Control of an axisymmetric jet using a passive ring

Control of an axisymmetric jet using a passive ring

R. Parker *, S. Rajagopalan, R.A. Antonia

Department of Mechanical Engineering, University of Newcastle, New South Wales 2308, Australia

Received 31 March 2002; accepted 30 June 2002

Abstract

This study examines the development of an axisymmetric jet, which is disturbed by placing a thin wire ring axisymmetrically in

the mixing layer close to the nozzle exit (0:16 x=D6 0:5). Mixing layer development up to x=D ¼ 5 is examined in detail using X-hotwire measurements, whereas far field development up to x=D ¼ 30 is examined using single hotwire measurements on the jet axis.The introduction of the ring in the mixing layer results in a significant reduction in both the mixing layer growth rate and jet spread

rate. The turbulence level in the mixing layer is substantially reduced for both axial and radial fluctuations (�20% and 15%respectively). In addition to these reductions, the normalised distributions of mean velocity, axial and radial velocity fluctuations

as well as the skewness and flatness factor of these fluctuations in the mixing layer all exhibit improved similarity.

Variations of the jet Reynolds number (and thus the wire Reynolds number of the ring) and the axial location of the ring yield

nearly identical changes to the characteristics of jet mixing layer. Increasing jet Reynolds number reduces the effect of the ring in the

region 5 < x=D < 30.� 2003 Elsevier Science Inc. All rights reserved.

Keywords: Axisymmetric jet; Passive control; Mixing layer

1. Introduction

The control of turbulence in jet flows has been the

subject of a large number of studies, with many different

methods, both active and passive, employed to modify

the turbulence characteristics of jet flows. In active

techniques energy is added at an appropriate frequencyand amplitude, using some form of actuator such as

speakers or mechanical actuators. Acoustic excitation

has been used extensively to both increase, Crow and

Champagne [1], and suppress, Zaman and Hussain [2],

turbulence levels.

Passive techniques depend on modifying critical

conditions, e.g. nozzle exit conditions or the introduc-

tion of objects into the flow that alter its development.Passive techniques have the advantage of requiring no

external energy source or control system. Reductions in

turbulence have been achieved through the introduction

of small diameter cylinders in the mixing layers of a

plane jet, Rajagopalan and Antonia [3]. Schrober and

Fernholz [4] used a trip wire to suppress the formation

of large structures in a wall jet.

Tong and Warhaft [5] examined the development of

an axisymmetric jet, up to x=D ¼ 25, modified by in-serting a fine wire ring in the shear layer of the jet. The

insertion of the ring resulted in a significant reduction

in RMS axial velocity fluctuations, a reduced spatialgrowth rate, a shift in the virtual origin of the jet and a

suppression of vortex formation and pairing.

Brown and Ahuja [6] investigated the use of a thick

wire ring in a high velocity coaxial jet, obtaining sig-

nificant increases in mixing when the vortex shedding

frequency from the ring was within the preferred range

for the jet, i.e. Strouhal number based on jet diameter of

0.2–0.5. The rate of decay of the jet axis Mach numberwas significantly increased, consistent with enhanced

mixing, when the vortex shedding frequency was within

this range. An 18% reduction in axis Mach number was

achieved at x=D ¼ 11 when the vortex shedding fre-quency corresponded to a jet Strouhal number of 0.47,

which was accompanied by a reduction in the potential

core length from 6 to 4.5 diameters.

The present work extends the investigation of Tongand Warhaft [5], by examining the effect of introducing

Experimental Thermal and Fluid Science 27 (2003) 545–552

www.elsevier.com/locate/etfs

*Corresponding author.

E-mail address: [email protected] (R.

Parker).

0894-1777/03/$ - see front matter � 2003 Elsevier Science Inc. All rights reserved.doi:10.1016/S0894-1777(02)00268-6

mail to: [email protected]

a fine (d ¼ 0:0244D) wire ring on axial (u0) and radialvelocity (v0) fluctuations, higher order moments of ve-locity fluctuations (skewness and flatness factor) and the

Reynolds shear stress. Detailed measurements in the

initial region of the jet (x=D6 5) are used to examine

the effect of the ring in the near field with particular

attention to shear layer development. Measurements on

the jet axis up to x=D ¼ 30 are used to examine effects onthe far field development of the jet.

2. Experimental set-up and conditions

2.1. Experimental set-up

The set-up consists of a 20.5 mm axisymmetric nozzleto which air is supplied by two coupled centrifugal fans

operating in parallel and driven by a D.C. motor. The

speed of the fans, and thus the jet exit velocity, can be

varied by changing the motor supply voltage. Air enters

a settling chamber and passes through a series of screens

and a honeycomb before entering the nozzle contrac-

tion, which has a diameter ratio of 6:1. The ring used

to disturb the mixing layer is constructed using a 0.5mm diameter wire and has an outer diameter of 20

mm. Measurements of mixing layer characteristics for

0 < x=D < 5 were performed using a 2.5 lm X-hotwireprobe, with an etched length of approximately 0.5 mm

and a wire separation of 0.5 mm, operated in constant

temperature mode. Measurements on the jet axis for

0 < x=D < 30 were performed using a 2.5 lm single hotwire probe, with an etched length of approximately 0.5mm, operated in constant temperature mode. Most de-

tailed measurements were carried out at a jet Reynolds

number (ReD ¼ UjD=m) of 23 500 with the ring posi-tioned concentrically with the nozzle exit at x=D ¼ 0:15.A few additional measurements at different Reynolds

numbers and ring positions were also carried out to

check the influence of ReD and ring location on flowdevelopment.

2.2. Nozzle exit conditions

Boundary layer conditions at the nozzle exit plane are

known to a have significant influence on jet develop-

ment. Mi et al. [7] have shown that initial conditions

have a significant influence on the structure of the self-

similar region of the jet as well as the development of the

initial region. It is therefore important to determine the

boundary layer conditions if meaningful comparisonsbetween measurements made in different test facili-

ties are to be made. Fig. 1 shows normalised boundary

layer profiles slightly upstream of the nozzle exit

(x=D ¼ �0:025) obtained with a miniature single hot

Nomenclature

D nozzle exit diameter (m)

d wire diameter (m)Fu flatness factor of axial velocity fluctuations

(dimensionless)

Fv flatness factor of radial velocity fluctuations

(dimensionless)

f frequency (Hz)

G shear gradient (s�1)

H boundary layer shape factor (dimensionless)

ReD jet Reynolds number (UjD=m) (dimensionless)Rew wire Reynolds number (Ud=m) (dimension-

less)

Rh momentum thickness Reynolds number (di-

mensionless)

Su skewness of axial velocity fluctuations (di-

mensionless)

Sv skewness of radial velocity fluctuations (di-

mensionless)StD jet Strouhal number (fD=U ) (dimensionless)StW wire Strouhal number (fd=U ) (dimensionless)U mean axial velocity (m/s)

UC mean centreline axial velocity (m/s)

Uj jet exit velocity (m/s)

u0 RMS axial velocity fluctuations (m/s)

v0 RMS radial velocity fluctuations (m/s)

x downstream displacement from jet exit plane(m)

y radial displacement from jet centre line (m)

m kinematic viscosity of air (m2/s)

d� boundary layer displacement thickness (mm)

h boundary layer momentum thickness (mm)

Fig. 1. Normalised boundary layer profiles at x=D ¼ �0:025.

546 R. Parker et al. / Experimental Thermal and Fluid Science 27 (2003) 545–552

wire probe. Calculation of the shape factor for these

profiles shows the boundary layer to be laminar at the

three values of Uj. These profiles show a small departurefrom the Blasius solution. This deviation could be due to

the presence of a small, favourable pressure gradient

along the nozzle contraction. The potential core turbu-

lence level, for all jet Reynolds numbers examined in this

study, is less than 1% (see Table 1). The velocity profileat the exit plane (excluding the boundary layer) is uni-

form within �0:005 Uj.

3. Results

3.1. Vortex shedding from ring

Examination of axial velocity fluctuations, in the

wake of the ring, and frequency spectra, show clear in-

dications of vortex shedding from the ring. However the

structures generated by this vortex shedding appear to

break down within a short distance. With the ring po-

sitioned at x=D ¼ 0:15 both the hot wire trace (Fig. 2)and the spectra (Fig. 3) show clear evidence of vortex

shedding at x=D ¼ 0:25. At x=D ¼ 0:5 the peak at thevortex shedding frequency is greatly reduced and by

x=D ¼ 0:75 there is no evidence of the vortex shedding.With the ring positioned at x=D ¼ 0:15 for ReD ¼23500, the observed vortex shedding frequency was 4550

Hz. For the same ring position and ReD ¼ 13300, thevortex shedding frequency was 1960 Hz. These fre-

quencies correspond to jet Strouhal numbers

(StD ¼ fD=Uj) of 5 and 3.9 respectively, well above thepreferred mode Strouhal number of the jet, �0.3–0.5,

and hence is unlikely to excite the jet in the manner

achieved by Brown and Ahuja [6].

An accurate estimate of the ring Reynolds number

and Strouhal number for the vortex shedding frequency

in this situation is difficult due to the high level of shear

in the region around the ring, which makes an accurateestimate of mean velocity difficult. A detailed velocity

profile of the shear layer was obtained at x=D ¼ 0:15,without the ring in place, and the velocity measured at

the location equivalent to the wire centreline is used here

to estimate wire Reynolds and Strouhal numbers. The

velocities obtained for ReD ¼ 23500 and ReD ¼ 13300were 10.6 and 4.6 m/s respectively which correspond to

ReW ¼ 341 and 148 and StW ¼ 0:214 and 0.213 respec-tively. The wire Strouhal numbers obtained for are

slightly higher than the calculated values. For ReW ¼148 the empirical relationship developed by Roshko [8]

St ¼ 0:212� 0:45=ReW for 50 < ReW < 150 ð1Þresulted in a Strouhal number of 0.209. For ReW ¼ 341,which is above the valid range for this relationship,

St ¼ 0:212 would normally be assumed. However thisrelationship and the assumed value for higher Reynolds

number were developed for uniform flows. Kiya et al. [9]examined the effect of a radial linear velocity gradient on

the vortex shedding frequency and thus the Strouhal

number. A non-dimensional shear parameter, K, wasproposed for quantifying the magnitude of the shear,

viz.

K ¼ Gd=UCL ð2Þwhere G is the gradient of the shear (dU/dy), d is thediameter of the cylinder and UCL is the mean velocity atthe cylinder centreline. Kiya et al. [9] found that for

100 < Re < 1000, increasing K in the range 0:1 <K < 0:25 resulted in an increase in the vortex sheddingfrequency. The values of K for the wire ring at these jetReynolds numbers are 0.91 and 0.89 respectively making

direct comparison with the results of Kiya et al. [9]

Time (Sec)

Vol

tage

(arb

itrar

ysc

ale)

Fig. 2. Hot wire trace in wake of ring, x=D ¼ 0:25.

10-11

10-9

10-7

10-5

10-3

0 2000 4000 6000 8000 10000

x/D = 0.25x/D = 0.5x/D = 0.75

Frequency (Hz)

Arb

itrar

ysc

ale

Fig. 3. Velocity spectra in wake of ring.

Table 1

Nozzle exit boundary layer statistics

Uj (m/s) d� (mm) h (mm) H Rh Potential core

turbulence (%)

4.1 0.813 0.275 2.96 751 0.93

10.3 0.444 0.171 2.60 1174 0.83

15.7 0.360 0.144 2.49 1507 0.82

R. Parker et al. / Experimental Thermal and Fluid Science 27 (2003) 545–552 547

impossible. It is clear however that the presence of a

shear gradient does affect the vortex-shedding frequency

and could be a contributing factor to the difference seen

here between the calculated and measured vortex shed-

ding Strouhal numbers.

3.2. Jet axis development up to x=D ¼ 30

3.2.1. Mean velocity ðUÞThe mean velocity distributions shown in Fig. 4 dis-

play qualitative agreement with those of Tong and

Warhaft [5] (shown as solid lines) with both sets of data

showing an extension of the potential core and a re-duced spread rate for the jet in the far field. Estimation

of the jet virtual origin by extrapolating (as in Tong and

Warhaft [5]) the self-similar region of the Uj=Uc plot toUj=Uc ¼ 0, shows an downstream shift from �1 to 2whereas Tong and Warhaft [5] obtained a shift from 0 to

3. The difference in these two results might be due to

some minor difference in the jet initial conditions but

nevertheless the shift in the virtual origin due to theintroduction of the ring appears to be similar. Figs. 5

and 6 show the same distributions obtained at ReD ¼18660 and ReD ¼ 23500 respectively and indicate thatthe effect of the ring on jet growth rate in the far field

reduces with increasing ReD.

3.2.2. Distribution of RMS axial velocity fluctuations (u0)The results of Tong and Warhaft [5] (shown in Fig. 4

as solid lines) and the present investigation (Fig. 4)

shows significant reductions namely 30% for Tong and

Warhaft and 22% for the present study, in the maximum

observed value of u0 as well as a significant upstreamshift in the location of this maximum. The peak values

observed in the present investigation are lower than

those in [5] particularly in the case of the unmodified

flow; this difference is most likely due to different initial

conditions in the two jets.

The axial distributions of u0 (Fig. 4) for the unmod-ified flows in both studies tend to exhibit an unexpecteddouble peak trend, one further upstream than would

normally be expected (x=D ¼ 4 for [5] and x=D ¼ 3 forthe present case) followed downstream by a second

higher peak in the present study and a short plateau in

[5]. The introduction of the ring modifies the u0 distri-bution by suppressing the first peak and by shifting the

second peak downstream to x=D ¼ 9. Increasing ReD,for the unmodified flow, as shown in Figs. 5 and 6 re-sults in a u0 distribution with a single distinct peak atapproximately x=D ¼ 7. Increasing ReD results in pro-gressively smaller differences in the maximum values of

u0 with and without the ring. The location of the peak inthe u0 distribution with the ring remains constant atapproximately x=D ¼ 10 for all values of ReD investi-gated, which agrees with the results of [5].

Fig. 4. Normalised mean velocity (U=Uj) and absolute turbulenceintensity (u0=Uj) on jet axis at ReD ¼ 13300.

0

2

4

6

0 10 20 300

0.05

0.10

0.15

0.20

u'/Uj - With Ringu'/Uj - No RingUj/Uc - With RingUj /Uc - No Ring

x/D

Uj/U

c

u'/U

j


0

2

4

6

0 10 20 300

0.05

0.10

0.15

0.20u'/Uj - With Ringu'/Uj - No RingU

j/Uc - With Ring

Uj/Uc - No Ring

x/D

Uj/U

c

u'/U

j



3.3. Mixing layer development

3.3.1. Mixing layer mean velocity distribution ðUÞThe growth of the mixing layer, estimated from mean

velocity measurements (Fig. 7), shows that the disturbed

flow exhibits an increase in thickness immediately be-

hind the ring but a reduced rate of mixing layer growth

further downstream resulting in a reduction in mixinglayer thickness for x=D > 1. This decreased rate ofmixing layer growth is mainly the result of a reduction in

the spread rate for the jet. However, a small increase

in the width of the potential core is also observable,

particularly downstream of x=D ¼ 4 which yields anincrease in the length of the potential core.

The modified flow exhibited a shift of the mixing

layer virtual origin from x=D ¼ 0 to x=D ¼ �0:6. Al-though there is a rapid increase in the mixing layer

thickness immediately downstream of the ring the re-

duced mixing layer growth rate yields an upstream shift

in the virtual origin. Fig. 8 shows the normalised mixing

layer mean velocity distribution where the modified flow

shows increased similarity consistent with a linear mix-

ing layer growth rate. The most significant improvement

in similarity occurs on the low speed side of the mixing

layer particularly in the region x=D6 1:25 where simi-larity of the undisturbed flow was relatively poor.

Antonia et al. [10] have examined the occurrence of

reverse flow in an axisymmetric jet and the subsequent

errors introduced in statistics obtained, using hot wireanemometry, due to rectification. The results presented

in [10] demonstrate that the errors introduced due to

flow reversal on statistics for U to be minimal, withinthe range presented in Fig. 7, and as such no corrections

have been applied to this data.

3.3.2. RMS velocity fluctuations

The modified flow exhibits a reduction in the maxi-mum RMS values for both the axial and radial velocity

fluctuations, of �20% in u0 (Fig. 9) and 15% in v0 (Fig.10). The disturbed flow showed greatly enhanced simi-

larity for the normalised (by the mixing layer virtual

origin) distributions of both u0 and v0 particularly on the

-1.0

-0.5

0.0

0.5

1.0

U/U

j-U

nmod

ified

Jet

0.0

0.5

1.0

1.5

2.0

U/U

j-M

odifi

edJe

t

-0.5 0 0.5 1 1.5y/(x-x0)

x/D = 5.0x/D = 4.5x/D = 4.0x/D = 3.75x/D = 3.5x/D = 3.25x/D = 3.0x/D = 2.75x/D = 2.5x/D = 2.25x/D = 2.0x/D = 1.75x/D = 1.5x/D = 1.25x/D = 1.0x/D = 0.75x/D = 0.5

Unmodified Jet

Modified Jet

Fig. 8. Normalised mean velocity distribution ðUÞ.

-0.10

-0.05

0.00

0.05

0.10

0.15

u'/U

j-U

nmod

ified

Jet

0.00

0.05

0.10

0.15

0.20

0.25

u'/U

j-M

odifi

edJe

t

-0.5 0.0 0.5 1.0 1.5y/(x-x0)

x/D = 5.0x/D = 4.5x/D = 4.0x/D = 3.75x/D = 3.5x/D = 3.25x/D = 3.0x/D = 2.75x/D = 2.5x/D = 2.25x/D = 2.0x/D = 1.75x/D = 1.5x/D = 1.25x/D = 1.0x/D = 0.75x/D = 0.5

Unmodified Jet

Modified Jet

Fig. 9. Distribution of RMS axial velocity fluctuations (u0).

Fig. 7. Mixing layer growth for natural and modified jet.


low speed side of the mixing layer similar to that ex-

hibited in the U distribution.. Both the reduced levels ofturbulence and the improved similarity are consistentwith suppression of large-scale structures in the mixing

layer and the reduced randomness in the occurrence, size

and convection velocity of these structures.

3.3.3. Skewness and flatness factor distributions

The distributions of skewness and flatness factors for

both axial and radial velocity fluctuations were obtained

in the modified and unmodified flows. The modified flowshows a significant enhancement in the similarity of the

distribution of the skewness, for both axial (Su, Fig. 11)and radial fluctuations (Sv, Fig. 12), in the mixing layer.As was the case for the mean and RMS distributions,

the improved similarity is most pronounced on the low

speed side of the mixing layer. An increase in magnitude

of Su occurred at most locations in the mixing layer.Such an increase was also evident in Sv.The flatness factor distributions of u and v display

enhanced similarity in the modified flow. Both the

distributions exhibited a reduced magnitude on the low-

speed side of the mixing layer and an increased magni-

tude on the high-speed side resulting in increased sym-

metry across the mixing layer (Figs. 13 and 14).

3.3.4. Reynolds shear stress

The levels of Reynolds shear stress throughout the

mixing layer (Fig. 15) are substantially reduced, with a

19% reduction in the maximum observed level. The

distribution of shear stress across the mixing layer of the

-0.5 0.0 0.5 1.0 1.5y/(x-x0)

-5

0

5

10

Su

Mod

ified

Jet

-10

-5

0

5

Su

Unm

odifi

edJe

t

x/D=5.0x/D=4.5x/D=4.0x/D=3.75x/D=3.5x/D=3.25x/D=3.0x/D=2.75x/D=2.5x/D=2.25x/D=2.0x/D=1.75x/D=1.5x/D=1.25x/D=1.0x/D=0.75

Unmodified Jet

Modified Jet

Fig. 11. Skewness factor distribution for axial velocity fluctuations.

-0.5 0.0 0.5 1.0 1.5y/(x-x0)

-5

0

5

10

15

Sv

Mod

ified

Jet

-20

-15

-10

-5

0

Sv

Unm

odifi

edJe

t

x/D=1.0x/D=1.25x/D=1.5x/D=1.75x/D=2.0x/D=2.25x/D=2.5x/D=2.75x/D=3.0x/D=3.25x/D=3.5x/D=3.75x/D=4.5x/D=5.0

Unmodified Jet

Modified Jet

Fig. 12. Skewness factor distribution for radial velocity fluctuations.

-0.5 0.0 0.5 1.0 1.5y/(x-x0)

0

20

40

60

80

Fu

Mod

ified

Jet

-40

-20

0

20

40

Fu

Unm

odifi

edJe

t

x/D=5.0x/D=4.5x/D=4.0x/D=3.75x/D=3.5x/D=3.25x/D=3.0x/D=2.75x/D=2.5x/D=2.25x/D=2.0x/D=1.75x/D=1.5x/D=1.25x/D=1.0x/D=0.75

Unmodified Jet

Modified Jet

Fig. 13. Flatness factor distribution for axial velocity fluctions.

-0.5 0.0 0.5 1.0 1.5y/(x-x0)

0

40

80

120

160

Fv

Mod

ified

Jet

-80

-40

0

40

80

Fv

Unm

odifi

edJe

t

x/D=1.0x/D=1.25x/D=1.5x/D=1.75x/D=2.0x/D=2.25x/D=2.5x/D=2.75x/D=3.0x/D=3.25x/D=3.5x/D=3.75x/D=4.0x/D=4.5x/D=5.0

Unmodified Jet

Modified Jet

Fig. 14. Flatness factor distribution for radial velocity fluctuations.

-0.5 0.0 0.5 1.0 1.5y/(x-x0)

0

0.1

0.2

0.3

0.4

v'/U

jMod

ified

Jet

-0.2

-0.1

0

0.1

0.2

v'/U

jUnm

odifi

edJe

t

x/D=5.0x/D=4.5x/D=4.0x/D=3.75x/D=3.5x/D=3.25x/D=3.0x/D=2.75x/D=2.5x/D=2.25x/D=2.0x/D=1.75x/D=1.5x/D=1.25x/D=1.0x/D=0.75x/D=0.5

Unmodified Jet

Modified Jet

Fig. 10. Distribution of RMS radial velocity fluctuations (v0).


modified flow exhibited increased similarity, particularly

on the low speed side of the mixing layer. Reduced levels

of Reynolds shear stress are due to a reduction in tur-

bulence levels and mixing, which can be attributed to the

suppression of large-scale structures in the modified

mixing layer.

3.4. Effect of ring position

Fig. 16 shows the effect on the mean velocity distri-

bution at x=D ¼ 3 caused by varying the downstreamposition of the ring over a range of 0:156 x=D6 0:5. It isclear that altering the ring position over this range

caused very little variation in the mean velocity distri-

bution.The effect of altering the ring position on the distri-

bution of axial and radial velocity fluctuations was also

examined by plotting their distributions as measured at

x=D ¼ 3 (not shown here). In both cases, similar to themean velocity distribution, there was no significant

variation between the distributions obtained for each

ring position. From these results it is apparent that the

distributions are effectively independent of ring positionover the range of positions investigated.

3.5. Effect of jet Reynolds number

The jet Reynolds number was varied from 11 450 to

23 500 and the effect of this variation was examined at

x=D ¼ 3. Only small variations in the distribution ofU=Uj, u0=Uj (Fig. 17) or v0=Uj were observed as a resultof varying Reynolds number over this range.

This seems to provide some support for the conclu-

sions of Tong and Warhaft [5], that the modification of

the mixing layer is driven primarily by the alteration in

the mixing layer mean velocity profile and that the ring

vortex shedding frequency does not seem to play a sig-

nificant role in the modification. Further measurementsover a wide range of ReW would be needed before anydefinite conclusions can be made.

3.6. Mixing layer spectra

Figs. 18 and 19 show frequency spectra for a range of

downstream locations for both the unmodified and

modified flows. All the data used in both figures weretaken at radial locations such that U=Uj ¼ 0:5. Fig. 13clearly shows dominant frequencies associated with

vortex formation in the region x=D6 1:5, for the un-modified jet. The evolution of the mixing layer spectra

Fig. 18. Evolution of mixing layer spectra for unmodified jet.

-0.5 0.0 0.5 1.0 1.5y/(x-x0)

-2

-1

0

1

2

3

<uv

>M

odifi

edJe

t

-5

-4

-3

-2

-1

0

<uv

>U

nmod

ified

Jet

x/D=5.0x/D=4.5x/D=4.0x/D=3.75x/D=3.5x/D=3.25x/D=3.0x/D=2.75x/D=2.5x/D=2.25x/D=2.0x/D=1.75x/D=1.5x/D=1.25x/D=1.0x/D=0.75x/D=0.5

Unmodified Jet

Modified Jet

Fig. 15. Reynolds shear stress distribution.

-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5y/D

0

0.4

0.8

1.2

U/U

j

x/D=0.15x/D=0.25x/D=0.5

Fig. 16. Effect of varying the axial location of ring on the mean velocity

distribution at x=D ¼ 3.

0.00

0.02

0.04

0.06

0.08

0.10

u'/U

j

0 0.5 1 1.5y/D

ReD = 11450

ReD = 15750

ReD = 18800

ReD = 23500

Fig. 17. Effect on distribution of u0 at (x=D ¼ 3) of varying jetReynolds number ReD (with ring at x=D ¼ 0:06).


with downstream position indicates the occurrence of

two vortex-pairing events. The spectra for the modified

flow, as shown in Fig. 19, do not display the presence ofdominant frequencies, indicating that the presence of the

ring suppresses the formation of large-scale organised

structures in the mixing layer.

4. Conclusions

The insertion of a thin ring axisymmetrically in themixing layer of a circular jet close to the nozzle exit

modifies the jet structure substantially, as evidenced by

the following results:

• Distributions of U , u0, v0, uv, S and F in the modifiedflow exhibit an improved quality of similarity

throughout the mixing layer, particularly in the re-

gion x=D6 1:25 where similarity in the unmodifiedflow is poor.

• The mixing layer growth rate of the modified jet issubstantially smaller than that of the unmodified

jet. This reduced growth rate is due predominantly

to a decrease in the spread rate of the jet. An increase

in the length of the potential core of the jet has been

observed.

• Turbulence intensities within the jet mixing layer, inboth axial and radial directions, are reduced substan-

tially, by 20% and 15% respectively.

• Different ring positions and Reynolds numbers yieldnearly identical changes to the jet characteristics. Thisindicates that this method of modification is robust,

thus lending itself to industrial applications.

• The suppression of vortex formation achieved withthis type of modification could find an application

when a reduction in aerodynamic jet noise is desired.

• There may also be potential applications in situationsrequiring a reduction in mixing.

Acknowledgement

The continuing support of the Australian Research

Council is acknowledged.

References

[1] S.C. Crow, F.H. Champagne, Orderly structure in jet turbulence,

J. Fluid Mech. 48 (1971) 547–591.

[2] K.B.M.Q. Zaman, A.K.M.F. Hussain, Turbulence suppression in

free shear flows by controlled excitation, J. Fluid Mech. 103 (1981)

133–159.

[3] S. Rajagopalan, R.A. Antonia, Turbulence reduction in the

mixing layer of a plane jet using small cylinders, Exp. Fluids 25

(1998) 96–103.

[4] M. Schrober, H. Fernholz, Turbulence manipulation in wall jets,

in: S. Gavrilakis (Ed.), Advances in Turbulence VI, Kluwer,

Dordrecht, 1996, pp. 529–530.

[5] C. Tong, Z. Warhaft, Turbulence suppression in a jet by means of

a fine wire ring, Phys. Fluids 6 (1994) 328–333.

[6] W.H. Brown, K.K. Ahuja, Jet mixing enhancement by hydro-

dynamic excitation, AIAA J 31 (1993) 599–601.

[7] J. Mi, D.S. Nobes, G.J. Nathan, Influence of jet exit conditions on

the passive scalar field of an axisymmetric free jet, J. Fluid Mech.

432 (2001) 91–125.

[8] A. Roshko, On the development of turbulent wakes from vortex

streets, N.A.C.A. Report 1191, 1954.

[9] M. Kiya, H. Tamura, M. Arie, Vortex shedding from a circular

cylinder in moderate reynolds number shear flow, J. Fluid Mech.

141 (1980) 721–735.

[10] R.A. Antonia, A.J. Chambers, A.K.M.F. Hussain, Errors in

simultaneous measurements of temperature and velocity in the

outer part of a heated jet, Phys. Fluids 23 (5) (1980) 871–874.

Fig. 19. Evolution of mixing layer spectra for modified jet.


Control of an axisymmetric jet using a passive ring

Documents

Transcript of Control of an axisymmetric jet using a passive ring