Coupling Modes in Supersonic Twin Rectangular Jets

Ata Esfahani*, Nathan Webb†, and Mo Samimy‡

Gas Dynamics and Turbulence Laboratory, Aerospace Research Center
The Ohio State University, Columbus, Ohio 43235

AIAA SciTech 2021 Meeting

January 11th-15th 2021

In the closely spaced twin-jet configuration often seen in military aircraft, the jets can
become coupled, resulting in significantly increased near-field pressure fluctuations and far-
field noise. The propulsion/airframe integration benefits of non-axisymmetric nozzles has led
to renewed interest in their integration into future generations of aircraft design. A thorough
understanding of the coupling dynamics and their effects on screech amplitude, intense near-
field unsteady pressure fluctuations and far-field noise levels and directivity is needed to
effectively implement active flow control strategies to disrupt the underlying mechanisms of
coupling. The results of experiments aimed at characterizing the coupling modes of closely
spaced, low aspect ratio, twin rectangular supersonic jets with a design Mach number 1.5 are
presented in this paper. The results of near-field experiments indicate that anti-symmetric
screech modes are present in both overexpanded and underexpanded flow regimes. The
results also indicate that the jets, in general, are coupled out-of-phase along the minor axis of
the jets in the overexpanded regime and in-phase in the underexpanded regime. Flow
visualization with a Z-type schlieren system showed that the jet spreading angle in the vertical
plane is higher when the jets are coupled out-of-phase. Visualization experiments also showed
that standing waves are present at Mach numbers where the screech amplitude and coherence
are increasing. More work is needed to understand the interplay between standing waves and
the strength of screech.

I. Nomenclature
AR = nozzle aspect ratio
BBSAN = broadband shock-associated noise
De = area-based equivalent nozzle diameter
f = screech frequency
Md = design Mach number
Mj = fully expanded jet Mach number
NPR = nozzle pressure ratio
s = nozzle center-to-center spacing in a twin-jet setup
SPL = sound pressure level in dB
St = screech Strouhal number, fDe/uj
θ = polar angle
ρ = near-field density
σ = standard deviation
φ = azimuthal angle

*
PhD Student, Department of Mechanical and Aerospace Engineering, AIAA student member
†
Research Scientist, Aerospace Research Center, AIAA Senior Member
‡
John B. Nordholt Professor of Mechanical and Aerospace Engineering, AIAA Fellow, Corresponding Author:
[email protected]

1
American Institude of Aeronautics and Astronautics
 II. Introduction
The propulsion/airframe integration benefits of non-axisymmetric nozzles has led to renewed interest in their
integration into future generations of aircraft design. The evolution of exhaust system design requirements in response
to a need for greater operational performance led to the rise of additional requirements such as minimizing drag, thrust
reversing, thrust vectoring in pitch, yaw and roll and improved flow mixing [1]. Integration of the aft body and nozzle
in an aircraft can severely affect drag and weight. Since the aft body of an aircraft is not typically axisymmetric,
integrating an axisymmetric nozzle with the airframe would introduce additional drag [2]. As a result, non-
axisymmetric nozzles are an attractive option for minimizing drag. Studies have shown [3,4] that the plume spreading
rate of a rectangular nozzle can be significantly higher than that of an equivalent axisymmetric nozzle. It is also easier
to implement a thrust vectoring system in a two-dimensional nozzle as the nozzle walls can simply be deflected to
provide thrust vectoring in pitch, yaw or even roll axis in addition to thrust reversing functions. Furthermore, a 4-flap,
two-dimensional nozzle system can be lighter than an equivalent axisymmetric design [1]. Aircraft with high- and
low-aspect-ratio nozzles have already been operational for years and recent interest in developing manned and
unmanned platforms with non-axisymmetric nozzles integrated with the airframe (e.g. F/A-XX, MQ-25 and X-47
programs, see Fig. 1) underscores the need for further development and characterization of such geometries. Another
issue that needs to be considered when designing future aircraft with twin rectangular nozzles that improve
propulsion/airframe integration is sonic/acoustic fatigue [5]. Due to widespread use of composites and special stealth
coatings in the latest generation of high-performance military aircraft, structural components in the vicinity of the
nozzles must be protected from the high levels of fluctuating pressures caused by coupling of the jets [6–8].

(a) (b) (c)

Fig. 1. Various Navy program demonstrators employing non-axisymmetric nozzle designs a) Boeing F/A-XX
rendering b) Northrop Grumman X-47B c) Boeing MQ-25

Despite the need for better understanding and characterizing the flow in twin rectangular geometries, it is surprising
to see that only a handful of studies have been directed at this topic over the past three decades. A list of studies
covering the topic of twin rectangular jets is provided in Table 1. The earliest work on the coupling between two low
aspect ratio rectangular jets was reported by Zilz and Wlezien [9]. The authors identified the four possible coupling
modes between low aspect ratio rectangular jets arranged side-by-side along their major axes: normal symmetric (in-
phase coupling in the vertical plane), normal asymmetric (out-of-phase coupling in the vertical plane), lateral
symmetric and lateral anti-symmetric (out-of-phase and in-phase coupling in the horizontal plane, respectively). They
also noted that symmetric coupling modes generate higher sound pressure levels (SPL) whereas anti-symmetric modes
led to a reduction in SPL in the plane of symmetry. Zilz and Wlezien reported that anti-symmetric modes are present
in closely spaced, underexpanded jets. Another notable finding of their study was the fact that higher spreading rates
in rectangular jets can lead to rapid merging of the plumes and result in screech suppression. Following their work,
Zeierman et al [10] studied the spreading rates of plumes in two adjacent rectangular nozzles with an aspect ratio of
3. They found that operating twin nozzles above their design Mach number does not alter the spreading rate compared
to a single rectangular nozzle. However, when operated at a highly overexpanded condition, there is a significant
increase in plume diffusion rate due to the addition of the second jet.
Raman and Taghavi [11] performed a comprehensive study of coupling between twin rectangular jets. The aspect
ratio in their work was 5, and the work itself is an extension of their earlier study of coupling between a linear array
of four jets [12]. In their work, while looking at mixing and noise benefits, Raman and Taghavi [8] synchronized the
four jets to achieve higher mass flux and plume spreading rates. In their follow-up study [11], Raman and Taghavi
confirmed the findings of Zilz and Wlezien [9], namely that anti-symmetric flapping of the jets minimizes the near-
field pressure fluctuation levels (at  = 90°) and symmetric flapping maximizes them.
Two more recent studies by Bozak [9] and Bozak and Wernet [14] were aimed at developing an empirical model
for prediction of noise in twin circular and rectangular nozzles. They also studied flow velocity in the inter-nozzle
region with PIV and noted that the addition of forward-flight effects led to the creation of a low velocity region

2
American Institude of Aeronautics and Astronautics
between the two nozzles which, in turn, led to higher turbulent kinetic energy levels for the jet and an increase in the
far-field sound pressure level (SPL) [14].

Table 1. List of studies on planar, twin rectangular jets (s: center-to-center nozzle spacing, h: nozzle exit
height, w: nozzle exit width, De: nozzle exit equivalent diameter)
Author Nozzles Design Nozzle Spacing Range of Mach
Numbers Investigated

Zilz and Wlezien, 1990 C-D, rectangular s/w = 1.55 – 3.2 Mj = 1.26 – 1.55
Md = 1.25, 1.45
AR = 1.15, 2.86
Walker, 1990 C-D, rectangular s/w = 2.50 – 3.7 Mj = 1.1 – 1.7
Md = 1.40
AR = 3.71
Zeierman et al., 1992 C-D, rectangular s/De = 1 – 4 Mj = 1.6 – 2.1
Md = 1.8
AR = 3
Raman and Taghavi, 1996 Converging nozzles, AR = 5 s/h = 4, 5.5, Mj = 1.2 – 1.75
7.5, 10, 11.5, 13.7

Raman and Taghavi, 1998 Converging nozzles, AR = 5 s/h = 5.5, 9, 15 Mj = 1.0 – 1.6

Bozak, 2014 Converging nozzles, AR = 2, s/De = 2.45, Mj = 0.9 – 1.33

8 3.32, 5.26

Almost all of the previous studies, with the works of Zilz and Wlezien [9] and [15] being an exception, have been
directed towards investigating high aspect ratio nozzles with significant inter-nozzle spacings, a combination which
may not be of interest for airframe/propulsion integration in the latest generation of tactical manned and unmanned
aircraft. While a significant body of work on the physics and control of twin circular jets exists in the literature (e.g.
recent work in our laboratory [16]) no attempt has been made, to the authors’ knowledge, to control and decouple twin
rectangular jets with active flow control techniques. The use of passive techniques, such as installing tabs in the
nozzles, to decouple twin rectangular jets has resulted in very limited success as reported by Walker [11]. A thorough
understanding of coupling modes between the jets is needed in order to implement active flow control strategies that
would disrupt the underlying mechanisms of coupling. The works published recently on low aspect ratio, closely-
spaced twin rectangular jets [17–19] have not investigated coupling between the jets across the wide range of NPRs
likely present in application.
The work presented here fills this gap in the literature by investigating a pair of closely spaced, low aspect ratio
rectangular nozzles at multiple NPRs. The objective of this work is to characterize coupling between the jets and the
near-field dynamics and far-field noise at various operating regimes (from highly overexpanded to underexpanded)
and establish a baseline for future experiments aimed at controlling the jets using Localized Arc Filament Plasma
Actuators (LAFPAs). LAFPAs have been developed at the Gas Dynamics and Turbulence Laboratory (GDTL) [20,21]
and have been implemented in various jets [18], including twin supersonic circular jets [12]. These actuators exert
control authority by producing periodic thermal perturbations using an arc discharge. The perturbations excite the
Kelvin-Helmholtz instability in shear layers of the jet, and the amplified perturbations eventually roll up into flow
structures. By modifying the frequency and relative timing of the LAFPAs perturbations, the spatial organization and
growth of flow structures can be controlled [22]. Large-scale structures are known to play a critical role in mixing and
acoustics in jet flows [14]. Thus, the LAFPAs can, using only perturbation-frequency inputs, exert significant control
authority over jets and other flows with shear layers[23–27]. More details on the mechanism through which LAFPAs
excite the flow can be found in a recent review article by Samimy et al. [28]. An understanding of dynamics and the
interplay between various phenomena such as screech, coupling, and standing waves is crucial to leveraging the

3
American Institude of Aeronautics and Astronautics
naturally occurring instabilities that can generate the desired flow control effects. This paper details the facility design
and characterization, along with physics of coupling in the near-field and its effects on far-field SPL and directivity.

III. Experimental Methodology

A. Nozzle Design
A new nozzle assembly was designed and built at the GDTL to study flow physics and aeroacoustics of low aspect
ratio, twin supersonic rectangular jets. The assembly uses a pair of sharp throated, military style rectangular
converging-diverging nozzles with a design Mach number of 1.5 (Md = 1.5) and aspect ratio of 2 (AR = 2). From the
outset, the assembly was designed to be modular to accommodate possible changes in nozzle spacing, design Mach
number and actuator configuration by replacing specific parts with alternate versions.
In its current configuration, the nozzle exit width (w), height (h) and aspect ratio (AR) are 24.13 mm, 12.06 mm
and 2, respectively. The center-to-center spacing between the nozzles is 43.38 mm (2.25De). The area-based equivalent
diameter is De = 19.25 mm. A cross section of the nozzle assembly can be seen in Fig. 2a.
Boron nitride housings for LAFPAs (“actuator blocks”) are used to house the plasma actuators and prevent arcing
to metal parts. In order stabilize the arc in the upcoming flow control experiments and allow its frequency to be
precisely controlled, 1 mm wide by 0.5 mm deep grooves were cut into the actuator block, approximately 1 mm
upstream of the nozzle exit plane. The 1 mm diameter tungsten electrode tips are contained within the groove and held
flush with the jet inner surface. Previous work [29] has shown that the existence of the groove does not affect the
control authority of the actuators. The actuators and the grooves are present in the baseline geometry examined in this
paper to ensure that the upcoming excited flow results are compared to baseline results from an identical geometry.
Figure 2b shows a cutaway of the actuator block depicting the electrode bores and setscrew holes that are present
in the baseline cases reported here. The nozzle assembly (Fig. 3a) is attached to a high-pressure air pipe at the center
of a co-flow duct (Fig. 3b), used for simulating forward flight effects, via a circular mounting plate. A custom designed
metallic O-ring between the interface of the nozzle assembly and mounting plate ensures a tight seal exists between
these parts even during heated flow experiments.
Following the installation of the new nozzle assembly in the anechoic chamber at the GDTL a Mach number sweep
in increments of 0.05 was conducted between Mj = 1.4 and Mj = 1.6 in order to determine the ideally expanded Mach
number of the new facility. The far-field array was used to collect the data at an azimuthal angle of 90° ( = 90° see
Fig. 3a) and polar angles of 30°, 90° and 120°. The shape of the spectra, specifically the amplitude of the screech tones
and broadband shock associated noise (BBSAN) peaks, were analyzed to infer the ideally expanded flow case. The
presence of a shock system, originating from the nozzle throat, in addition to a shock cells present in non-ideally
expanded cases complicates the task of interpreting the spectra and establishing the ideally expanded case.
Nevertheless, based on the trends seen in the spectra (not shown here), we determined that the ideally expanded Mach
number is very near the design Mach number of Md = Mj = 1.50.

(a) (b)
Fig. 2. a) Nozzle cross-section, b) actuator block cutaway

4
American Institude of Aeronautics and Astronautics
 (a) (b)
Fig. 3. a) Twin-jet assembly showing minor axis, y, major axis, z, and azimuthal angle, ; b) twin-jet nozzle assembly
installed inside the co-flow channel

B. Facility and Instrumentation

All the experiments reported in this work have been conducted in the jet anechoic facility at the GDTL within the
Aerospace Research Center at the Ohio State University. The facility is a 6.2 m × 5.6 m × 3.4 m anechoic chamber
with a cut-off frequency of 160 Hz. High-pressure air is supplied by three five-stage reciprocating compressors that
store the air in two cylindrical tanks with a total storage capacity of 43 m3 at a maximum pressure of 16 MPa. The air
is then dried, filtered and passed through a series of screens in the settling chamber before being fed to the jet plenum.
A computer-controlled valve maintains the plenum pressure at a setpoint to maintain a constant NPR The nozzle
assembly is attached to the plenum using a circular mounting plate to allow the azimuthal angle of the nozzle assembly
() with respect to the far-field microphone array to be adjusted in 15° increments. The facility is capable of heating
the air in the stagnation chamber up to a total temperature ratio of 2.5. Forward flight conditions can be simulated
using a co-flow duct that can provide a stream of air at up to 100 m/s. The co-flow duct and the twin-jet nozzle
assembly were covered with acoustic foam (as seen in Fig. 4) during the experiments to prevent acoustic reflections.
A series of experiments at various nozzle pressure ratios (NPRs), ranging from moderately overexpanded (Mj =
1.2) to moderately underexpanded (Mj = 1.85) flow regimes, was conducted. This paper focuses on three jet operating
conditions that represent different jet behaviors observed in the course of our detailed investigation of low aspect ratio
twin rectangular jets. Near- and far-field acoustic data and schlieren images were acquired for NPR of 2.97, 3.67, and
5.0, corresponding to the following fully expanded jet Mach numbers (Mj): 1.35, 1.50, and 1.70, all at a jet total
temperature ratio of 1. These three NPRs correspond to an overexpanded case, an ideally expanded case and an
underexpanded case.
1
A near-field array, consisting of a 38 cm × 38 cm rectangular frame and six Brüel and Kjær 4939 in. microphones,
4
was used to collect data to study the coupling between the jets at various regimes. By computing the coherence and
phase difference between the pressure traces from individual microphones in the array, it is possible to assess the
coupling and screech modes of the jets at different operating Mach numbers. The number of microphones was chosen
so that it would be possible to infer both the screech mode of each jet [30,31] and the coherence and phase of the jets
with respect to each other, as reported by Raman and Taghavi [11]. All of the near-field microphones were positioned
at an axial location of x/De = 0 and radially (measured from the nearest jet centerline) at r/De = 4. These axial and
radial distances are sufficiently far from the jet boundary to be free of the hydrodynamic signature of the jet and record
only the acoustic signature of screech. The array and microphone mounts were wrapped in acoustic foam to prevent
reflections, as seen in Fig. 4. The array microphones were numbered in a counterclockwise (looking upstream) fashion
starting with the right horizontal microphone.

5
American Institude of Aeronautics and Astronautics
 1
A schematic of the far-field microphone array is shown in Fig. 5. Eight Brüel and Kjær 4939 in. microphones
4
mounted at polar angles (), ranging from 30° to 120°, measured from the downstream jet axis, are aimed at the
midpoint between the centerline of the twin-jet nozzles and are used to collect far-field data. The individual
microphone distances from the jet centerline range from 1.90 m to 3.78 m (98.9De to 196.5De). The results were
propagated to an observer distance of 100De from the point centered between the two jets along the major axis.
A standard Z-type schlieren system was used to visualize the flow-field along the major and minor axes of the jets.
To acquire high-resolution, time-averaged maps of density gradient at a wide range of Mach numbers, a 5.5 MP
LaVision Imager sCMOS camera was used. The knife edge was oriented vertically in order to highlight the horizontal
density gradients and 300 images with a window size of 2500×2150 pixels were acquired at 50 frame per second for
each test condition. An HPLS-36 high-powered pulsed LED from Lightspeed Technologies was used as the light
source. The LED pulse width was set to 500 ns which was short enough to freeze the flow features even for the highest
observed screech frequency. The Imager sCMOS camera was swapped with a Phantom v1210 camera for high-speed
schlieren imaging experiments. The goal of these experiments was to capture the motion of the screeching jets at
various Mach numbers. The frame acquisition rate was fixed at 60,000 frames per second for all the tested Mach
numbers and a set of 1000 frames with a window size of 512×340 pixels was recorded for each case. A Nikon zoom
lens was used to set the field of view to include only the flow features of interest, namely the shock system while
reducing the window size to allow an increase in maximum frame rate.

Fig. 4. Photograph of near-field microphone array covered with foam to prevent acoustic reflections

Fig. 5. Schematic of far-field microphone array and polar angle, 

6
American Institude of Aeronautics and Astronautics
C. Data Acquisition and Processing
Near- and far-field acoustic signals collected by microphones were amplified and band-pass filtered between 20
Hz and 100 kHz, with Nexus 2690 signal conditioners. Prior to acquiring a set of near-field or far-field data, an
experiment was conducted to ascertain the appropriate level of gain. Microphone calibrations were then performed at
the set gain values using a Brüel and Kjær model 4231 acoustic calibrator. The sampling frequency was 200 kHz. For
near-field acoustic measurements, 100 blocks of 32,768 samples were acquired for each data point while for far-field
measurements 400 blocks of 8,192 samples were recorded for each test case. These acquisition parameters result in a
24.4 Hz frequency resolution for the far-field results. The calculated Power Spectral Density plots are expressed as
sound pressure level in decibels referenced to 20 Pa. As was mentioned earlier, the far-field results were propagated
to an observer distance of 100De assuming a spherical propagation.
The jet screech and coupling modes were determined from the relative phase and coherence of the signals acquired
from various near-field microphones. In order to extract this information wavelet (Morlet) coherence and phase
between various pairs of microphone signals were calculated. The temporal averages of the coherence and phase are
performed using only time instants with a coherence greater than 0.7. This allows the average phase of the signals to
be examined only when the coherence values are high (and therefore the phases are reliable).
Maps of average density gradients and standard deviation of density gradients were calculated from the time-
averaged schlieren images in DaVis 8.2 and then imported in MATLAB where Log 10 of standard deviation of density
gradients was computed and plotted to visualize the standing wave patterns at various Mach numbers.

IV. Results and Discussion

The purpose of this paper is to present the results of experiments aimed at characterizing the baseline of a low
aspect ratio twin rectangular supersonic jets. Understanding the underlying physics of coupling in the baseline case
will enable us to implement effective flow control strategies.

A. Acoustic Measurements

1. Near-field Results
A detailed, fully-expanded jet Mach number sweep was conducted to obtain screech frequency in both
overexpanded and underexpanded jet operating regimes. Various scaling parameters for the screech frequency were
explored and it was found that the normalized screech frequencies using the geometrical nozzle exit equivalent
diameter and fully expanded jet velocity (St = fDe/uj) collapse the data from our experiments as well as data from other
researchers using rectangular jets with small (AR = 2) to moderate (AR = 5) aspect ratios. Figure 6 shows a comparison
between the screech tones reported by Shih et al. [34], Panda et al. [35], Raman [36], Alkislar [3], Karnam et al. [32]
and our results. Our data points presented in Fig. 6 have been acquired using the near-field microphone array and
compared with spectra from data gathered simultaneously with far-field microphones located at θ = 30°, 90° and 120°
to ensure that screech tone frequencies are accurate. Surprisingly, the data collapse is good given that the results
include both jets issuing from converging and converging-diverging nozzles and both single and twin-jet
arrangements. Of greatest importance is the twin-jet setup investigated by Karnam et al. [32] and numerically
simulated by Jeun et al. [33] and Viswanath et al [19] as these geometries are very similar to that investigated in this
paper. Additionally, a survey of the literature was conducted to gather screech data from single and twin rectangular
jets with various aspect ratios for comparison with the present results. Shih et al. [34] studied the variation of screech
Strouhal number, when normalized using nozzle exit height (h), against Mj from the results of several authors and
concluded that the variation of normalized screech tones against Mj is independent of nozzle aspect ratio. We
investigated this finding using two scaling parameters, namely h and De, to collapse the screech data. The scaling
parameter De was found to result in the best data collapse for low to moderate nozzle aspect ratios. However, the data
from high aspect ratio nozzles (e.g. AR = 10, 16.7 [4], AR = 9.63[31]) do not collapse either with each other or with
the results from low to moderate aspect ratio nozzles, i.e. AR = 2 – 5, regardless of whether h or De are used for
normalization. To avoid cluttering the graph, results from the higher aspect ratio nozzles are omitted from the figure
below.
Although the general collapse of the data from nozzles with a relatively wide range of aspect ratios and internal
geometries (converging or converging-diverging) and single or twin jets is good, it is instructive to observe that the
screech data from converging-diverging nozzles show more scatter in the overexpanded regime but demonstrate better
collapse at underexpanded Mach numbers. We have also observed more day-to-day variations in screech peak

7
American Institude of Aeronautics and Astronautics
frequency in our own results at overexpanded Mach numbers. More research is needed to examine this issue. However,
it should be noted that in the overexpanded flow regime, the nozzle exit pressure is lower than ambient pressure and
an oblique shock wave is attached to the nozzle exit. Depending upon the strength of the pressure gradient at the
nozzle exit, the boundary layer inside the nozzle could separate and introduce unsteadiness to the flow, shock system,
and screech process. Furthermore, when the high AR cases (not included in this figure) are considered, a trend of
higher St number for a given Mj with AR is observed, which shows that the relevancy of De as a length scale is reduced
as AR is increased.

Fig. 6 Variation of screech Strouhal number vs. Mach number

As mentioned earlier, a thorough understanding of the coupling dynamics and mode at different Mach numbers is
critical in implementing active flow control in order to reduce noise levels or decouple the jets. An extensive
experimental campaign, aimed at characterizing the jets’ coupling modes at a wide range of Mach numbers was
conducted. All the results presented in this section have been acquired using the near-field microphone array described
in section III.B. Time-averaged wavelet coherence and phase between near-field acoustic results are presented in Fig.
7. In order to infer the coupling mode of the jets, phase difference and coherence between signals from microphones
2 and 3 (see Fig. 4) were calculated and are presented in Fig. 7a. The jets start screeching for the first time at Mj =
1.20. The coherence between the jets increases progressively as the Mach number increases to the design Mach
number, except for a dip at Mj = 1.25. A screech frequency mismatch between jets is observed only at lower
overexpanded Mach numbers (i.e. Mj = 1.20 – 1.35). While initially the difference in screech frequencies between the
jets is approximately 140 Hz at Mj = 1.20, it decreases to 33 Hz (its minimum value) as Mj increases to 1.35. This
frequency mismatch causes the relative phase between the jets drifts in time which complicates the task of calculating
a time-averaged value for phase difference. This phenomenon is further discussed below. As Mach number increases
from Mj = 1.40 to 1.50, the frequency mismatch disappears and the coherence between the jets maintains its high
value, leading to a steady phase difference between the jets at approximately ±π. This indicates out-of-phase coupling.
The general upward trend in coherence in the overexpanded regime with increasing Mach numbers is, in large part,
caused by the screech intermittency at the low Mach numbers (Mj = 1.20 to 1.35), which results in alternately high
and low coherence levels. The higher Mach numbers, as already mentioned, produce consistent screech and long
periods of sustained high coherence. As a result, the time averaged coherence levels between Mj = 1.20 to 1.30 are
lower than those of Mj = 1.40 to 1.50 (see Fig. 7a). After reaching the design Mach number, further increasing the
Mach number causes an immediate drop in time-averaged coherence value at Mj = 1.55 (Fig. 7a), followed by screech
strength augmentation and increases in coherence between the jets. In this regime, the jets are strongly coupled in-
phase as indicated by the 0° phase difference between the jets (see Fig. 7a).

8
American Institude of Aeronautics and Astronautics
 It is worth mentioning that while Zilz and Wlezien [9] and Walker [15] have reported the existence of lateral
flapping in the horizontal plane (for low aspect ratio nozzles), we did not observe any evidence of coupling between
the jets in the horizontal plane (major axis). These observations were reinforced by high-speed schlieren imaging
experiments (not reported in this paper) which did not show any evidence of lateral flapping or coupling between the
jets in the horizontal plane.
It is instructive to examine the coherence and phase difference between signals from microphones located above
and below each jet (Fig. 7b). While examining coherence and phase difference between microphones 2 and 3 reveals
the nature of coupling between the jets, the relative strength of coherence and phase difference between microphones
2 and 6, and 3 and 5 indicates the nature (i.e. screech is anti-symmetric or symmetric) as well as coherence of the
screech of each jet. As we can see in Fig. 7b, the general trends in coupling strength are reflected in the coherence of
the screech of each jet. High coherence for each individual jet at Mj = 1.35 – 1.45 and 1.70 – 1.85 is accompanied by
strong coupling between the jets. It is also interesting to see that the coupling strength between the jets is almost never
significantly higher than the coherence level of either of the jets. At Mach numbers where a frequency mismatch exists
between the jets, i.e. Mj = 1.20 – 1.35, significant differences between the coherence level of the jets exist in some
cases. Such significant differences are not observed at any Mach number in the underexpanded regime.
Having examined the coupling behavior of the jets for a wide range of Mach numbers, we can now move on to
study the screech mode of each individual jet. Gutmark et al. [30], Shih et al. [34] and Raman and Rice [31] have
noted that single rectangular jets can display either a flapping motion, also known as an anti-symmetric mode, or a
pulsating movement in the axial direction which is commonly referred to as a symmetric mode. A plot of coherence
and phase difference between microphones 2 and 6 positioned above and below the right jet is presented in Fig. 7c. In
contrast to the coupling between the jets where values of time-averaged phase difference between microphones 2 and
3 were somewhat ambiguous at low overexpanded Mach numbers due to phase drift, it is interesting to see that for
nearly all the Mach numbers tested, the time-averaged phase differences are close to ±π (see Fig. 7c). Even at Mach
numbers where coherence in the jet on the right is relatively low, e.g. Mj = 1.20, 1.30, 1.50 and 1.65, the phase
difference is close to ±π. This indicates a strong flapping motion that is present throughout the test envelope. The
results from Fig. 7a and c indicate that in general, the jets tend to display a flapping motion (anti-symmetric screech
mode) and in overexpanded regime are largely coupled out-of-phase and in underexpanded regime are generally
coupled in-phase. While some authors have observed both anti-symmetric and symmetric screech modes in their single
jet experiments [30,31,34], our results are more in line with Raman and Taghavi’s observations in a high aspect ratio
twin-jet setup where they only observed an anti-symmetric screech mode for each jet [11].
The trends in near-field screech amplitude closely follow those of coherence between microphones 2 and 6 as
shown in Fig. 7d. The amplitudes of screech peaks from time-averaged spectra recorded by microphone 2 are plotted
over time-averaged coherence values in Fig. 7d. Raman and Taghavi [11] reported a 20 dB increase in the near-field
screech amplitude as their high aspect ratio twin-jets switched from out-of-phase flapping to in-phase flapping. They
also reported a significant reduction in screech amplitude in out-of-phase flapping mode compared to a single jet
screech amplitude and observed a modest tone amplification for in-phase mode cases compared to their single-jet
measurements. Zilz and Wlezien [9] also noted a 15 dB increase in screech tone amplitude as the their twin jets
switched from out-of-phase coupling to in-phase coupling while Walker [15] observed insignificant differences in
screech amplitudes as jets switched modes. While we do not have any single jet data available at this time, our
experiments indicate that when the coupling mode between the jets is switched from out-of-phase to in-phase, the
near-field screech peak amplitude increases by about 5 dB for cases where coherence is high. This might suggest a
modest degree of screech tone amplification in the in-phase coupling mode. Future single jet experiments will shed
light on some of these issues.

9
American Institude of Aeronautics and Astronautics
 (a) (b)

(c) (d)

Fig. 7. Twin-jet coupling modes and single jet screech modes observed in the twin rectangular jet setup
along the minor axis for a wide range of Mach numbers

To better illustrate the differences between various coupling modes we have observed in the jets, we have selected
three cases that are representative of the unique characteristics of each mode for further discussion. Mj = 1.35
represents the coupling characteristics observed in low overexpanded Mach numbers, namely intermittent screech,
frequency mismatch and the resultant phase drift. A spectrogram of the signal from microphone 2 is shown in Fig. 8a.
The screech peak occurs at St = 0.40 and, as can be seen in the plot, is highly intermittent. This case demonstrates the
highest screech peak amplitude in the overexpanded regime (see Fig. 7d). Our results indicate that generally, very
high coherence values are accompanied by a steady and consistent phase difference in time. That does not appear to
be the case for Mj = 1.35 where, despite high time-averaged coherence values (Fig. 8b), the large error bars on phase
values indicate that there are considerable variations in phase in the 0.16s time window of measurements. While the
coherence values appear to be consistently high in the measurement window (Fig. 8c), the time-resolved phase plot

10
American Institude of Aeronautics and Astronautics
(Fig. 8d) indicates that the phase difference between the signals is continuously varying from 0° to -π to +π and back
to 0°. The period between instances of time when phase is 0° is approximately 0.030 s which corresponds to the
frequency difference between the jets: 33 Hz. In fact, the frequency mismatch between the jets is the likely reason for
the presence of intermittent screech peaks seen in Fig. 8a. This phenomenon is further explained in Fig. 9. Each
microphone registers not only the acoustic signature of nearest jet, but that of the other jet as well. Thus, the resultant
waveform registered by microphone 2 is the superposition of the jets’ acoustic signatures (Fig. 9a). As there is a slight
(33 Hz) mismatch in screech frequency, the interference between these waveforms generates the beat pattern observed
in the spectrogram of microphone 2 signal (see Fig. 9b, same data presented in Fig. 8a). The instances of time where
the amplitude of the beat pattern is maximum correspond to a 0° phase difference between the jets and when the
difference is ±π, the amplitude of the beat pattern is at its minimum. Raman and Taghavi [11] reported that when two
adjacent jets are run simultaneously, they adopt a frequency that is slightly different from the frequency of each jet if
it is run individually. They stressed that when jets are coupled, they would always adopt the same frequency. This
does not appear to be the case at Mj = 1.35 where despite high coherence (see Fig. 8c), each jet still maintains its own
screech frequency. Examining Fig. 7b indicates that each jet has high coherence level which perhaps cannot be
overridden by the coupling process.

(a) (b)

(c) (d)

Fig. 8 Mj = 1.35 near-field results from microphones 2 and 3 a) spectrogram of microphone 2 signal, b)
time-averaged phase and coherence, c) time-resolved coherence, d) time-resolved phase

The coupling characteristics displayed by Mj = 1.50 are shared by overexpanded Mach numbers that are close to
the design Mach number. These characteristics include high coherence values between the jets, low but relatively
steady screech peaks and phase difference and the absence of frequency mismatch. As can be seen in Fig. 10a, the
relative amplitude of the screech peaks are much lower than those of Mj = 1.35 (see Fig. 8a) as expected in a case
where Mach number is the design Mach number, but the intermittency of the peaks is not as pronounced. This is
because the previously seen frequency mismatch does not exist at this Mach number.

11
American Institude of Aeronautics and Astronautics
 (a)

(b)

(c)

Fig. 9 a) Interference between acoustic waveforms from the jets, b) spectrogram of microphone 2 signal, c)
phase difference between microphone 2 and 3

(a) (b)

(c) (d)

Fig. 10. Mj = 1.50 near-field results from microphones 2 and 3 a) spectrogram of microphone 2 signal, b)
time-averaged phase and coherence, c) time-resolved coherence d) time-resolved phase

12
American Institude of Aeronautics and Astronautics
 Fig. 10b shows a strong coherence peak between the signals from microphones 2 and 3 around the screech Strouhal
number indicating a consistent and strong coupling between the jets in the minor-axis plane (also see Fig. 10c). The
smaller error bars on time-averaged phase at Mj = 1.50 compared to Mj = 1.35 indicate that variations in phase are
minimal in time. This is confirmed by the time-resolved phase plot seen in Fig. 10d, which indicates the coupling
phase is consistently ±π.This is in part due to the matched frequency of the jets.
Compared to Mj = 1.35 and 1.50, at Mj = 1.70 the screech peaks at St = 0.2 have a higher amplitude and are
consistent in time (see Fig. 11a). This Mach number is representative of underexpanded cases with high coherence
between the jets (Fig. 11b) indicating a strong, in-phase coupling in the minor-axis plane. As expected, high coherence
values are accompanied by a time-consistent phase difference which is evident from the small error bars of time-
averaged phase in Fig. 11b and time-resolved phase plot of Fig. 11d. Note that the amplitude of screech peaks at Mj =
1.70 are not significantly higher than that of those at Mj = 1.35 (Fig. 11a vs. Fig. 8a) but the peaks are more consistent.
That appears to be the case for all underexpanded Mach numbers between 1.70 and 1.85. Zilz and Wlezien [9], Walker
[15] and Raman and Taghavi [11,12] have noted that out-of-phase coupling between the jets results in higher screech
peak amplitudes in the near-field compared to in-phase coupling. Our current results indicate that while individual
screech peaks at Mj = 1.35 (a case which displays intermittent coupling) are comparable to those observed at Mj =
1.70, the consistency of screech peaks in time is higher in cases where jets are coupled in-phase.

(a) (b)

(c) (d)

Fig. 11 Mj = 1.70 near-field results from microphones 2 and 3: a) spectrogram of microphone 2 signal, b)
time-averaged phase and coherence, c) time-resolved coherence, d) time-resolved phase

2. Far-field Results
Far-field data sets were gathered to characterize the jets’ baseline acoustic signature at various nozzle pressure
ratios along the major and minor axes (azimuthal angles, φ = 0 and 90 ). The three cases representing the various
coupling characteristics of the jets, discussed in the previous section, will be examined here so that the effect of near-
field coupling dynamics on far-field noise level and directivity can be ascertained. Maps of far-field spectra for a range

13
American Institude of Aeronautics and Astronautics
of polar angles from θ = 30° – 120°, two different azimuthal angles and three Mach numbers are presented in Fig. 12.
Note that θ is measured from the downstream jet axis as shown in Fig. 5.

(a) (b)
(a)

(c) (d)
(a) (b
(a)

)
)

(e) (f)
(b (b
(a) (a)

Fig. 12 Maps of far-field SPL for a range of polar angles indicating noise levels and directivity a,b) Mj =
) ) e,f) Mj = 1.70
1.35, c,d) Mj = 1.50,

14
American Institude of Aeronautics and Astronautics
 Fig. 12a and b show the far-field spectra for Mj = 1.35 along the major and minor axes. Starting with the directivity
of the turbulent mixing noise component of the spectra, it appears that the directivity of this component of noise is not
strongly dependent on azimuthal angle suggesting that shielding does not play a major role for this component of the
noise at this Mach number. This is in line with reports of far-field noise directivity from a geometrically similar setup
investigated by Karnam et al. [32]. The mixing noise lobe, however, appears to have a slightly higher peak amplitude
along the major axis (φ = 0°). Although the signature of screech tone (f) is registered strongly at all polar angles along
the major axis, it is the first (2f) and to some degree second harmonics (3f) of the screech that have strong signatures
along the minor axis. We can also observe that the 1st harmonic radiates predominantly to the sideline as reported in
the literature [37]. At Mj = 1.35, which represents an out-of-phase coupling mode, for every complete flapping cycle
(observed by far-field microphones between  = 75° and 120° for  = 90°) the far-field microphones register two
flapping events as the jets are not coupled in-phase. Thus, far-field microphones register a strong peak at the first
harmonic (2f) of the screech fundamental tone along the minor axis (Fig. 12b). This assertion is supported by the fact
that the screech mode of the jets at this Mach number is a strong anti-symmetric mode (Fig. 7c). A relatively minor
BBSAN peak is only present between  = 100° and 120° for  = 90°. This feature becomes more dominant as Mach
number is increased and is registered along both azimuthal angles at higher Mach numbers. In conclusion, the fact
that the amplitudes of screech harmonic tones and peaks of BBSAN are higher along the minor axis compared to those
along the major axis at Mj = 1.35 reflects the minor axis flapping observed in near-field measurements.
As the Mach number is increased to Mj = 1.50, the mixing noise lobe grows in size and amplitude compared to Mj
= 1.35. While the spreading rate of the jets is higher at lower Mach numbers, as will be discussed in section IV.B, it
appears mixing noise levels become higher as Mach number is increased suggesting that the spreading rate of the jets
and mixing noise might not be directly related. The high amplitude area of the lobe along the minor axis is larger and
more intense (see Fig. 12d) suggesting some effects of acoustic shielding along the major axis. As before, the
amplitude of the first and second harmonics of screech are highest along the minor axis between  = 80° and 120°
reflecting the flapping motion of the jet due to screech. One can also notice a relatively stronger BBSAN lobe at higher
frequency than the primary lobe between  = 90° and 120° along the minor axis which is absent from the major axis
spectra (Fig. 12c and d). This suggests that jet shielding may also hinders the propagation of shock associated noise
components. It is also interesting to note the presence of a secondary BBSAN lobe around the second screech harmonic
between  = 105° and 120° (Fig. 12d).
At Mj = 1.70, the screech tone shows a strong downstream directivity and its well-defined peak stands out against
the prominent mixing noise lobe for both azimuthal angles (Fig. 12e and f). The mixing noise lobe at  = 90° also
shows a wide high-amplitude peak between St = 0.1 and 0.3 (Fig. 12f) whereas for  = 0°, the high-amplitude lobe is
somewhat smaller suggesting a possible shielding effect in propagation of mixing noise component at this Mach
number. It is interesting to note that at Mj = 1.70, in contrast to the previous cases, the signature of the screech primary
tone and its harmonic are stronger along the major axis. This is rather unexpected as the jets are strongly coupled in-
phase at this Mach number and based on the existing reports in the literature [9,11,15] one might expect to observe
higher noise levels along the minor axis normal to the twin-jet plane. Another observation from the results acquired
along the major axis is that the high-amplitude broadband peaks between St = 0.3 and 0.5 stretch from  = 90° to θ =
120° whereas along the minor axis, the broadband lobe is only dominant between  = 105° and 120°. In short, the
trends in far-field noise directivity indicate that at Mj = 1.35 and 1.50, mixing noise appears to be the dominant noise
component and the amplitude of the screech tones is highest along the minor axis. At Mj = 1.70, however, in an
unexpected turn, the amplitude of the screech tone and its first harmonic become higher along the major axis and
BBSAN becomes a significant source of noise on par with turbulent mixing noise.
As active flow control strategies for addressing different components of far-field noise are dissimilar, the insights
gained from examining the maps of far-field noise are invaluable and would assist us in the process of choosing the
excitation parameters for implementing active control.

B. Optical Diagnostics

1. Time-averaged Schlieren
In order to further complement the insights gained from near- and far-field acoustic measurements, an extensive
schlieren imaging campaign was conducted. High-resolution images of the jets along the minor and major axes were
acquired for Mach numbers between 1.20 and 1.90 in increments of 0.05. In this section, a limited set of results for
the Mach numbers that have been selected for discussion are presented. In the course of our investigation of twin-jet
flow-field, we noticed the presence of standing waves at some specific Mach numbers. In order to further discuss

15
American Institude of Aeronautics and Astronautics
standing waves, a brief background is presented here. Panda [38] reported the presence of standing waves (SWs) in
an underexpanded axisymmetric jet for the first time and attributed their presence to an interference between
downstream propagating hydrodynamic waves just outside of the jet (signature of large-scale structures convecting
downstream in the shear layer of the jet) and upstream propagating acoustic waves. He argued that the correct length
scale in Powell’s screech formula [39] is SW wavelength. This postulate would give the formation of standing waves
a key role in the generation of screech, and potentially the resultant coupling phenomenon. Accordingly, an
understanding of the presence and role of standing waves in the rectangular twin-jet flow field is an important part of
the overall understanding of the dynamics necessary to effectively implement flow control.
Panda managed to identify standing waves by obtaining maps of root mean square of pressure fluctuations. An
alternative approach to identifying standing waves was proposed by Edgington-Mitchell et al. [40], which involved
calculating the standard deviation of intensity values from sets of schlieren images. Bell et al. [41] later on visualized
standing waves in their particle image velocimetry (PIV) data by calculating Log10 of standard deviation from maps
of axial velocity. Aside from providing a length scale that can be used to predict screech frequency, the importance of
the SWs lies in the fact that their presence might be tied to the dynamics of coupling between the jets. By comparing
the Mach numbers at which standing waves were present against a plot of coherence level of screech for a single jet
(Fig. 7d), we noticed that standing waves appear whenever the coherence level of screech is rising, i.e. Mj = 1.20 –
1.35 and 1.65 – 1.80. Whenever coherence and amplitude attain their maximum values for each flow regime, i.e.
overexpanded and underexpanded, standing waves tend to disappear. We have not been able to find any reports in the
literature that tie the existence of standing waves to the coherence level of screech and believe this might be the first
account of such a relationship. Furthermore, none of the investigations in which standing waves have been addressed
[38,40–42] have been conducted using a rectangular nozzle. Therefore, more research is warranted to provide a better
understanding of these issues.
Maps of streamwise intensity gradients and Log10 of standard deviation of intensity gradients acquired along the
major axis of the twin-jet assembly are presented in Fig. 14 while Fig. 15 shows maps of Log10 of standard deviation
of intensity gradients along the minor axis. Fig. 14a shows a Mj = 1.35 jet which is representative of the coupling
characteristics observed in low overexpanded Mach numbers. As was shown in Fig. 7a, b and d, the coherence level
of screech is quite high at this Mach number and the screech amplitude is highest among the overexpanded Mach
numbers. Furthermore, the coherence of signals from microphones located above the jets is high indicating strong
coupling between the jets (see Fig. 8b and c). Several observations can be made from examining Fig. 14a. First, the
centerline fluctuations begin to decrease monotonically at approximately 6De. This location is likely to be the end of
the potential core where coherent structures break down and dissipate. Next, as Panda reported, there is a mismatch
between the location of the shock cell tips and the crests of the standing wave lobes [38]. While the crests sometimes
are nearly aligned with the shock cell tips (Fig. 14a, x/De = 1.8), in general they fall somewhere between the cell tips.
According to Panda, significant pressure differences between the crest and the trough of the SWs exist along the jet
boundary [38]. At Mj = 1.35, there is a ~20% difference in intensity fluctuation levels between the crest of the lobe
and its trough (see Fig. 14a).
Panda indicated that the SW pattern is three dimensional and usually would be wrapped around an axisymmetric
jet [38]. Knast et al. [42] observed the presence of SWs in a twin circular jet setup when the jets were viewed along
the twin-jet plane and normal to it. Examining Figs. 15a, b and c indicates that SWs are never present when the jets
are viewed along their minor axis. As mentioned previously, we did not detect evidence of presence of a screech loop
in the horizontal plan and high-speed schlieren imaging experiments did not show any lateral motion of the jets due
to shedding of vortical structures along the minor axis of the jet, as a result, the ingredients needed for the presence of
standing waves in the horizontal plan are non-existent.
We noticed that as the Mach number is increased from Mj = 1.35, the standing waves disappear and do not emerge
again until Mj = 1.65 (not shown here). Fig. 14b confirms the absence of standing waves at Mj = 1.50. The furthest
axial location displaying the maximum value of intensity fluctuation is now x/De = 7 indicating that the potential core
length has increased compared to Mj = 1.35. The spreading angle of the jet in the minor axis plane has also been
reduced. For Mj = 1.35, the edge of the jet at x/De = 9 is approximately at z/De = 2.9. For Mj = 1.50, however, at the
same axial location, the edge of the jet lines up with z/De = 2.3. That corresponds to a 20.7% lower spreading rate for
Mj = 1.50 (Fig. 14b).
Following the reemergence of standing waves at Mj = 1.65, the lobes become well-defined and dominate the near-
field at Mj = 1.70. For this case, the furthest axial location of maximum intensity fluctuation is at x/De = 8.2. There is
a 9.7% difference between the value of intensity fluctuations between the crest located at x/De = 3.4 and the adjacent
trough at x/De = 4.7. Due to the presence of standing waves, it is not possible to exactly measure the jet’s spreading
angle in the minor axis plane at Mj = 1.70, however, it appears that the jets spreading angle has not significantly
changed compared to Mj = 1.70. As reported by Shih et al. [34] and Alkislar et al. [3], the maximum spreading rate of

16
American Institude of Aeronautics and Astronautics
the jet occurs when the jets are flapping out-of-phase. The wavelength of the standing waves, on other hand, has been
increased compared to Mj = 1.35. For Mj = 1.35, the spacing between the crests is approximately 1De whereas for Mj
= 1.70, it is close to 2De. These increases appear to be comparable to growths in shock cell length going from Mj =
1.35 to 1.70. This is in line with observations of Knast et al. [42] in a twin circular jet setup where they reported that
changes in wavelength of standing waves closely follows those of shock cell spacing. Panda’s observations from a
converging single circular nozzle[38], however, were not similar.

(a)

(b)

(c)

Fig. 14 Maps of Log[(k)] where k ~ dρ/dx (top) and schlieren images (sensitive to axial density gradients
dρ/dx ) (bottom) for a) Mj = 1.35, b) Mj = 1.50, c) Mj = 1.70

17
American Institude of Aeronautics and Astronautics
 (a) (b)

(c)

Fig. 15 Maps of Log[(k)] where k ~ dρ/dx for a) Mj = 1.35, b) Mj = 1.50, c) Mj = 1.70

V. Conclusions
A new, low aspect ratio, twin rectangular jet assembly has been designed, constructed and tested across a wide
range of fully expanded jet Mach numbers to gain a thorough understanding of the coupling modes of the jets and the
effects of those modes on near-field sound pressure levels and far-field noise directivity. The insights gained will aid
in implementing active flow control for noise reduction and decoupling the jets in the upcoming experiments at the
Gas Dynamics and Turbulence Laboratory. When comparing the normalized screech frequencies from our facility
with the results of other researchers using rectangular nozzles with low to moderate aspect ratios (AR = 2 to 5) reported
in the literature, we noticed that area-based equivalent diameter (De), not nozzle height h, collapsed the screech data.
Wavelet coherence and phase between near-field acoustic measurements were used to investigate the screech and
coupling modes of the twin-jets at various Mach numbers. The results indicate that each individual jet always adopts

18
American Institude of Aeronautics and Astronautics
an anti-symmetric (flapping) screech mode regardless of the Mach number. The screech amplitude and coherence are
reduced at Mach numbers close to the design Mach number. A frequency mismatch was observed at low overexpanded
Mach numbers, which led to intermittent screech peaks and a phase drift between the jets. As the Mach number
increases, the frequency mismatch disappears and the strength of the screech amplitude increases. Coupling in
overexpanded Mach numbers in the vicinity of the design Mach number is out-of-phase while in the majority of
underexpanded cases, the jets are strongly coupled in-phase. Better understanding these dynamics will be crucial to
effectively implementing instability-leveraging flow control to control the jet coupling.
Dynamics of coupling in the near-field affect noise directivity patterns in the far-field. Far-field acoustic
measurements were conducted for azimuthal angles of φ = 0° and 90° and polar angles ranging from θ = 30° to 120°.
The amplitude of the BBSAN components in the far-field were found to be higher along the minor axis for Mj = 1.35
and 1.50 whereas for Mj = 1.70, this trend is reversed. Turbulent mixing noise is the dominant noise component at Mj
= 1.35 and 1.50. At Mj = 1.70, however, BBSAN amplitude approaches that of mixing noise. Flow-field visualizations
with a standard Z-type schlieren system revealed that standing waves are present at Mach numbers where strength of
the screech loop and screech amplitude was increasing. More work is needed to investigate and understand the
underlying reasons for the presence of standing waves.

Acknowledgements
The support of this research by the Office of Naval Research under contract No. N00014-19-1-2207 through Dr.
Steve Martens is gratefully acknowledged. Fruitful discussions with Dr. John Spyropoulos of NAVAIR are very much
appreciated. We would also like to thank Andrew Sais and Jack Tyszkiewicz for their assistance in setting up the
schlieren system and post-processing the near-field acoustic data.

References
[1] Dusa, D., Speir, D., Rowe, R., and Leavitt, L. Advanced Technology Exhaust Nozzle Development. In 19th
Joint Propulsion Conference, AIAA paper 1983-1286, June 1983.
[2] Wiegand, C. F-35 Air Vehicle Technology Overview. In 2018 Aviation Technology, Integration, and
Operations Conference, AIAA paper 2018-3368, June 2018
[3] Alkislar, M. B., Krothapalli, A., and Lourenco, L. M. “Structure of a Screeching Rectangular Jet: A
Stereoscopic Particle Image Velocimetry Study.” Journal of Fluid Mechanics, Vol. 489, 2003, pp. 121–154.
https://doi.org/10.1017/S0022112003005032.
[4] Krothapalli, A., Baganoff, D., and Karamcheti, K. “On the Mixing of a Rectangular Jet.” Journal of Fluid
Mechanics, Vol. 107, 1981, pp. 201–220. https://doi.org/10.1017/S0022112081001730.
[5] Mixson, J. S., and Roussos, L. A. Acoustic Fatigue: Overview of Activities at NASA Langley. NASA
Technical Memorandum NTM-89143, April 1987.
[6] Berndt, D. E. Dynamic pressure fluctuations in the internozzle region of a twin-jet nacelle. No. 841540. SAE
Technical Paper, 1984.
[7] Shaw, L. “Twin-Jet Screech Suppression.” Journal of Aircraft, Vol. 27, No. 8, 1990, pp. 708–715.
https://doi.org/10.2514/3.25344.
[8] Harper-Bourne, M. Twin-Jet near-Field Noise Prediction. In 6th Aeroacoustics Conference and Exhibit,
AIAA paper 2000-2084, June 2000.
[9] Zilz, D., and Wlezien, R. The Sensitivity of Near-Field Acoustics to the Orientation of Twin Two-
Dimensional Supersonic Nozzles. In 26th Joint Propulsion Conference, AIAA paper 1990-2149, July 1990.
[10] Zeierman, S., Gutmark, E., and Nosseir, N. Characteristics of Two Adjacent Rectangular Jets. In 30th
Aerospace Sciences Meeting and Exhibit, AIAA paper 1992-237, January 1992.
[11] Raman, G., and Taghavi, R. “Coupling of Twin Rectangular Supersonic Jets.” Journal of Fluid Mechanics,
Vol. 354, 1998, pp. 123–146. https://doi.org/10.1017/S0022112097007441.
[12] Raman, G., and Taghavi, R. “Resonant Interaction of a Linear Array of Supersonic Rectangular Jets: An
Experimental Study.” Journal of Fluid Mechanics, Vol. 309, 1996, pp. 93–111.
https://doi.org/10.1017/S0022112096001577.
[13] Bozak, R. Twin Jet Effects on Noise of Round and Rectangular Jets: Experiment and Model. In 20th
AIAA/CEAS Aeroacoustics Conference, AIAA paper 2014-2890, June 2014.

19
American Institude of Aeronautics and Astronautics
[14] Bozak, R., and Wernet, M. P. Subsonic Round and Rectangular Twin Jet Flow Effects. In 50th
AIAA/ASME/SAE/ASEE Joint Propulsion Conference, AIAA paper 2014-3736, July 2014.
[15] Walker, S. Twin Jet Screech Suppression Concepts Tested for 4.7 Percent Axisymmetric and Two-
Dimensional Nozzle Configurations. In 26th Joint Propulsion Conference, AIAA paper 90-2150, July 1990.
[16] Kuo, C.-W., Cluts, J., and Samimy, M. “Exploring Physics and Control of Twin Supersonic Circular Jets.”
AIAA Journal, Vol. 55, No. 1, 2017, pp. 68–85. https://doi.org/10.2514/1.J054977.
[17] Karnam, A., Baier, F., and Gutmark, E. J. Nature of Flow Field & Acoustics of Twin Supersonic Rectangular
Jets. In AIAA Scitech 2020 Forum, AIAA paper 2020-0500, January 2020.
[18] Kalagotla, D., Karnam, A., and Gutmark, E. J. Comparison of Flow Characteristics of Single and Twin
Rectangular Jets Using OVERFLOW Code. In AIAA Scitech 2020 Forum, AIAA paper 2020-1334, January
2020.
[19] Viswanath, K., Liu, J., Ramamurti, R., Karnam, A., Baier, F., and Gutmark, E. J. Noise Characteristics of
Low Aspect Ratio Supersonic Twin Jet Configuration. In AIAA Scitech 2020 Forum, AIAA paper 2020-0496,
January 2020.
[20] Utkin, Y. G., Keshav, S., Kim, J.-H., Kastner, J., Adamovich, I. V., and Samimy, M. “Development and Use
of Localized Arc Filament Plasma Actuators for High-Speed Flow Control.” Journal of Physics D: Applied
Physics, Vol. 40, No. 3, 2007, pp. 685–694. https://doi.org/10.1088/0022-3727/40/3/S06.
[21] Samimy, M., Adamovich, I., Webb, B., Kastner, J., Hileman, J., Keshav, S., and Palm, P. “Development and
Characterization of Plasma Actuators for High-Speed Jet Control.” Experiments in Fluids, Vol. 37, No. 4,
2004, pp. 577–588. https://doi.org/10.1007/s00348-004-0854-7.
[22] Samimy, M., Webb, N., and Crawley, M. “Excitation of Free Shear-Layer Instabilities for High-Speed Flow
Control.” AIAA Journal, Vol. 56, No. 5, 2018, pp. 1770–1791. https://doi.org/10.2514/1.J056610.
[23] Samimy, M., Kim, J.-H., Kastner, J., Adamovich, I., and Utkin, Y. “Active Control of High-Speed and High-
Reynolds-Number Jets Using Plasma Actuators.” Journal of Fluid Mechanics, Vol. 578, 2007, pp. 305–330.
https://doi.org/10.1017/S0022112007004867.
[24] Samimy, M., Kearney-Fischer, M., and Kim, J.-H. “High-Speed and High-Reynolds-Number Jet Control
Using Localized Arc Filament Plasma Actuators.” Journal of Propulsion and Power, Vol. 28, No. 2, 2012, pp.
269–280. https://doi.org/10.2514/1.B34272.
[25] Esfahani, A., Webb, N., and Samimy, M. “Flow Separation Control over a Thin Post-Stall Airfoil: Effects of
Excitation Frequency.” AIAA Journal, Vol. 57, No. 5, 2019, pp. 1826–1838.
https://doi.org/10.2514/1.J057796.
[26] Yugulis, K., Hansford, S., Gregory, J. W., and Samimy, M. “Control of High Subsonic Cavity Flow Using
Plasma Actuators.” AIAA Journal, Vol. 52, No. 7, 2014, pp. 1542–1554. https://doi.org/10.2514/1.J052668.
[27] Webb, N., and Samimy, M. “Control of Supersonic Cavity Flow Using Plasma Actuators.” AIAA Journal,
Vol. 55, No. 10, 2017, pp. 3346–3355. https://doi.org/10.2514/1.J055720.
[28] Samimy, M., Webb, N., and Esfahani, A. “Reinventing the Wheel: Excitation of Flow Instabilities for Active
Flow Control Using Plasma Actuators.” Journal of Physics D: Applied Physics, Vol. 52, No. 35, 2019, p.
354002. https://doi.org/10.1088/1361-6463/ab272d.
[29] Hahn, C., Kearney-Fischer, M., and Samimy, M. “On Factors Influencing Arc Filament Plasma Actuator
Performance in Control of High Speed Jets.” Experiments in Fluids, Vol. 51, 2011, pp. 1591–1603.
https://doi.org/10.1007/s00348-011-1172-5.
[30] Gutmark, E., Schadow, K. C., and Bicker, C. J. “Near Acoustic Field and Shock Structure of Rectangular
Supersonic Jets.” AIAA Journal, Vol. 28, No. 7, 1990, pp. 1163–1170. https://doi.org/10.2514/3.25187.
[31] Raman, G., and Rice, E. J. “Instability Modes Excited by Natural Screech Tones in a Supersonic Rectangular
Jet.” Physics of Fluids, Vol. 6, No. 12, 1994, pp. 3999–4008. https://doi.org/10.1063/1.868389.
[32] Karnam, A., Baier, F., and Gutmark, E. J. Near Field Acoustic Analysis of Cold Supersonic Rectangular Jets.
In AIAA Scitech 2019 Forum, AIAA paper 2019-0809, January 2019.
[33] Jeun, J., Wu, G. J., and Lele, S. K. Towards Large-Eddy Simulations of Twin Rectangular Jets Including
Screech. In AIAA Scitech 2020 Forum, AIAA paper 2020-0998, January 2020.
[34] Shih, C., Krothapalli, A., and Gogineni, S. “Experimental Observations of Instability Modes in a Rectangular
Jet.” AIAA Journal, Vol. 30, No. 10, 1992, pp. 2388–2394. https://doi.org/10.2514/3.11238.
[35] Panda, J., Raman, G., Zaman, K., Panda, J., Raman, G., and Zaman, K. Underexpanded Screeching Jets from
Circular, Rectangular and Elliptic Nozzles. In 3rd AIAA/CEAS Aeroacoustics Conference, American Institute
of Aeronautics and Astronautics.
[36] Raman, G. “Cessation of Screech in Underexpanded Jets.” Journal of Fluid Mechanics, Vol. 336, 1997, pp.
69–90. https://doi.org/10.1017/S002211209600451X.

20
American Institude of Aeronautics and Astronautics
[37] Tam, C. K. W. “Supersonic Jet Noise.” Annual Review of Fluid Mechanics, Vol. 27, No. 1, 1995, pp. 17–43.
https://doi.org/10.1146/annurev.fl.27.010195.000313.
[38] Panda, J. “An Experimental Investigation of Screech Noise Generation.” Journal of Fluid Mechanics, Vol.
378, 1999, pp. 71–96. https://doi.org/10.1017/S0022112098003383.
[39] Powell, A. “On the Noise Emanating from a Two-Dimensional Jet Above the Critical Pressure.” The
Aeronautical Quarterly, Vol. 4, No. 2, 1953, pp. 103–122. https://doi.org/10.1017/S0001925900000822.
[40] Edgington-Mitchell, D., Honnery, D. R., and Soria, J. “Staging Behaviour in Screeching Elliptical Jets:”
International Journal of Aeroacoustics, 14(7), 2015, pp.1005-1024. https://doi.org/10.1260/1475-
472X.14.7.1005.
[41] Bell, G., Soria, J., Honnery, D., and Edgington-Mitchell, D. “An Experimental Investigation of Coupled
Underexpanded Supersonic Twin-Jets.” Experiments in Fluids, Vol. 59, No. 9, 2018, p. 139.
https://doi.org/10.1007/s00348-018-2593-1.
[42] Knast, T., Bell, G., Wong, M., Leb, C. M., Soria, J., Honnery, D. R., and Edgington-Mitchell, D. “Coupling
Modes of an Underexpanded Twin Axisymmetric Jet.” AIAA Journal, Vol. 56, No. 9, 2018, pp. 3524–3535.
https://doi.org/10.2514/1.J056434.

21
American Institude of Aeronautics and Astronautics