An experimental investigation of the flow dynamics of streamwise vortices of various

strengths interacting with a supersonic jet

N. Heeb, E. Gutmark, and K. Kailasanath

Citation: Physics of Fluids (1994-present) 26, 086102 (2014); doi: 10.1063/1.4892008

View online: http://dx.doi.org/10.1063/1.4892008
View Table of Contents: http://scitation.aip.org/content/aip/journal/pof2/26/8?ver=pdfcov
Published by the AIP Publishing

Articles you may be interested in

Flow structure and acoustics of supersonic jets from conical convergent-divergent nozzles
Phys. Fluids 23, 116102 (2011); 10.1063/1.3657824

Detailed flow physics of the supersonic jet interaction flow field

Phys. Fluids 21, 046101 (2009); 10.1063/1.3112736

Numerical study of screech generation in a planar supersonic jet

Phys. Fluids 19, 075105 (2007); 10.1063/1.2747225

Experimental study of a supersonic jet-mixing layer interaction

Phys. Fluids 16, 765 (2004); 10.1063/1.1644574

On streamwise vortices in a turbulent wall jet that flows over a convex surface
Phys. Fluids 13, 1822 (2001); 10.1063/1.1366678

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 PHYSICS OF FLUIDS 26, 086102 (2014)

An experimental investigation of the flow dynamics

of streamwise vortices of various strengths interacting
with a supersonic jet
N. Heeb,1,a) E. Gutmark,1,b) and K. Kailasanath2,c)
1
Department of Aerospace Engineering & Engineering Mechanics, University of Cincinnati,
Cincinnati, Ohio 45220, USA
2
Naval Research Laboratory, Washington, D.C. 20375, USA
(Received 23 February 2014; accepted 22 July 2014; published online 11 August 2014)

The results of an experimental study involving the introduction of streamwise vor-

ticity into a supersonic jet are presented. Both streamwise and cross stream PIV
measurements of a baseline jet and three vortex generator configurations of varied
penetration were acquired in the overexpanded, ideally expanded, and underexpanded
regimes. Streamwise vortex pairs were shown to persist no further than two diam-
eters downstream independent of initial magnitude. Integration of the modulus of
streamwise vorticity was shown to better correlate with secondary flow field modi-
fications than maximum values. Reductions in shock cell spacing and downstream
turbulence, and increases in initial spread rate and upstream integrated turbulence
were correlated with increases in streamwise vorticity for all operating conditions.
Streamwise vorticity was shown to achieve opposite effects on shock strength in the
over/under expanded regimes due to introduction of secondary shock structures in
the underexpanded case. Additionally, limited effect on potential core length was
quantified. C 2014 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4892008]

I. INTRODUCTION
Introduction of streamwise vortices into a supersonic jet has been shown to be a practical method
of enhancing mixing. This is in part due to the limited effect of compressibility on streamwise
vorticity as compared to spawnwise structures.1, 2 Additionally, the underlying vortex dynamics
particular to axisymmetric jets3 are highly dependent on the jet’s initial conditions,4 a feature which
is easily modified by streamwise vortices. Multiple technologies have been developed with the aim of
producing streamwise vorticity to enhance jet mixing. For example, tabs,2 fluidic injection,5 notched
nozzles,6 and nozzle cross section7 have all been used to this end. The primary challenge facing
adoption of these technologies is achieving appreciable mixing enhancement without incurring large
performance penalties.
In an effort to achieve this goal, a blending of tab and notched nozzle design methodology was
performed resulting in vortex generators. Within the literature, vortex generators are also referred
to as serrations, chevrons, and sometimes encompass notches and tabs as well. Zaman et al.1
were able to generate streamwise vorticity and enhance mixing in a supersonic jet using vortex
generators. The quasi-periodic shock cell structure was also found to be significantly shortened and
weakened. Furthermore, they identified the primary mechanism behind streamwise vortex production
as the circumferential pressure gradient generated by the presence of the vortex generator in the
flow field. Hileman and Samimy8 identified an increase in the regularity of spanwise large scale
structure formation, resulting from the introduction of streamwise vorticity. The potential core

a) Electronic mail: heebns@mail.uc.edu

b) Electronic mail: gutmarej@ucmail.uc.edu
c) Electronic mail: kailas@lcp.nrl.navy.mil

1070-6631/2014/26(8)/086102/25/$30.00 26, 086102-1 

C 2014 AIP Publishing LLC

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-2 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

was also found to be shortened as compared to an unmodified jet. Opalski et al.9 found that
independent of initial vorticity magnitude, at roughly 1.5 diameters downstream of the nozzle exit
all investigated vortex pairs had reached similar peak values. Additionally, they found that trends
relating to streamwise vorticity introduction identified in unheated jets were directly comparable
to those obtained at elevated temperatures. Behrouzi and McGuirk10 identified streamwise vortex
proximity as detrimental to enhanced mixing due to negative vortex interaction. Additionally, they
found the complexity of induced shock cell structure increased with streamwise vortex introduction
in underexpanded jets. Similar findings were reported by Tide and Srinivasan11 for less aggressively
designed vortex generators. Bridges et al.12 identified issues in generating streamwise vorticity in
overexpanded conditions due to the inherent contraction of the flow field at the nozzle exit. Similar
findings were presented by Munday et al.13 and Kuo et al.14 Bridges et al.12 also found that the peak
vorticity magnitudes roughly two diameters downstream of the nozzle were independent of initial
values, though the spatial extent of the structures varied depending on the initial configuration, similar
to what was presented by Opalski et al.9 Burak et al.15 noted the presence of a secondary counter
rotating streamwise vortex pair in addition to the primary pair reported in most studies. Investigation
of streamwise vorticity introduction into supersonic jets has typically been performed using sonic
nozzles, limiting results to sonic and under expanded conditions. Therefore, limited investigation
regarding the overexpanded regime has been performed. In fact, of the above mentioned studies, Kuo
et al.,14 Bridges et al.,12 and Munday et al.13 were the only studies which investigated overexpanded
operation, of which only Bridges et al.12 quantified streamwise vorticity. Consequently, additional
investigation using convergent divergent nozzles is needed.
The objective of the present work is to improve understanding of the effect of streamwise
vorticity on the flow field of a supersonic jet. To this end, several vortex generator configurations
along with an unmodified nozzle were experimentally investigated during imperfectly expanded
operation, as outlined in Sec. II. In particular, emphasis is placed on the overexpanded regime,
though ideally and underexpanded conditions are included. The acquired streamwise and cross
stream particle imaging velocimetry (PIV) results are presented in Sec. III. Therein, a comparison
of the resulting vorticity fields is provided, followed by an investigation into the effect of vorticity
magnitude on the overall flow field along with specific jet features such as shock cell spacing and
potential core length. Finally, Sec. IV provides a conclusion of the work presented.

II. EXPERIMENTAL METHODS

A. Facility
The Aeroacoustic Test Facility (ATF) at the University of Cincinnati was used for all experiments
presented in this work. The AFT consists of a 24 ft × 25 ft anechoic chamber and a coaxial test
rig which was offset from the chamber’s centerline. Due to the configuration of the nozzles which
are being modeled, only the core section of the coaxial rig was used to simulate the exhaust jets.
During PIV measurements, secondary flow was used at very low speeds (M ≈ 0.05) to improve the
consistency of tracer particles in the ambient air, which in turn increases the data reliability within
the shear layer of the jet following the methodology of Samimy et al.16 More complete information
about the test facility and the test rig can be found in Callender et al.17

B. Model hardware
1. Baseline nozzle
This study uses a single biconic converging-diverging nozzle to represent a military style
variable area nozzle, a schematic of which is shown in Figure 1. The nozzle’s area ratio is 1.181,
which corresponds to a design Mach number of Md = 1.5. The nozzle’s exit diameter, Dd , is 2.868 in.
which roughly equates to a 1/8th scale model. This nozzle was originally studied by Munday et al.18
and additional information about the design and flow features can be found therein.
Munday et al.18 showed that the nozzle under investigation was prone to screech: an aeroacoustic
feed back mechanism which can result in flow field modification in particular shortening of the

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-3 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

FIG. 1. Baseline nozzle geometry (dimensions in inches).

potential core,19, 20 especially during overexpanded operation. In the past, nozzle trailing edge
modifications have been used to eliminate screech and the associated flow field modifications.6, 20–22
Effectiveness of this technique has been shown to be limited to the underexpanded flow regime
as an excess of pressure at the nozzle exit is needed to expel fluid through the notches. This in
turn results in generation of streamwise vortex pairs and modification of the jet cross section.6, 21
Consequently, the use of notches was deemed unacceptable for the current study as the introduction
of vorticity would muddle baseline flow field comparisons and as the study focuses primarily on the
overexpanded regime.

2. Vortex generators
The main parameters necessary to define the geometry of a vortex generator are length, pene-
tration, and width (as defined in Figure 2). It is important to note that due to the conical diverging
section of the nozzle, penetration is not measured from the nozzle lip line. Instead, penetration is
defined as the normal distance between the fictitiously extended nozzle inner contour and the vortex
generator tip. The vortex generator length is also affected by this definition and is consequently
defined as the distance between the nozzle lip and the vortex generator tip, measured along the
fictitiously extended nozzle inner contour.
To facilitate the measurement of multiple configurations, three sets of vortex generators were
constructed using fused deposition modeling (FDM) rapid prototyping. Based on previously pre-
sented streamwise vorticity results,9, 12 it was chosen to vary penetration while holding the length

FIG. 2. Definition of vortex generator geometric parameters.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-4 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

TABLE I. Vortex generator parameters for Md = 1.5 nozzles.

Dd [in.] Number Penetration/Dd Length/Dd Width [◦ ]

Low 2.868 12 0.04 0.26 30

Medium 2.868 12 0.08 0.26 30
High 2.868 12 0.12 0.26 30
Bridges et al.12 4.840 12 0.04 0.26 30
Bridges et al.12 4.840 12 0.09 0.26 30
Kuo et al.14 0.676 12 0.05 0.23 30
Schlinker et al.23 2.000 12 0.04 ... ...

and width constant. The three values of penetration were selected as multiples (1, 2, and 3) of the
previously measured18 shear layer thickness. Correspondingly, the three sets of vortex generators
are denoted “low,” “medium,” and “high” throughout this work. The geometric specifics of each of
these vortex generators and several other vortex generator geometries12, 14, 23 are located in Table I,
for comparison purposes.

C. Test matrix
Operating conditions were monitored to ensure that the measured values are within 1.5% of the
target value during experiments. Following the procedures outlined by the AIAA,24, 25 the facility
parameters were post-processed to determine the total uncertainty of each measured quantity. For
all results shown here, the jet velocity was determined with 95% confidence to be within ±7.3
m/s of the target value. Two overexpanded, the design condition, and one underexpanded condition
were chosen for investigation following the previously selected conditions of Munday et al.18 The
corresponding nozzle total to ambient pressure and temperature ratios (NPR and NTR) are listed
in Table II, along with the fully expanded parameters of Mach number, Mj , jet velocity, uj , and
Reynolds number, Rej . Additionally, the fully expanded diameter to design diameter ratio, Dj /Dd ,
and non-ideally expand parameter, β, are given.

D. PIV measurements
Velocity and turbulence measurements were performed using a LaVision Flow Master system.
This system was used to acquire both streamwise (aligned with the centerline) and cross stream
measurement planes. The streamwise planes were collected using the two dimensional PIV config-
uration, resulting in measurement of streamwise and radial components of velocity, while the cross
stream planes were acquired using the stereoscopic PIV (SPIV) configuration, affording measure-
ment of all three velocity components. A data acquisition computer, a 120 mJ dual pulse New Wave
Research Nd:Yag laser, and two Imager Intense CCD 1 megapixel 12-bit cameras were the main
hardware of the Flow Master system. Additionally, beam forming optics (cylindrical and spherical
lenses) were employed to create a laser sheet of the desired width and thickness. Due to the reduced
residence time of the particles within the laser sheet in the cross stream application, the laser sheet
was chosen to be approximately 2 mm in thickness and a time delay of 1.5μs was used. As the laser

TABLE II. Investigated operating conditions.

NPR NTR Mj uj [m/s] Dj /Dd β Rej /106

2.50 1.00 1.22 373 0.94 0.87 2.60

3.00 1.00 1.36 403 0.96 0.64 3.17
3.67 1.00 1.50 433 1.00 0.00 3.93
4.00 1.00 1.56 444 1.02 0.42 4.30

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-5 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

FIG. 3. PIV measurement locations. Cross stream planes are perpendicular to the page and clustered towards the near nozzle
region. Streamwise planes overlap in the axial direction and encompass only half of the jet downstream of 2.5Dd .

sheet was aligned with the dominant velocity direction in the streamwise measurements, the laser
sheet was reduced to approximately 0.5 mm thick and a time delay in the range of 5μs was used.
Throughout the measurements a Nikon lens (28, 35, or 50 mm), a 532 nm wavelength filter, and a lens
adapter that allowed micro adjustment of the angle at which the lens was positioned were attached
to the cameras. These adapters allowed angular corrections to be introduced for non-perpendicular
observation of the measurement plane following the Scheimpflug principle. For the cross stream
measurements, the cameras were located with a 45◦ viewing angle of the laser sheet, following the
results of Alkislar et al.26 The cameras were arranged side by side with a nearly perpendicular view
of the laser sheet for the streamwise measurement, and only slight angular adjustment of the lenses
was required to ensure ample field of view overlap.
Due to the size of the jet column and the desire to measure the spatial evolution of quantities
such as streamwise vorticity and shock spacing, multiple locations of data were collected. Use of a
three axis traverse system allowed the entire PIV setup to be traversed in the axial direction to ease
experimental setup. Nine cross stream planes clustered towards the nozzle exit and four streamwise
axial positions, as outlined in Figure 3, were chosen for this task. Due to jet spread and camera
limitations, the aft streamwise positions were elevated to encompass only half of the jet, and the
field of view of the aft cross stream measurements was enlarged. In all cases, the spatial resolution
of the post-processed vector fields was below 2 mm.
In an effort to remove reflections off hardware, laser masking and surface treatment consisting
of a mat black undercoat and a Rhodamine-B suspension as the finish treatment were used. Olive oil
tracer particles that were atomized by custom built Laskin nozzles were used for flow visualization.
The corresponding particle sizes were roughly 1μm, resulting in a worst case Stoke’s number
of 0.08.
Pre-processing, vector calculation, and post-processing of the data was performed using LaVi-
sion’s DaVis software. Pre-processing of the raw image pairs was performed following the method-
ology of Deen et al.27 to improve the vector calculations due to the presence of any hardware
reflections that were not eliminated using the methods discussed above. Vector calculation was
performed using a multi-pass method with an initial window size of 64 × 64, which was reduced to
a 16 × 16 window with a 50% overlap for the final iteration. Post-processing of the resulting vector
fields consisted of applying an allowable vector range and computation of ensemble statistics.
Due to the low data rate of the PIV hardware (5 Hz), only mean quantities will be presented
here. Following the methodology of Carr et al.,28 it was determined from initial data sets that 500
image pairs were required near nozzle (x/D < 5) and 1000 downstream to acceptably converge
the mean results. The analysis estimated the precision error of the results to be 1.2%uj and 0.9%uj
for the mean velocity and RMS fluctuating velocity respectively at a 95% confidence level. Total
experimental error was estimated based on the work of Lazar et al.29 and the iTTC guidelines30 to
be below 7%uj for the time average velocity magnitude and less than 15%uj for the RMS fluctuating
velocity error. This estimate includes facility set point error, precision error, and the errors from

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-6 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

the PIV measurement itself such as centerline misalignment, camera calibration, and particle lag
through shock waves.

III. RESULTS
Before discussing the flow field results, far-field narrowband acoustic spectra, experimentally
acquired at an upstream observation angle following the methodology of Heeb et al.31 are presented
in Figure 4 for reference. This allows identification of operating conditions where the aeroacoustic
feedback mechanism of screech occurs which, as noted above, can lead to flow field modification.
The presented frequency range was normalized using the fully expanded jet velocity and diameter
into the Strouhal number and has an unnormalized resolution of 50 Hz. For all of the considered
jet Mach numbers and configurations, the presence of screech is identifiable in the 0.2–0.7 Strouhal
number range. Investigation of the baseline spectra indicate that the Mj = 1.36 and 1.50 conditions
are dominated by screech, while the Mj = 1.22 and 1.56 conditions exhibit screech amplitudes that
only slightly exceed the broadband shock associated noise peak amplitude. On the other hand, none
of the vortex generator spectra are dominated by a screech tone as the broadband shock associated
noise exceeds or is roughly equivalent to the measured screech amplitudes. The results of André
et al.32 indicate a direct relation between screech amplitude and the extent of jet column undulation.
Consequently, it is thought that screech will have a limited effect on the investigated results, except
for the baseline jet at the Mj = 1.36 and 1.50 conditions, due to the large screech amplitude which
dominates the acoustic spectrum.
The focus of this study is on the effect of streamwise vorticity magnitude on the flow features
of a supersonic jet. Streamwise vorticity was determined though post-processing of cross stream
PIV measurements. Visualization of normalized streamwise vorticity near the nozzle is presented

120 120
Base Base
110 110
Low Low
100 Med 100 Med
SPL [dB ]

SPL [dB ]

High High
90 90
80 80
70 70
60 60
50 50
−1 0 1 −1 0 1
10 10 10 10 10 10
Strouhal Num ber fs Dj u−
j
1
Strouhal Num ber fs Dj u−
j
1

(a) Mj = 1.22 (b) Mj = 1.36

120 120
Base Base
110 110
Low Low
100 Med 100 Med
SPL [dB ]

SPL [dB ]

High High
90 90
80 80
70 70
60 60
50 50
−1 0 1 −1 0 1
10 10 10 10 10 10
Strouhal Num ber fs Dj u−
j
1
Strouhal Num ber fs Dj u−
j
1

(c) Mj = 1.50 (d) Mj = 1.56

FIG. 4. Far-field narrowband acoustic spectra illustrating screech amplitudes. Measured 47Dd from the nozzle exit and 35◦
from the upstream jet axis.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-7 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

Base. Low Base. Low

5

|ω z |D j
0 uj

−5
High Med. High Med.

(a) Mj = 1.22 (b) Mj = 1.50 (c) Contour Scale

FIG. 5. Contours of streamwise vorticity in a cross stream plane at z/Dd = 0.5. Quadrants depict the baseline, low, medium,
and high penetration vortex generators progressing clockwise from the upper left, respectively.

in Figure 5 for the baseline and vortex generator configurations at an overexpanded and ideally
expanded operating condition. For reference, the angular location of the vortex generator tip is
highlighted by a point, and the nozzle exit is illustrated as a dashed line. As seen, each individual
vortex generator introduces a counter rotating vortex pair, even the low penetration vortex generators
at the overexpanded condition though, with a low magnitude. Additionally, an individual vortex pair
for each of the vortex generator configurations was isolated and enlarged in Figure 6 to aid in
visualization.
In the overexpanded condition, penetration is directly correlated with near nozzle maximum
vorticity and vortex size. The limited effect of the low penetration vortex generators is linked to
insufficient effective penetration resulting from the overexpansion and consequent inward turning
of the flow.12, 13 Due to the spatial extent of the vortices generated by the high penetration vortex
generators there is limited azimuthal space for vortex growth leading to the highly elongated vortex

(a) Low. Mj = 1.22 (b) Medium. Mj = 1.22 (c) High. Mj = 1.22

(d) Low. Mj = 1.36 (e) Medium. Mj = 1.36 (f) High. Mj = 1.36

−8 −6 −4 −2 0 2 4 6 8
|ω z |D j
uj

(g) Contour scale

FIG. 6. Enlarged view of individual streamwise vortex pairs along with corresponding velocity vector field. z/De = 0.5.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-8 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

20
Base
Low
15 Med.
High
|ωx |Dj uj −1

10

0
0 0.5 1 1.5 2 2.5 3 3.5
zDj −1

FIG. 7. Variation of maximum streamwise vorticity with downstream position. Mj = 1.22.

shape and the relatively high speed expulsion of fluid, as indicated by the vectors at the top and
bottom of Figure 6(c). Additionally, interaction between vortex pairs occurs due to limited space.
This results in reduced vortex persistence and peak magnitudes downstream, which will be discussed
below. The ideally expanded configuration does not show the same correlation between penetration
and vorticity, with all three configurations possessing similar peak magnitudes. The radial extent
of vortices did increase with penetration, though a direct correlation to the azimuthal size was not
achieved. This was again thought to be caused by the spatial proximity of the vortex pairs. In
addition to the main vortex pair, each configuration induces small secondary vortex pairs located
between adjacent main pairs. These secondary structures rotate in an opposite fashion compared to
the primary pair, as is particularly evident for the low penetration vortex generator configuration
(Figure 6(d)). Similar vorticity features were reported in the computational study of Burak et al.15
In an effort to quantify the effect of penetration on the vorticity field, the maximum vorticity was
initially tabulated at each available axial location as presented in Figures 7 and 8. An exponential
function was also fit to each of the vortex generator configurations to illustrate the decay rate
following the results of Alkislar et al.33 As expected from the above vorticity contour plots, the
overexpanded results show initial maximum vorticity that is much higher for the medium and high
penetration vortex generators than the baseline and low penetration vortex generators. From that
point downstream, the maximum vorticity exponentially decays to an indistinguishable level by z/Dj
= 2.5. The exponential fits indicate that the high penetration vortex generators’ vorticity decays at a
rate roughly 1.5 times faster than the medium penetration vortex generators. This is an indication of
the negative vortex interaction noted above. For all configurations, independent of initial vorticity
level, the peak value decays to the baseline in the 1.5-2 Dj range. Similar results were presented
by Opalski et al.9 and Bridges et al.12 At the ideally expanded condition, the maximum vorticity
and the decay rates roughly collapse, indicating an independence of penetration. Though maximum

20
Base
Low
15 Med.
High
|ωx |Dj uj −1

10

0
0 0.5 1 1.5 2 2.5 3 3.5
zDj −1

FIG. 8. Variation of maximum streamwise vorticity with downstream position. Mj = 1.50.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-9 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

2.5
Base
2 Low
Med.
High
Γ(Dj uj )−1

1.5

0.5

0
0 0.5 1 1.5 2 2.5 3 3.5
zDj −1

FIG. 9. Variation of integrated vorticity modulus () with downstream position. Mj = 1.22.

magnitudes are similar, the results presented in Figure 6 indicate that penetration does have an effect
on vortex size. Therefore, maximum vorticity appears to be an over simplified performance metric
due to the exclusion of vortex extent.
In an effort to more accurately describe the evolution of the vorticity field, a metric that included
both vorticity magnitude and spatial extent was used. This metric involved integrating the modulus
of streamwise vorticity over a given jet cross section as defined in Eq. (1):

 = |ωz | dcs. (1)
cs

The resulting evolution profiles are presented in Figures 9 and 10, which upon detailed investigation
were not found to decay exponentially. As expected, the primary difference between the newly
calculated metric and the maximum vorticity is a better representation of the spatial extent of the
induced streamwise vorticity pairs. This is particularly evident when comparing the near nozzle
high and medium penetration vorticity fields in the overexpanded case. In all cases, the near nozzle
values are shown to be function of penetration. Consequently, the low, medium, and high penetration
configurations will also be referred to as low, medium, and high vorticity cases throughout the
remainder of this work.
The most discernible effect of the induced streamwise vorticity on the jet flow field is modifica-
tion of the cross sectional shape. This is illustrated through the comparison of the spatial evolution
of jet cross section, as shown in Figures 11 and 12, for the overexpanded Mj = 1.22 and ideally
expanded Mj = 1.50 conditions. These figures are arranged such that the baseline configuration
is located in the upper left quadrant and in each consecutive quadrant, progressing clockwise, the
induced vorticity magnitude is increased. The comparison for the overexpanded condition shows a
nearly axisymmetric baseline jet which is relatively unchanged by the low vorticity configuration as

3
Base
2.5 Low
Med.
2 High
Γ(Dj uj )−1

1.5

0.5

0
0 0.5 1 1.5 2 2.5 3 3.5
zDj −1

FIG. 10. Variation of integrated vorticity modulus () with downstream position. Mj = 1.50.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-10 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

Base. Low

High Med.

(a) z/Dj = 0.5 (b) z/Dj = 1.0 (c) z/Dj = 1.5

(d) z/Dj = 2.0 (e) z/Dj = 2.5 (f) z/Dj = 3.0

1.0

0.8

0.6 |u|
uj

0.4

0.2

(g) z/Dj = 4.0 (h) z/Dj = 6.0 (i) Contour Scale

FIG. 11. Cross-sectional velocity magnitude evolution. Quadrants depict the baseline, low, medium, and high penetration
vortex generators progressing clockwise from the upper left, respectively. Mj = 1.22.

expected from Figure 5. Further increase in vorticity levels resulted in large crenelation of the jet
shear layer, the radial extent of which is correlated with the penetration level. These crenelations, or
jetlets,34 evolve and reduce with downstream propagation due to mixing, but are discernible at loca-
tions up to z/Dj = 3.0. Asymmetries in the flow field are visible at z/Dj = 4.0, but further downstream
at z/Dj = 6.0 the jet cross sections are nearly indistinguishable, with only a slight crenelation caused
by the high vorticity case. The comparison of the ideally expanded operating condition (Figure 12)
indicates appreciable crenelation of the shear layer by all three configurations. Additionally, the
secondary vortex pairs discussed above complicate the primary expulsion of fluid and lead to aux-
iliary lobes located at the extent of the main jetlet as seen in the near nozzle planes. Consistent
with the results in Figures 6(d)–6(f), the radial expulsion of fluid produced by the low vorticity
configuration is shorter in radial distance, but larger in azimuthal extent than either the higher two
levels. Asymmetries exist downstream of z/Dj = 2.5, though the typical jetlet pattern is washed out
by z/Dj = 4.0 for all configurations. Similar to the overexpanded condition, the low and medium
vorticity levels display nearly axisymmetric cross sections by z/Dj = 6.0, while a non-circular flow
field is still present for the high vorticity configuration.
As the gross effect vortex generators have an axial jet features, in particular the shock cell
structure, are not determinable using the cross stream measurements, streamwise contours of time
mean axial velocity and axial fluctuating velocity are presented in the following figures.
Figure 13 depicts the mean and fluctuating velocity in the upper and lower halves, respectively,
for the four configurations at the overexpanded condition of Mj = 1.22. Additionally, enlarged views

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-11 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

Base. Low

High Med.

(a) z/Dj = 0.5 (b) z/Dj = 1.0 (c) z/Dj = 1.5

(d) z/Dj = 2.0 (e) z/Dj = 2.5 (f) z/Dj = 3.0

1.0

0.8

0.6 |u|
uj

0.4

0.2

(g) z/Dj = 4.0 (h) z/Dj = 6.0 (i) Contour Scale

FIG. 12. Cross-sectional velocity magnitude evolution. Quadrants depict the baseline, low, medium, and high penetration
vortex generators progressing clockwise from the upper left, respectively. Mj = 1.50.

of the near nozzle mean velocity are shown in Figure 14. One of the most noticeable features of the
baseline overexpanded jet is the large Mach disk and corresponding slip lines located near the jet
centerline. Furthermore, the shock pattern does not display the typical double diamond structure of
conical nozzles due to coalescence of the individual diamonds, as was originally shown in Munday
et al.18 The fluctuating velocity of the baseline jet illustrates a turbulence distribution which grows
with downstream position and which peaks near the end of the potential core. The maximum value
is approximately 17%, which corresponds well with the results of Bridges et al.12 at a similar
operating condition. As expected from the above cross stream results, the low vorticity configuration
has limited effect on the velocity field, though a slight decrease in peak turbulence is observed.
On the other hand, the medium level configuration drastically affects the flow field. The shock cell
spacing is reduced, the centerline Mach disk strength is reduced, the initial spread rate of the jet
is increased, and finally a large area of increased turbulence is generated in the near nozzle region
while the downstream peak levels and radial extent are reduced. The high vorticity configuration
achieves similar modifications, but to a more drastic level. The primary difference between the two
is the large secondary turbulent structure related to the expulsion of fluid between the individual
vortex pairs and an increased downstream peak turbulence level.
Similarly, streamwise planes of time mean, and fluctuating velocity fields are presented in
Figure 15, along with enlargements of the near nozzle time mean velocity field in Figure 16 for
the ideally expanded condition. The baseline jet displays the typical double shock diamond pattern,
though the large centerline Mach disk present in the overexpanded regime is no longer apparent.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-12 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

FIG. 13. Contours of mean and fluctuating axial velocity located in the plane containing the vortex generator valleys.
Mj = 1.22.

Additionally, the region of high turbulence is increased in axial and radial extent along with a slight
increase in peak amplitude (19%). The low vorticity configuration appreciably affects the flow field
at the design condition: the shock cell spacing is reduced, the initial centerline Mach disk strength is
increased, secondary flow features are introduced due to expulsion of fluid between the vortex pairs,
the upstream spread rate is increased, near nozzle turbulence levels are increased in radial extent and
amplitude, and finally downstream turbulence levels are drastically reduced. The medium vorticity
configuration introduces similar flow features. The primary changes are an introduction of a complex
shock structure, increased centerline Mach disk strength, upstream turbulence, and upstream spread
rate, along with decreased downstream turbulence and shock strength. Finally, the high vorticity
configuration further complicates the shock cell structure and shortens its spacing, the expulsion of
fluid between vortex pairs is of much higher velocity and the corresponding turbulence is increased,
and the downstream turbulence levels are increased compared to the other two configurations.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-13 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

0.6 0.6
r Dj −1

r Dj −1
0.4 0.4

0.2 0.2

0.0 0.0
0.5 1.0 1.5 2.0 0.5 1.0 1.5 2.0
z Dj −1 z Dj −1

(a) Baseline (b) Low

0.6 0.6
r Dj −1

r Dj −1
0.4 0.4

0.2 0.2

0.0 0.0
0.5 1.0 1.5 2.0 0.5 1.0 1.5 2.0
z Dj −1 z Dj −1

(c) Medium (d) High

0.5 0.6 0.7 0.8 0.9 1.0 1.1

u z u j −1

FIG. 14. Enlarged view of the near nozzle mean velocity field presented in Figure 13. Mj = 1.22.

Additionally, the introduction of any level of vorticity appears to have increased the potential core
length compared to the baseline. Considering the acoustic results presented above, this is possibly
due to screech reducing the potential core length of the baseline as originally outlined by Glass.19
Comparison of radial profiles of axial velocity at several downstream positions are presented in
Figures 17 and 18 to quantitatively enhance the comparisons given above. In each of the figures, the
data were extracted from the streamwise PIV results, and the open symbols correspond to data taken
in the valley plane of the vortex generators, while the filled symbols correspond to the tip plane.
Additionally, only every fourth data point is presented to ease viewing. Cross stream SPIV baseline
results are also included as a solid line to validate the independently acquired results. As seen, the
comparison between the two measurements is excellent. Deviations are seen near the centerline in the
slip portion of the jet, which is in part due to lack of resolution of the streamwise measurement. Addi-
tionally, it is possible that due to the unsteady nature of air supply system the Mach disk is not always
present per the work of Irie et al.35 and consequently the time average results could display large
difference in the slip line region. As expected from the above contour plots the low vorticity config-
uration has almost no effect on the baseline velocity field at the overexpanded condition (Figure 17).
Near the nozzle exit, the medium vorticity case slightly modifies the jet radius by decreasing the
width in the tip plane and increasing the width at the valley plane. The high vorticity configuration
performs in the same manner, but with larger modifications. Progressing downstream, the velocity
profiles become more similar and by z/Dj = 2.5 there are limited differences, indicating the jet is
nearly axisymmetric. Though axisymmetric, the medium, and high configurations still increased
jet spread up until the axial location of z/Dj = 6.0, where all four profiles are nearly identical.
Similar results are shown for the ideally expanded operating condition (Figure 18) except that the
low vorticity case has an effect on the baseline flow field, particularly at the upstream positions.
Comparisons of the axial turbulence distributions are presented in Figures 19 and 20 in a
similar manner to the velocity figures above. Due to an artificial inflation of the fluctuating velocity

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-14 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

FIG. 15. Contours of mean and fluctuating axial velocity located in the plane containing the vortex generator valleys. Mj =
1.50.

component by inconsistent tracer particle entrainment from the ambient flow field, the profiles have
been truncated at positions where the axial velocity was lower than 10 m/s. Again as expected, the low
vorticity case has limited influence on the overexpanded jet. Increases in vorticity further increase
the near nozzle amplitude and radial extent of high turbulence, both inward and outward, compared
to the baseline jet. Additionally, the location of maximum turbulence is moved radially outward in
the valley plane. This effect is somewhat reduced by z/Dj = 2.5 where the peak value of turbulence
is roughly the same for all configurations. Further downstream, the baseline and low vorticity case
have slightly higher peak turbulence values, though the profiles are roughly the same shape. Results
are again similar for the ideally expanded case except that the difference between the downstream
peak levels is increased and the low vorticity configuration is able to achieve modification. The
increase in bulk turbulence upstream and decrease down stream is attributed to increased small scale

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-15 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

0.6 0.6
r Dj −1

r Dj −1
0.4 0.4

0.2 0.2

0.0 0.0
0.5 1.0 1.5 2.0 0.5 1.0 1.5 2.0
z Dj −1 z Dj −1

(a) Baseline (b) Low

0.6 0.6
r Dj −1

r Dj −1
0.4 0.4

0.2 0.2

0.0 0.0
0.5 1.0 1.5 2.0 0.5 1.0 1.5 2.0
z Dj −1 z Dj −1

(c) Medium (d) High

0.5 0.6 0.7 0.8 0.9 1.0 1.1

u z u j −1

FIG. 16. Enlarged view of the near nozzle mean velocity field presented in Figure 15. Mj = 1.50.

production near the nozzle, resulting in reduced large scale structure overall energy and growth rate,
which in turn reduces the bulk turbulence downstream.36–38
The complexity of the shock cell/Mach disk system makes direct comparison of the centerline
profiles of the different configurations difficult as seen in Figure 21. In Figure 21, the velocities
were normalized by the mean axial velocity within the potential core and the axial coordinate
was normalized by the potential core length, the determination of which will be outlined below.

Base
Low
1.0
Med
High
0.8
r/Dj

0.6

0.4

0.2

0.0
0.0 0.5 1.0 0.0 0.5 1.0 0.0 0.5 1.0 0.0 0.5 1.0 0.0 0.5 1.0
uz uj −1 uz uj −1 uz uj −1 uz uj −1 uz uj −1

(a) z/Dj = 0.5 (b) z/Dj = 1.5 (c) z/Dj = 2.5 (d) z/Dj = 4.0 (e) z/Dj = 6.0

FIG. 17. Radial profiles of axial velocity at several downstream locations. Mj = 1.22. Filled symbols: Vortex generator tip
profile. Open symbols: Vortex generator valley profile.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-16 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

Base
Low
1.0
Med
High
0.8
r/Dj

0.6

0.4

0.2

0.0
0.0 0.5 1.0 0.0 0.5 1.0 0.0 0.5 1.0 0.0 0.5 1.0 0.0 0.5 1.0
uz uj −1 uz uj −1 uz uj −1 uz uj −1 uz uj −1

(a) z/Dj = 0.5 (b) z/Dj = 1.5 (c) z/Dj = 2.5 (d) z/Dj = 4.0 (e) z/Dj = 6.0

FIG. 18. Radial profiles of axial velocity at several downstream locations. Mj = 1.50. Filled symbols: Vortex generator tip
profile. Open symbols: Vortex generator valley profile.

As was discussed above, the examined overexpanded conditions (Figures 21(a) and 21(b)) exhibit
a single shock cell structure as is apparent from the centerline profiles. On the other hand, the
double shock diamond structure extends through roughly half of the jet potential core at the ideally
and under expanded conditions. Afterwords, a single shock structure persists past the end of the
theoretical potential core termination. Reductions in shock strength and spacing are apparent though
visual comparison of the amplitude and frequency of velocity fluctuation at the most overexpanded
condition. Qualitative comparison of the other operating conditions is inconclusive. Consequently,
the axial profiles were used to quantitatively determine shock cell spacing, shock strength, and
potential core length as outlined in Figure 22.
The potential core length was extracted by least squares fitting a slightly modified profile
developed by Lau et al.,39 shown below in Eq. (2), to the measured centerline velocity:
⎧
⎨1 z ≤ zc
u z (z) /u m =   : α = 1.5, u m = u z (z ∈ [0, z c ]). (2)
⎩1 − exp α
z > zc
1−z/z c

Base
Low
1.0
Med
High
0.8
r/Dj

0.6

0.4

0.2

0.0
0.05 0.10 0.15 0.05 0.10 0.15 0.05 0.10 0.15 0.05 0.10 0.15 0.05 0.10 0.15
uz uj −1 uz uj −1 uz uj −1 uz uj −1 uz uj −1

(a) z/Dj = 0.5 (b) z/Dj = 1.5 (c) z/Dj = 2.5 (d) z/Dj = 4.0 (e) z/Dj = 6.0

FIG. 19. Radial profiles of axial turbulence intensity at several downstream locations. Mj = 1.22. Filled symbols: Vortex
generator tip profile. Open symbols: Vortex generator valley profile.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-17 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

Base
Low
1.0
Med
High
0.8
r/Dj

0.6

0.4

0.2

0.0
0.05 0.10 0.15 0.05 0.10 0.15 0.05 0.10 0.15 0.05 0.10 0.15 0.05 0.10 0.15
uz uj −1 uz uj −1 uz uj −1 uz uj −1 uz uj −1

(a) z/Dj = 0.5 (b) z/Dj = 1.5 (c) z/Dj = 2.5 (d) z/Dj = 4.0 (e) z/Dj = 6.0

FIG. 20. Radial profiles of axial turbulence intensity at several downstream locations. Mj = 1.50. Filled symbols: Vortex
generator tip profile. Open symbols: Vortex generator valley profile.

The modification involves the inclusion of the um term, which accounts for the non-constant velocity
within the potential core due to shock waves and non-ideal expansion. The minimum coefficient of
determination of all of the least squares fits was 0.96, indicating excellent agreement with the profile
of Lau et al.39 With the appropriate fit determined, the potential core length was easily found as
indicated by the square in Figure 22.
The result of completing this routine for all available streamwise data is shown in Figure 23.
Additionally, the potential core length prediction of Lau et al.,39 shown below in Eq. (3), was
included for comparative purposes:
z c (M j )
= 4.2 + 1.1M j 2 . (3)
Dj
The prediction closely follows a majority of the data and captures the trend of increasing zc with
Mj well, though the baseline jet at both the Mj = 1.36 and Mj = 1.50 conditions significantly
deviates from the profile. This is thought to be a result of screech shortening the potential core length
following the results of Glass19 and Bridges and Wernet20 at these specific conditions. Interestingly,
the conditions in question were the only cases where screech was identified to dominate the acoustic
spectrum, though screech occurred for all configurations and operating conditions (Figure 4). This
indicates that potential core length reduction and screech are not necessarily mutually inclusive, but
instead requires screech to be the dominant noise source for reduction to occur. The low vorticity
case at Mj = 1.36 illustrates this particularly well, as a distinct screech tone is easily identified in
the acoustic spectrum yet the measured potential core length is predicted within 1% by the relation
of Lau et al.39 This follows the results of André et al.32 discussed above, which indicate large
amplitude screech is required for significant flow field modification. The comparison of the results
to the prediction of Lau et al.39 indicates vorticity has a limited effect in terms of potential core
length change, as compared to the effect of screech. This possibly explains the conflicting reports of
shortening9, 12, 40 and lengthening13, 41 of the potential core by introduction of streamwise vorticity
in the supersonic regime, i.e., reduction of screech opposes the effect of increased mixing which in
turn allows an increase or decrease in zc depending on the severity of the individual effect.
The shock cell spacing was also determined from axial velocity profiles following the method
outlined in Heeb et al.31 This involves finding all of the axial locations corresponding to local peak
velocity values, as shown by the circles in Figure 22, subtracting subsequent locations, and then
averaging to determine the mean shock cell spacing. Due to difficulty determining local maximums
near the end of the potential core as well as the effect of screech on the downstream shock structure,22
a maximum of 7 shock cells were used to determine the mean length. Additionally, to allow

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-18 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

Base.
Low
uz um −1
1 Med.
High

0.8

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

z zc −1
(a) Mj = 1.22

Base.
Low
Med.
uz um −1

1
High

0.8

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

z zc −1
(b) Mj = 1.36

Base.
Low
Med.
uz um −1

1
High

0.8

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

z zc −1
(c) Mj = 1.50

Base.
Low
Med.
uz um −1

1
High

0.8

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

z zc −1
(d) Mj = 1.56

FIG. 21. Centerline profiles of axial velocity normalized by the mean centerline velocity within the potential core. The axial
coordinate is normalized by the potential core length.

direct comparison of all configurations the second shock cell system which occurs at the higher
operating conditions was excluded from the mean shock cell length calculations and only the
structure emanating from the nozzle lip was used.
The results of this calculation are shown in Figure 24. The theoretical Prandtl-Pack42 vortex-
sheet model, shown below in Eq. (4),
Ls π 2
= M j − 1(μ = 2.405) (4)
Dj μ
was also included for comparison purposes. Additionally, the relation between fully expanded jet
Mach number and average shock spacing developed by Norum and Seiner43 for CD nozzles, shown

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-19 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

1.2 Profile
Lau
z
1 c
uz uj −1

0.8

0.6

0.4
0 2 4 6 8 10 12 14 16 18 20
z Dj −1

FIG. 22. Centerline quantity determination. Circles correspond to local velocity maximums. Diamonds correspond to local
velocity minimums. Baseline jet. Mj = 1.22.

below in Eq. (5), was compared to the measured values:

Ls b/2
= a Mj2 − 1 ; a = 1.1 and b = 1.17. (5)
Dj
The comparison with Norum and Seiner43 very closely follows the baseline data, with only a small
deviation in the underexpanded regime. Prandtl-Pack42 on the other hand, over predicts the shock
spacing by roughly 20%, a limitation which has been documented in the past.18, 31, 44
The results presented in Figure 24 indicate that vorticity introduction reduces shock cell spacing
at all operating conditions. Additionally, a trend of reducing shock cell spacing with increases in
vorticity level occurs across all observed jet Mach numbers, though in a non-uniform manner. Similar
results were shown by Tide and Srinivasan11 for sonic nozzles.
Finally, the shock strength was estimated using the centerline velocity profiles. In this work,
shock strength is defined as the difference between the upstream velocity maximum and down-
stream minimum velocity divided by the fully expanded jet velocity (u1 − u2 )uj −1 . An example
of the velocity maximums and minimums are indicated by the circles and diamonds in Figure 22,
respectively. The effect of vorticity magnitude on average shock strength for the available jet Mach
numbers is located in Figure 25. Additionally, a slightly modified shock strength characterization
parameter (Eq. (6)) originally developed by Tam45 was included for comparison purposes. The
modification to the original parameter is in the form of an additive constant used to account for
the non-ideal expansion of the flow field and the consequent shock structure at the nozzle’s design
condition. A good comparison between the baseline’s shock strength and the characterization pa-
rameter is achieved in the moderately imperfectly expanded range, while the highly overexpanded
condition is over predicted. A possible explanation for the over prediction is particle lag through

6
j
z /D
c

Base.
5 Low
Med.
High
4 Lau et al.

1.2 1.25 1.3 1.35 1.4 1.45 1.5 1.55 1.6

Mj

FIG. 23. Potential core length as a function of fully expanded Mach number. The empirical relation of Lau et al.39 is provided
for comparison.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-20 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

1.8 Base.

Shock Spacing [Ls Dj −1 ]

1.6 Low
Med.
1.4 High
Prandtl
1.2
Norum
1
0.8
0.6
0.4
1.2 1.25 1.3 1.35 1.4 1.45 1.5 1.55
Mj

FIG. 24. Evolution of shock cell spacing with fully expanded jet Mach number. The Prandtl-Pack relation42 and the relation
of Norum and Seiner43 are provided for comparison.

the Mach disks reduces the measured velocity gradient which consequently lowers the calculated
shock strength.46–48 Investigating the effect of vorticity on the calculated shock strength indicates
a reduction for all investigated cases. In the overexpanded regime, increases in vorticity magnitude
improve shock strength reductions, while in the ideally/under expanded regimes lower values are
optimal in terms of shock strength reduction. This switch in efficacy is attributed to the generation
of complex shock structures (Figure 16) as was discussed above:
⎧
⎪  2 2  2  2 3 −1
⎪
⎪ M jγ +1 −Md 2 D
+
M j −Md 2
+ c M j − Md ≤ 0,
⎨ 1+ 2 M j 2 Dj
1 3 γ +1
1+ 2 M j 2
2
A =  2 2  2 5 −1 (6)
⎪
⎪
⎪
⎩
M j −Md 2
γ +1 2 1+6
M j −Md 2
γ +1 2 +c M j − Md ≥ 0.
1+ 2 Mj 1+ 2 Mj

In addition to the above quantities, an investigation of spread rate characteristics was performed
using the streamwise velocity data. Each of the metrics used in Schadow et al.49 were computed
and it was determined that the jet half width, r1/2 , best illustrated the effect of vorticity magnitude
on the flow field. The jet half width for all operating conditions and configurations is located in
Figure 26. Due to the high number of axial data points, only every tenth point was included to ease
visualization. The axial location in the figure was nondimensionalized using the axial sonic point
location calculated using the theoretical relation of Zaman,50 shown below in Eq. (7):
z sonic
= 7 + 1.2M j 2 . (7)
Dj

0.6
Average Sho ck Strength

Base.
0.5 Low
Med.
0.4 High
Tam
0.3

0.2

0.1

0
1.2 1.25 1.3 1.35 1.4 1.45 1.5 1.55 1.6
Mj

FIG. 25. Average shock strength as a function of fully expanded jet Mach number. The shock strength characterization
parameter of Tam45 is provided for comparison.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-21 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

1.6 1.6
Base Base
1.4 Low 1.4 Low
Med. Med.
r1/2 Dj −1 1.2 High 1.2 High

r1/2 Dj −1
1 1

0.8 0.8

0.6 0.6

0.4 0.4
0 0.5 1 1.5 0 0.5 1 1.5
z zsonic −1 z zsonic −1

(a) Mj = 1.22 (b) Mj = 1.36

1.6 1.6
Base Base
1.4 Low 1.4 Low
Med. Med.
1.2 High 1.2 High
r1/2 Dj −1

1 r1/2 Dj −1 1

0.8 0.8

0.6 0.6

0.4 0.4
0 0.5 1 1.5 0 0.5 1 1.5
z zsonic −1 z zsonic −1

(c) Mj = 1.50 (d) Mj = 1.56

FIG. 26. Normalized jet half width development. Filled symbols: Vortex generator tip profile. Open symbols: Vortex generator
valley profile.

This nondimensionalization was chosen as it segregates the developing and asymptotic jet spread
regions to the left and right of unity, respectively. The results follow similar trends to the half width
presented by Schadow et al.,49 but due to the mismatch in jet Mach number the rates are not directly
comparable as rate decreases with Mach number.50 Excluding near the nozzle, the effect of vorticity
on the jet half width and consequently spread rate inside the developing region is limited. The near
nozzle effects are limited to z/zsonic < 0.25 or z/Dj < 3.0, where the expulsion of fluid is manifested
as an increase in half width and a slight decrease in half width results from the presence of the vortex
generator tip. Furthermore, operating conditions without a large change in potential core length, as
shown in Figure 23 (Mj = 1.22 and 1.56), possess asymptotic spread rates that are unchanged (within
9%) due to the effect of vorticity introduction. On the other hand, the introduction of any level of
vorticity roughly halves the spread rate for conditions with large potential core length changes.
Finally, integration of the fluctuating axial velocity over the streamwise cross sections was
completed in an effort to quantify overall mixing enhancement. The results of this effort are located
in Figure 27, where the integration has been normalized by the fully expanded jet velocity and
diameter. As was done above, only every tenth data point was included for clarity. The most
noticeable feature of the profiles is the large amplitudes of the baseline profiles at the Mj = 1.36
and 1.50 conditions, which is again attributed to the undulation of the jet column due to screech.
For nearly all configurations, the integrated turbulence is increased as compared to the baseline in
the near nozzle region. Additionally, the effect roughly increases with vorticity magnitude. In the
downstream region the effect is reversed, with the baseline possessing the largest amplitudes which
reduce with increased vorticity. Similar non-screeching results were presented by Bridges et al.12
for an overexpanded jet.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-22 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

FIG. 27. Integrated axial turbulence development normalized by the fully expanded jet diameter and velocity.

IV. CONCLUSIONS
An experimental investigation into the effects of induced streamwise vorticity magnitude on a
supersonic jet was presented to improve overall understanding of flow field modifications. To this
end, three vortex generator configurations were employed to achieve differing levels of streamwise
vorticity in the overexpanded, ideally expanded, and underexpanded regimes, though particular em-
phasis was placed on the overexpanded regime due to the limited number of previous investigations
within the literature. Both streamwise and cross stream PIV measurements were presented, which
allowed quantification of axial and radially aligned jet features such as shock spacing and streamwise
vorticity, respectively.
Investigation of the induced streamwise vorticity indicated that vortex generator penetration
was positively correlated with the radial and azimuthal size of the vortices as well as the maxi-
mum magnitude in the overexpanded regime. At the ideally expanded condition, vortex generator
penetration was only correlated with the radial extent of vortices. This was attributed to limited
azimuthal space for vortex growth leading to vortex interaction. Consequently, integration of the
streamwise vorticity modulus  was used as the metric for overall modification as it contains both
vorticity amplitude and spatial extent. Additionally, use of this metric includes the secondary vortex
structures that were measured in addition to the primary pair of counter rotating vortices. Analysis of
the downstream evolution of streamwise vorticity indicated that persistence was limited to roughly
2 diameters downstream of the nozzle exit, independent of magnitude. The decay of maximum
streamwise vorticity magnitude was found to be exponential, though the decay rate of  was not.
The potential core length was calculated through use of the centerline profile developed by
Lau et al.39 Comparison of the three levels of induced vorticity did not indicate a significant trend
regarding modification of the potential core length. Due to the presence of screech at the slightly

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-23 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

overexpanded and ideally expanded conditions, the baseline jet’s potential core length was found to
be drastically lower than the empirical relation Lau et al.39 It is thought that this phenomenon could
be the cause of conflicting reports within the literature regarding increased/decreased potential core
length caused by introduced streamwise vorticity.
The shock cell spacing was calculated using the centerline profiles and compared to theoretical
and empirical relations. Agreement between the relation of Norum and Seiner43 and the baseline
results was found to be excellent. Comparison to the vortex generator configurations indicated
reductions in spacing were directly related to the induced vorticity. Considering the results of André
et al.,22 the effect of screech on the calculated shock cell spacing was judged to be limited as the
computation method employed a maximum of the first seven shock cells.
Similarly, the average shock strength was determined using the centerline velocity profiles.
Comparison to the strength characterization parameter of Tam45 indicated reasonable agreement
with the baseline results, but highlighted the limitation of PIV to measure very large velocity
gradients as exist in the Mach disk region of the highest overexpanded condition. The effect of
vorticity on shock strength was found to be positively correlated in the overexpanded regime, and
negatively correlated in the underexpanded regime. This was attributed to introduction of secondary
shock structures by the presence of the larger vortices in the underexpanded regime.
The spread rate of all investigated configurations was illustrated through calculation of the jet
half width. The effect of vorticity on the half width was primarily limited to axial locations less
than three diameters downstream of the nozzle exit. In cases where large amplitude baseline screech
was not observed, the vortex generator jets achieve nearly identical asymptotic spread rates as the
baseline jet. The baseline asymptotic spread rates for conditions with large amplitude screech were
roughly halved by any induced vorticity.
Finally, the integration of axial turbulence within the measured streamwise planes indicated that
near nozzle mixing increased and downstream turbulence levels decreased with vorticity introduc-
tion.

ACKNOWLEDGMENTS
This work was sponsored by the Office of Naval Research (ONR) through the Jet Noise
Reduction (JNR) Project under the Noise Induced Hearing Loss (NIHL) program. N. Heeb would
like to thank the Ohio Space Grant Consortium (OSGC) for personal support during this work.
1 K. B. M. Q. Zaman, M. F. Reeder, and M. Samimy, “Control of an axisymmetric jet using vortex generators,” Phys. Fluids
6, 778–793 (1994).
2 M. Samimy, K. B. M. Q. Zaman, and M. F. Reeder, “Effect of tabs on the flow and noise field of an axisymmetric jet,”

AIAA J. 31, 609–619 (1993).

3 F. F. Grinstein, E. Gutmark, T. P. Parr, D. M. Hanson-Parr, and U. Obeysekare, “Streamwise and spanwise vortex interaction

in an axisymmetric jet: A computational and experimental study,” Phys. Fluids 8, 1515–1524 (1996).
4 A. K. M. F. Hussain and M. F. Zedan, “Effects of the initial condition on the axisymmetric free shear layer: Effect of the

initial fluctuation level,” Phys. Fluids 21, 1475–1481 (1978).

5 P. J. Morris, D. K. McLaughlin, and C. W. Kuo, “Noise reduction in supersonic jets by nozzle fluidic inserts,” J. Sound

Vib. 332, 3992–4003 (2013).

6 S. S. Pannu and N. H. Johannesen, “The structure of jets from notched nozzles,” J. Fluid Mech. 74, 515–528 (1976).
7 J. M. Seiner, L. S. Ukeiley, and B. J. Jansen, “Aero-performance efficient noise reduction for the f404-400 engine,” AIAA

Paper 2005-3048, 2005.

8 J. Hileman and M. Samimy, “Effects of vortex generating tabs on noise sources in an ideally expanded mach 1.3 jet,” Int.

J. Aeroacoust. 2, 35–63 (2003).

9 A. Opalski, M. P. Wernet, and J. Bridges, “Chevron nozzle performance characterization using stereoscopic DPIV,” AIAA

Paper 2005-444, 2005.

10 P. Behrouzi and J. J. McGuirk, “Effect of tab parameters on near-field jet plume development,” J. Propul. Power 22,

576–585 (2006).
11 P. Tide and K. Srinivasan, “Effect of chevron count and penetration on the acoustic characteristics of chevron nozzles,”

Appl. Acoust. 71, 201–220 (2010).

12 J. Bridges, M. P. Wernet, and F. C. Frate, “Piv measurements of chevrons on f400-series tactical aircraft nozzle model,”

AIAA Paper 2011-1157, 2011.

13 D. Munday, N. Heeb, E. Gutmark, J. Liu, and K. Kailasanath, “Acoustic effect of chevrons on supersonic jets exiting

conical convergent-divergent nozzles,” AIAA J. 50, 2336–2350 (2012).

14 C. Kuo, J. Veltin, and D. K. McLaughlin, “Advanced acoustic assessment of small-scale military-style nozzles with

chevrons,” AIAA Paper 2010-3923, 2010.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-24 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

15 M. O. Burak, L. E. Eriksson, D. Munday, E. Gutmark, and E. Prisell, “Experimental and numerical investigation of a
supersonic c-d chevron nozzle,” AIAA Paper 2009-4004, 2009.
16 M. Samimy, J. H. Kim, J. Kastner, I. Adamovich, and Y. Utkin, “Active control of high-speed and high-reynolds-number

jets using plasma actuators,” J. Fluid Mech. 578, 305–330 (2007).

17 B. Callender, E. Gutmark, and R. DiMicco, “The design and validation of a coaxial nozzle acoustic test facility,” AIAA

Paper 2002-0369, 2002.

18 D. Munday, E. Gutmark, J. Liu, and K. Kailasanath, “Flow structure and acoustics of supersonic jets from conical

convergent-divergent nozzles,” Phys. Fluids 23, 116102 (2011).

19 D. R. Glass, “Effects of acoustic feedback on the spread and decay of supersonic jets,” AIAA J. 6, 1890–1897

(1968).
20 J. Bridges and M. P. Wernet, “Turbulence associated with broadband shock noise in hot jets,” AIAA Paper 2008-2834,

2008.
21 A. J. Saddington, N. J. Lawson, and K. Knowles, “Numerical predictions and experiments on supersonic jet mixing from

castellated nozzles,” ICAS Paper 2002-362, 2002.

22 B. André, T. Castelain, and C. Bailly, “Broadband shock-associated noise in screeching and non-screeching underexpanded

supersonic jets,” AIAA J. 51, 665–673 (2013).

23 R. H. Schlinker, J. C. Simonich, and R. A. Reba, “Flight effects on supersonic jet noise from chevron nozzles,” AIAA

Paper 2011-2703, 2011.

24 AIAA, “Assessment of experimental uncertainty with application to wind tunnel testing,” in AIAA Standard S-071A-1999

(AIAA, Reston, VA, 1999).

25 AIAA, “Assessing experimental uncertainty: Supplement to AIAA S-071A-1999,” in AIAA Guide G-045-2003 (AIAA,

Reston, VA, 2003).

26 M. B. Alkislar, L. M. Lourenco, and A. Krothapalli, “Stereoscopic PIV measurements of a screeching supersonic jet,” J.

Visualization 3, 135–143 (2000).

27 N. G. Deen, P. Willems, M. van Sint Annaland, J. A. M. Kuipers, R. G. H. Lammertink, A. J. B. Kemperman, M. Wessling,

and W. G. J. van der Meer, “On image pre-processing for PIV of single- and two-phase flows over reflecting objects,” Exp.
Fluids 49, 525–530 (2010).
28 Z. R. Carr, K. A. Ahmend, and D. J. Forliti, “Spatially correlated precision error in digital particle image velocimetry

measurements of turbulent flows,” Exp. Fluids 47, 95–106 (2009).

29 E. Lazar, B. DeBlauw, N. Glumac, C. Dutton, and G. Elliott, “A practical approach to PIV uncertainty analysis,” AIAA

Paper 2010-4355, 2010.

30 J. T. Park, A. Derrandji-Aouat, B. Wu, S. Nishio, and E. Jacquin, “Uncertainty analysis: Particle imaging velocimetry,”

in ITTC Recommended Procedures and Guidelines (International Towing Tank Conference) (The Japan Society of Naval
Architects and Ocean Engineers, Fukuoka, Japan, 2008).
31 N. Heeb, E. Gutmark, J. Liu, and K. Kailasanath, “Fluidically enhanced chevrons for supersonic jet noise reduction,”

AIAA J. 52, 799–809 (2014).

32 B. André, T. Castelain, and C. Bailly, “Shock-tracking procedure for studying screech-induced oscillations,” AIAA J. 49,

1563–1566 (2011).
33 M. B. Alkislar, A. Krothapalli, and G. W. Butler, “The effect of streamwise vortices on the aeroacoustics of a Mach 0.9

jet,” J. Fluid Mech. 578, 139–169 (2007).

34 C. K. W. Tam and K. B. M. Q. Zaman, “Subsonic jet noise from nonaxisymmetric and tabbed nozzles,” AIAA J. 38,

592–599 (2000).
35 T. Irie, T. Yasunobu, H. Kashimura, and T. Setoguchi, “Characteristics of the mach disk in the underexpanded jet in which

the back pressure continuously changes with time,” J. Therm. Sci. 12, 132–137 (2003).
36 K. Gudmundsson and T. Colonius, “Spatial stability analysis of chevron jet profiles,” AIAA Paper 2007-3599,

2007.
37 B. Malla, D. Cuppoletti, J. Kastner, and E. Gutmark, “Proper orthogonal decomposition on a subsonic jet exhausted from

an axisymmetric nozzle and chevron nozzles with varying penetration,” AIAA Paper 2011-278, 2011.
38 D. Violato and F. Scarano, “Three-dimensional evolution of flow structures in transitional circular and chevron jets,” Phys.

Fluids 23, 124104 (2011).

39 J. C. Lau, P. J. Morris, and M. J. Fisher, “Measurements in subsonic and supersonic free jets using a laser velocimeter,” J.

Fluid Mech. 93, 1–27 (1979).

40 D. C. Kenzakowski and C. Kannepalli, “Jet simulation for noise prediction using advanced turbulence modeling,” AIAA

Paper 2005-3086, 2005.

41 J. M. Seiner, L. S. Ukeiley, B. J. Jansen, C. Kannepalli, and S. Dash, “Noise reduction technology for f/a-18 e/f aircraft,”

AIAA Paper 2004-2972, 2004.

42 D. C. Pack, “A note on Prandtl’s formula for the wavelength of a supersonic gas jet,” Q. J. Mech. Appl. Math. 3, 173–181

(1950).
43 T. D. Norum and J. M. Seiner, “Broadband shock noise from supersonic jets,” AIAA J. 20, 68–73 (1980).
44 C. K. W. Tam, J. M. Seiner, and J. C. Yu, “Proposed relationship between broadband shock associated noise and screech

tones,” J. Sound Vib. 110, 309–321 (1986).

45 C. K. W. Tam, “Broadband shock-associated noise of moderately imperfectly expanded supersonic jets,” J. Sound Vib.

140, 55–71 (1990).

46 A. Krothapalli, E. Rajkuperan, F. Alvi, and L. Lourenco, “Flow field and noise characteristics of a supersonic impinging

jet,” J. Fluid Mech. 392, 155–181 (1999).

47 B. Henderson, J. Bridges, and M. Wernet, “An experimental study of the oscillatory flow structure of tone-producing

supersonic impinging jets,” J. Fluid Mech. 542, 115–137 (2005).

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18
 086102-25 Heeb, Gutmark, and Kailasanath Phys. Fluids 26, 086102 (2014)

48 D. Mitchell and D. Honnery, “Particle relaxation and its influence on the particle image velocimetry cross-correlation
function,” Exp. Fluids 51, 933–947 (2011).
49 K. C. Schadow, E. Gutmark, and K. J. Wilson, “Compressible spreading rates of supersonic coaxial jets,” Exp. Fluids 10,

161–167 (1990).
50 K. B. M. Q. Zaman, “Asymptotic spreading rate of initially compressible jets-experiment and analysis,” Phys. Fluids 10,

2652–2660 (1998).

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.88.136.149 On: Mon, 01 Dec 2014 09:56:18