarity curve for the annular jet is very nearly the
the same as that for a single jet, while those for the other two jets are slightly
different.
I n the fully merged zone similarity of the turbulence intensity profiles with the
non-dimensional distance y/y, has been found not to be as good as that of the mean
velocity. Thus the results are not included in this paper.
Results for an exit velocity of 30 ms-1 are also found to agree very well with the
results presented above. Thus they are excluded from the present investigation.

5. Discussion
The mean velocity and turbulence intensity measurements made in the outer
mixing region within the initial region of the three annular jets permit the isolation
of three separate zones: the initial merging, the intermediate and the fully merged
zone. Irrespective of the type of jet, i.e. annular, conical or ellipsoidal, the initial
merging zone is found within the first two outside diameters downstream of the
nozzle exit. It is within this zone that the annular potential core exists. The inter-
mediate zone, where the mixing of the high velocity flow inherited from the potential
core occurs, is found within the next three outside diameters downstream. After the
intermediate zone there is the fully merged zone, where the flow has become fully
merged and behaves like a combined jet.
I n all three of theee zones similarity of both the mean velocity and the turbulence
intensity profiles in the outer mixing region has been found. In each zone, the same
non-dimensional radial distance is found for both the mean velocity and the turbu-
lence intensity. The similarity observed in the different zones further supports their
division.
666 N . W . M . KO and W . T.Chan
The difference in the entrainment in the centre of the jet in the three configurations
is responsible for the difference in the displacement of both the mean velocity and the
turbulence intensity profiles towards the central axis. This difference is more obvious
in the initial merging zone, where the effect of entrainment is more obvious.
I n the intermediate zone a much smaller displacement is found. I n the fully merged
zone no difference is observed.
Very good agreement between the similarity curves for the present three jets and
for a single jet is found. The initial merging zone corresponds to the initial region of
the single jet where the potential core and mixing region are found. The intermediate
zone corresponds to the transition region and the fully merged zone to the fully
developed region.
REFERENCES
ABRAMOVICH, N. 1963 The Theory of Turbulent Jeta. M.I.T. Press.
AYUKAWA, K. & SEAKOUCHI, T . 1976 Analysis of a jet attaching to an offset parallel plate.
Bull. Japan SOC.Mech. E y r s 19, 395-401.
CHIGIER,N. A. & BEER,J. M. 1964 The flow region near the nozzle in double concentric jets.
Tram. A.S.M.E., J . Bask Engng 86, 794-804.
DAVIES,P. 0. A. L. & DAVIS,M. R. 1966 Hot-wire anemometer. Inst. Sound Vib. Rm., Uniu.
Southampton Rep. no. 155.
DAVIES,P. 0. A. L., FISHER,M. J. & BARRATT, M. J . 1963 The characteristics of the turbulence
in the mixing region of a round jet. J . Fluid Mech. 15, 337-367.
ELDRED, K. M. et al. 1971 Far-field noise generation by coaxial flow jet exhaust. 1 . Detailed
discussion. Wyle Lab. Rep. no. FAA-RD-71- 101, p. 1.
KO,N. W. M. & DAVIES,P. 0. A. L. 1971 The near field within the potential cone of subsonic
cold jets. J . Fluid Mech. 50, 49-78.
KO, N. W. M. & KWAN,A. S. H. 1976 The initial region of subsonic coaxial jets. J . Fluid
Mech. 73, 305-332.
LAURENCE, J. C. 1956 Intensity, scale and spectra of turbulence in the mixing region of a free
subsonic jet. N.A.C.A. Rep. no. 1292.
MILLER,D. & COMINGS,E. W. 1960 Force-momentum fields in a dual-jet flow. J . FZuid Mech.
7,237-256.
MOLLO-CHRISTENSEN,E., KOLPIN,M. A. & MARTUCCELLI, J. R. 1964 Experiments on jet flows
and jet noise for field spectra and directivity patterns. J . Fluid Mech. 18, 285-301.