You are on page 1of 12

AIAA JOURNAL

Vol. 60, No. 7, July 2022

Experimental Investigation of Flow and Noise Control


by Porous Coated Tandem Cylinder Configurations

Thomas F. Geyer∗
Technical Acoustics Group, Brandenburg University of Technology Cottbus - Senftenberg,
03046 Cottbus, Germany
https://doi.org/10.2514/1.J061180
The aerodynamic noise generated by aircraft landing gear components is often studied using a so-called tandem
cylinder configuration, consisting of two circular cylinders arranged in a streamwise row and separated by a certain
distance. The subject of the present study is the experimental investigation of the effect of porous coatings on the noise
generated by such a tandem cylinder configuration. This was done by performing acoustic and flow measurements in
Downloaded by INDIAN INST OF TECHN. BOMBAY on September 12, 2022 | http://arc.aiaa.org | DOI: 10.2514/1.J061180

a small aeroacoustic wind tunnel at Reynolds numbers (based on cylinder diameter) between 30,000 and 130,000. It is
found that the porous coatings lead to a decrease of both the tonal noise and the broadband noise and to a notable shift
of the resulting tonal noise peak toward lower frequencies. The noise reduction is mainly caused by the stabilizing
effect of the coating of the upstream cylinder on the wake flow and the absorption of turbulence by the coating of the
downstream cylinder. The observed shift of the frequency is most likely due to the mean velocity reduction in the wake
of the porous coated upstream cylinder and the interaction of the turbulent flow with the porous coating of the
downstream cylinder.

I. Introduction simulations. This includes effects such as cylinder spacing [10,15],


different cross-sectional areas [16–20], a nonzero angle of incidence
T ANDEM cylinder configurations are often examined in aero-
acoustic studies, as they constitute a simplified model for air-
craft landing gear components. In such a configuration, a circular
(staggering of the cylinders) [21–23], the yawing of the upstream [24]
or downstream cylinder [25], and measures to actively or passively
cylinder in a crossflow generates a turbulent wake that interacts with a reduce noise or related vibrations [26–29]. The current review is by
second cylinder located at a certain distance downstream of the first no means complete, and the reader is rather referred to the reviews
cylinder. This interaction leads to a strong generation of both tonal given in [30,31] to obtain a more complete overview on the flow
noise and broadband noise. In one of the configurations of the official around various tandem cylinder configurations.
tandem cylinder benchmark, the two cylinders have the same diam- In recent years, the application of flow-permeable materials to
eter d and the distance between them is 3.7d [1]. One of the results in bluff bodies with the aim of reducing the aerodynamic noise has
the work of Khorrami et al. [2] are spectra of the velocity fluctuations more and more become a research field of great interest. This
in the wake of this configuration and in the wake of a single cylinder. includes, but is not limited to, porous material applied to cylinders
It has to be mentioned, though, that in this configuration transition (see, e.g., [32–36]) and airfoils (see, e.g., [37–46]). Based on prom-
strips were applied to the upstream cylinder to produce a turbulent ising results obtained for the reduction of vortex shedding noise from
boundary layer. The single-cylinder configuration leads to a narrow- single cylinders coated with porous materials by the author [47–49],
band peak at a Strouhal number Sr (based on cylinder diameter) of the main purpose of the present study is to experimentally investigate
0.29. Adding the second cylinder produces a narrow peak as well, but the potential reduction of tandem cylinder noise using porous
at a lower Strouhal number of Sr  0.24. According to Wu et al. [3], material. Thereby, the study focuses on the tonal noise generation,
the vortex shedding from the upstream cylinder is suppressed when the which is due to the periodic shedding of vortices from the cylinders.
second cylinder is placed at distances less than 4d behind the upstream So far, the numerical studies by Liu et al. [50,51] are the only
cylinder. This leads to an increased spanwise coherence and a reduc- published cases where the use of porous coatings on tandem cylin-
tion of the shedding frequency compared to the single-cylinder case. der configurations was investigated. Their work consisted of two-
Benchmark tandem cylinder experiments were conducted at the dimensional unsteady Reynolds-averaged Navier–Stokes (URANS)
Quiet Flow Facility (QFF) at NASA [4] with diameter d  57.15 mm, simulations, coupled with the Ffowcs Williams-Hawkings analogy
an aspect ratio (ratio of length to diameter) of 16, and a Reynolds [52], on porous cylinder coatings in a tandem cylinder configuration,
number of Re  1.66 ⋅ 105 (based on d). As the benchmark contains at the same Reynolds number of 1.66 ⋅ 105 that was used in the
a variety of experimental data, including surface pressures on the benchmark study [4]. Both the porosity and the thickness of the
cylinders, velocity fluctuations in the wake, and far-field sound pres- porous material were varied, with the first taking values of 0.8 and
sure levels, there are a large number of experimental and numeri- 0.95 and the latter taking values of h∕d  0.1, 0.15, and 0.2 (nor-
cal investigations available that aim to replicate these results (e.g., malized with the inner diameter). Their results showed that porous
[1,2,4–12]). coatings decrease the peak level and the frequency of the dominant
Since landing gear noise is one of the major noise sources during tone and that high porosities and large thicknesses are especially
the landing phase [13,14], tandem cylinder configurations have been beneficial. They identified three overall mechanisms that contrib-
studied extensively in the past by means of both experiments and ute to the far-field noise reduction: The most efficient one is that a
porous coating of the upstream cylinder stabilizes the wake (upstream
Presented as Paper 2021-2266 at the AIAA Aviation 2021 Forum, Virtual cylinder wake control). In addition, a coating of the downstream
Event, August 2–6, 2021; received 15 August 2021; revision received 3 cylinder leads to the absorption of turbulent energy from the flow
February 2022; accepted for publication 18 February 2022; published online (impingement absorption) and to a stabilization of the wake from the
15 March 2022. Copyright © 2021 by the authors. Published by the American downstream cylinder (downstream cylinder wake control). Thus, a
Institute of Aeronautics and Astronautics, Inc., with permission. All requests modification of the upstream cylinder was found to have a much
for copying and permission to reprint should be submitted to CCC at www.
stronger effect than a modification of the downstream cylinder.
copyright.com; employ the eISSN 1533-385X to initiate your request. See
also AIAA Rights and Permissions www.aiaa.org/randp. Inspired by the results of these numerical investigations [50,51]
*Senior Researcher; currently Institute of Electrified Aero Engines, and because no experimental results are currently available, the aim
German Aerospace Center, 03050 Cottbus, Germany; thomas.geyer@dlr.de. of the present study is to provide experimental evidence for the noise
Member AIAA. reducing potential of porous cylinder coatings in such a tandem
4091
4092 GEYER

cylinder configuration. To this end, a variety of measurements on cylinders, neglecting any effect of the porous material on the solid–
tandem cylinder configurations modified with porous materials were fluid boundary. Second, the blockage of the wind tunnel should be as
conducted in an aeroacoustic wind tunnel. Thereby, the goal is to small as possible. And third, the aspect ratio should at least equal 7, as
identify the influence of the porous material parameters on the tonal that is the value for which periodic Kármán vortex shedding along the
noise generation and to help understanding the underlying physics. span can be expected [54]. It was finally decided to use 10 mm thick
However, it is important to note that due to practical restrictions the porous coatings, resulting in an outer cylinder diameter of D 
present investigation does not exactly replicate the cases studied by 40 mm. This yielded an aspect ratio of L∕D  7 and a wind tunnel
Liu et al. [50,51]. blockage ratio of approximately 0.17 (note that blockage ratios less
This paper is organized as follows: In Sec. II, the experimental than 0.2 are shown to have a negligible impact on the recorded
setup is explained in detail, including the wind tunnel, the cylinder Strouhal number [55]). The maximum Reynolds number that was
models as well as acoustic and flow measurement techniques. Results achieved with this diameter was 1.27 ⋅ 105 , which is below the value
from the far-field noise measurements and flow measurements are of 1.66 ⋅ 105 used by Jenkins et al. [1], Lockard et al. [4], and Liu et al.
then shown and discussed in Sec. III, while the last section gives [50,51]. However, it is important to note that both Reynolds numbers
conclusions of the study. belong to the so-called subcritical state of cylinder vortex shedding
[56], which is characterized by the fact that transition appears along
the free shear layers. The maximum Mach number in the present
II. Experimental Setup experiments is 0.14. Also note that in the present study, no trip tape
Downloaded by INDIAN INST OF TECHN. BOMBAY on September 12, 2022 | http://arc.aiaa.org | DOI: 10.2514/1.J061180

A. Wind Tunnel has been applied to the cylinders, whereas the upstream cylinder was
Experiments were conducted in the small aeroacoustic open jet wind tripped in the benchmark case [1,2].
tunnel at the Brandenburg University of Technology in Cottbus, Four different porous materials were used in this study, and they
Germany, which has a very low background noise [53]. For the are mainly characterized by their airflow resistivity r. The airflow
measurements on tandem cylinder configurations, the wind tunnel resistivity is a parameter that describes the resistivity of a porous
was equipped with a nozzle with rectangular exit area of 0.23 m× material against the fluid flow attempting to pass through it and is
0.28 m, which allows for maximum flow velocities of approximately inversely proportional to the porous material permeability K. The air
60 m∕s. At a flow speed of 50 m∕s, the turbulence intensity in front of flow resistivity of the materials was measured according to ISO 9053
the nozzle is in the order of 0.2%. The velocity is set by adjusting the [57]. In addition, the porosity σ of the materials was determined based
pressure inside the settling chamber, which is measured using a on the density of the foams and the density of the respective skeletal
pressure transducer with a high accuracy of 0.073% Full Scale (FS). material. The properties of the cylinder models used in the present
The cylinders were attached to side plates made of acrylic glass, study are summarized in Table 1.
which were mounted to the top and bottom of the nozzle (see Fig. 1a). In the experiments, the cylinders were arranged in a tandem con-
During acoustic measurements, the test section is surrounded by figuration with a distance of 3.7D  148 mm between the cylinder
sound absorbing side walls made of 0.25 m thick melamine foam, axes (see Fig. 1b). Although there are various studies on tandem
which lead to a quasi-anechoic environment for frequencies of cylinder configurations with other distances (see, e.g., [1,10,58), the
125 Hz and greater. chosen configuration is the one focused on in the benchmark [7].
Following the approach in the numerical studies by Liu et al. [50,51],
B. Porous Cylinders subsequent experiments were then performed to analyze the effect of
The porous coated cylinders of the present study consist of a bare a) a porous coating applied to the upstream cylinder, b) a porous coating
inner cylinder with a diameter of d  20 mm, covered by a layer of applied to the downstream cylinder, and c) a porous coating applied to
flow permeable material. The length of the cylinder models, both cylinders. The results are always compared to the baseline case of
L  0.28 m, is determined by the dimension of the wind tunnel two bare cylinders (with the same diameter as the outer diameter D of
nozzle. To select a value for the thickness of the porous coatings, the porous coated cylinders) in tandem configuration.
and hence the outer diameter D of the cylinders, three different
requirements had to be considered: First, it was intended to obtain C. Acoustic Measurement Technique
as high a Reynolds number as possible that is close to the value of In total, acoustic measurements were performed at seven flow
1.66 ⋅ 105 used by Lockard et al. at the QFF [4], within the restrictions velocities with corresponding Reynolds numbers ranging from
of the available facility. Note that the Reynolds number in the current 3.18 ⋅ 104 to 1.27 ⋅ 105 . They were done using three 1/4-inch free-
study is based on the geometric outer diameter D of the porous coated field measurement microphones of the type MI-17 that possess a

in m
0.76
A 0.44 m B 0.51 m C

microphones

0.4
A
B
nozzle
C
0.115
0
-0.115 side wall

3.7
-0.4
a) Photograph (Note that measurements were taken wit
the three microphones on the right side of the nozzle only.) in m
8
.1

14
0
-0

0.

b) Schematic (top view)


Fig. 1 Setup for the acoustic measurements inside the aeroacoustic wind tunnel.
GEYER 4093

Table 1 Properties of the cylinder models used in the present study

Outer diameter, Inner diameter, Length, Airflow resistivity,


Name Description D (mm) d (mm) L (m) r (Pa ⋅ s∕m2 ) Porosity, σ
Bare Reference 40 —— 0.28 ∞ 0
Foam 1 Regufoam vibration 150 plus 40 20 0.28 64,500 0.85 : : : 0.88
Foam 2 Basotect 40 20 0.28 9,800 ≈0.99
Panacell
Foam 3 40 20 0.28 4,000 0.96 : : : 0.98
90 ppi
Panacell
Foam 4 40 20 0.28 700 0.97 : : : 0.99
45 ppi

sensitivity of about 50 mV∕Pa and a typical frequency response of The microphone signals were recorded with a sampling frequency
0.5 dB in a frequency range from 20 Hz to 4 kHz. They were of 51.2 kHz and a duration of 60 s using a National Instruments 24 Bit
located on one side of the test section, with angles relative to multichannel measurement frontend. In postprocessing, the data were
the tandem cylinder configuration that equal those used in the converted to the frequency domain using a fast Fourier transformation
Downloaded by INDIAN INST OF TECHN. BOMBAY on September 12, 2022 | http://arc.aiaa.org | DOI: 10.2514/1.J061180

benchmark [1,4], but smaller distances due to the limited space (FFT) with Welch’s method [68], applied to Hanning-windowed
available. Figure 1b shows a schematic of the acoustic measurement blocks of 65,536 samples with 75% overlap. This results in a relative
setup, including the positions of the three microphones (the origin of error of the spectral estimate of 0.13 [69].
the coordinate system is located at the axis of the upstream cylinder).
The spectra obtained with the microphones were corrected in the D. Flow Measurement Technique
present work to represent values obtained at a distance corresponding In addition to the acoustic measurements, constant temperature
to the one used by Lockard et al. [4]. In addition, as the aspect ratio anemometry (CTA) measurements were performed at the maximum
in the current study (L∕D  7) also differs from that by Lockard Reynolds number of Re  1.27 ⋅ 105 for a subset of the configura-
et al. (L∕D  16), an additional correction was included for the tions investigated, using a Dantec P11-type single-wire probe with
differences in cylinder diameter and cylinder length. These correc- the wire parallel to the cylinder axis. In this configuration the probe
tions are based on the early work of Phillips [59], which states that for picks up both the streamwise component (x) of the flow velocity as
an observer at a distance R and an angle of 90° to the flow, the mean well as the component in the direction perpendicular to the flow and
square acoustic pressure in the far field the cylinder axis (the y component). As the latter frequently switches
directions (positive and negative), the resulting time-averaged veloc-
LDU60 Sr2 ities have to be interpreted with care. However, it has to be noted that
p2 R ∝ (1) this alignment of the wire was selected due to the possibility to get as
c2 R2 close to the cylinders as possible while providing access to the probe
holder at the same time.
where U0 is the flow velocity and c is the speed of sound. Based on The probe was positioned using a 3D traverse system made by Isel,
this theoretical dependence, the sound pressure level spectra obtained which has a minimum step size of 0.1 mm. In the majority of cases,
in the measurements on tandem cylinders in the current study were the velocity fluctuations were recorded with a sampling frequency of
corrected to account for the differences in R, L, and D compared to 25.6 kHz and a measurement duration of 8 s per point, while for a few
the measurements by Lockard et al. in the QFF [4]. A similar measurements the duration was increased to 60 s. The data were
correction procedure was also used in a study on the noise generation recorded using a Dantec multichannel CTA hardware system with a
by cylinders with different cross sections by Iglesias et al. [60] to 24-bit National Instruments digital signal acquisition module. Cali-
correct for differences in cylinder diameter. bration was done against a pitot tube using a fourth-order polynomial
When the sound, which is generated at the cylinders, travels to the fit, while a 90P10 temperature probe was used for temperature
microphones located outside the flow, it has to pass the wind tunnel correction.
shear layers and thereby it gets refracted. This is known to affect both In postprocessing, the first second of each data set was omitted to
the direction of travel (the angle between the sound source and the account for possible vibrations of the probe after each step of the
receivers) and the amplitude of the sound and depends on flow speed, traverse system. From the remaining samples, the mean velocity U as
shear layer shape and thickness, as well as frequency. Different well as power spectral densities of the turbulent velocity fluctuations
methods exist to account for this refraction effect. One of the most were calculated. The latter was done using an FFT on Hanning-
commonly used correction models is the model by Amiet [61], windowed blocks of 32,768 samples and an overlap of 50%, leading
which is based on geometrical acoustics and the assumption of zero- to the same frequency resolution as in the acoustic measurements.
thickness parallel shear layers of plane or cylindrical shape. Other
methods extend this approach to include, e.g., off-axis sources [62]
or different (convex) shear layer geometries [63,64]. In addition, III. Results
other correction schemes are often used, including ray-tracing [65] A. Far-Field Noise
or numerical [66,67] approaches. In the present study, however, no In this section, the results of the acoustic measurements on the
shear layer correction has been applied. This is due to two reasons. tandem cylinder configuration will be analyzed in order to investigate
First, the effects of such a correction were found to be very small for the effect of porous cylinder coatings on the noise generation. To
the current setup and the flow velocities considered here. For exam- allow for comparison with results from other studies, these analyses
ple, the corrected (retarded) angle for microphone B would be 75.4° were done using the data recorded by microphone B (see Fig. 1b).
compared to the measured angle of 74.3° at the highest flow velocity.
Second, the focus of the current acoustic measurements is the iden- 1. Modification of the Upstream Cylinder
tification of differences regarding the far-field noise due to a porous In a first step, the effect of a modification of only the upstream
coating compared to a bare cylinder, and a common refraction cylinder on the noise generated by the tandem cylinder configuration
correction would reveal identical corrections, both with respect to is examined. Figure 2 shows the corresponding sound pressure level
the amplitude of the far-field sound pressure and the angle, for both spectra for a Reynolds number (based on the outer cylinder diameter
cases. Thus, it was decided to not perform a shear layer refraction D) of 1.27 ⋅ 105 . The levels were corrected to correspond to the same
correction and hence to keep the paper more focused on the distance and aspect ratio as in the study by Lockard et al. [4] using
differences in the emitted noise due to a porous coating. Eq. (1). It is immediately visible that the different porous coatings
4094 GEYER

Fig. 2 Sound pressure level spectra obtained for the tandem cylinder configurations with microphone B (see Fig. 1b) with only the upstream cylinder
Downloaded by INDIAN INST OF TECHN. BOMBAY on September 12, 2022 | http://arc.aiaa.org | DOI: 10.2514/1.J061180

modified, at Re  1.27 ⋅ 105 ( bare–bare, foam 1–bare, foam 2–bare, foam 3–bare, foam 4–bare, background noise).

a) b)
Fig. 3 Maximum sound pressure level (a) and Strouhal number (b) of the vortex shedding peak obtained for the tandem cylinder configurations with
microphone B (see Fig. 1b) with only the upstream cylinder modified ( bare–bare, foam 1–bare, foam 2–bare, foam 3–bare, foam
4–bare).

affect the tonal peak in a different way. When the upstream cylinder is In addition, the porous coatings allow for a low-velocity flow through
coated with the porous material with the highest air flow resistivity the porous material, which enables a regularization of pressure
(foam 1 with r  64;500 Pa s∕m2 ), the center frequency of the differences. As a consequence, porous cylinder coatings were found
tonal peak is close to (but slightly higher than) that of the reference to lead to a notable reduction of both the mean velocity and the
tandem configuration (bare–bare), with a Strouhal number of 0.166 turbulence intensity in the wake [48]. Studies on porous airfoils have
compared to 0.163 for the reference. However, the width of the peak also shown that porous materials enforce a slowing down of the eddies
is smaller for the case with the porous coated cylinder, which agrees inside the boundary layer over the porous surface and an increased
with previous findings for porous coated single cylinders [47,48]. anisotropy in the flow over the porous surface [44]. However, past
The maximum sound pressure level for the porous case is higher than measurements on single cylinders also revealed that the porous coat-
that for the reference case. The remaining porous materials, when ings do not lead to a reduction of the spanwise coherence [48], which
applied to the upstream cylinder, lead to a notable shift of the has been observed for porous airfoils [42,44]. Combined, these mech-
tonal peak to lower frequencies. The corresponding peak Strouhal anisms result in a stabilization of the wake from the upstream cylinder,
numbers are 0.124 for foam 2 (r  9800 Pa ⋅ s∕m2 ) and 0.119 for which is in accordance with the numerical study by Liu et al. [51] and
foam 3 (r  4000 Pa ⋅ s∕m2 ) and foam 4 (r  700 Pa ⋅ s∕m2 ). will be analyzed in more detail in Sec. III.B.
Thus, it is important to note that the effect of the porous coatings is The effect of a porous coating applied to the upstream cylinder on
not a transition of the flow in the shear layers as one would expect the sound pressure level and the Strouhal number of the first tone
from tripping devices such as tape or wire, as according to experi- observed for the tandem cylinder configuration is shown in Fig. 3 as a
ments by Hutcheson et al. [10] that would lead to an increase of the function of Reynolds number (based on the outer cylinder diameter).
Strouhal number, not a decrease. The peak levels are again slightly Thereby, the sound pressure level maxima in Fig. 3a are scaled with the
increased compared to the bare–bare reference configuration, while sixth power of the flow Mach number, assuming a theoretical dipole
at the same time a notable narrowing of the tonal noise peak can be behavior of the noise generation [70,71]. This is in agreement with the
observed. In addition, the peaks of some of the configurations with velocity dependency utilized in Eq. (1). It has to be noted that past
coated upstream cylinder seem to be skewed compared to the refer- studies on airfoil self-noise-reduction using porous material [37] have
ence cylinder, which is especially visible for the configuration where shown that the porous consistency has an effect on the velocity
the upstream cylinder is coated with foam 1, but also for foam 3 and dependence, and hence different scalings for the individual coatings
foam 4. Thus, the coatings seem to affect the left side of the peak may lead to better results. However, the current approach has been
(toward lower frequencies) more than the right side of the peak, an chosen for the sake of simplicity, as it will at least allow for a better
effect that is not fully understood yet. comparison of the resulting levels. It can be seen that the scaled peak
The physical mechanisms that lead to the noise reduction are levels decrease with increasing Reynolds number, revealing that a
believed to be a viscous damping of turbulent flow pressure amplitudes smaller exponent for the Mach number may be more fitting. More
in the shear layer by the porous coating of the upstream cylinder. interesting, however, is the fact that the three foams with the low to
GEYER 4095

Fig. 4 Sound pressure level spectra obtained for the tandem cylinder configurations with microphone B (see Fig. 1b) with only the downstream cylinder
Downloaded by INDIAN INST OF TECHN. BOMBAY on September 12, 2022 | http://arc.aiaa.org | DOI: 10.2514/1.J061180

modified, at Re  1.27 ⋅ 105 ( bare–bare, bare–foam 1, bare–foam 2, bare–foam 3, bare–foam 4, background noise).

medium air flow resistivities† (foam 2 with r  9800 Pa ⋅ s∕m2 , foam assumed that it is not related to differences in noise directivity.
3 with r  4000 Pa ⋅ s∕m2 , and foam 4 with r  700 Pa ⋅ s∕m2 ) lead However, this effect shows a strong dependence on Reynolds num-
to notable reductions of the peak level at a large range of Reynolds ber, as it was only observed for the two highest Reynolds numbers of
numbers between approximately 30,000 and 70,000, which are in the the current study.
order of 10 dB. At the higher Reynolds numbers of the present study, Based on studies on the effect of porous leading edges on the
the peak levels measured for the porous cases approach that of the reduction of airfoil–turbulence interaction noise [46,73] it can be
bare–bare configuration, until at the two highest Reynolds numbers assumed that the physical mechanisms responsible for the reduction
all porous cases yield even higher peak levels than the reference of both the tonal noise (through a narrowing of the tonal noise peaks)
configuration. and the broadband noise, which is visible in Fig. 4 for the configu-
Figure 3b then shows the Strouhal numbers of the tonal peak as a rations with a porous coated downstream cylinder, are a combination
function of Reynolds number. In general, the Strouhal numbers for all of the following: an absorption of kinetic energy of the turbulent
configurations decrease with increasing Reynolds number. A slight eddies by the porous medium, a breaking of larger eddies into smaller
decrease in Strouhal number would also be expected for single ones when convecting over the porous surface, and a low-velocity
cylinders in the subcritical flow regime [72]. The most interesting internal flow through the material, leading to the increase of the
observation, however, is the fact that the difference between the boundary layer around the cylinder and a subsequent shift of the
Strouhal numbers obtained for the configurations with a porous vortices away from the cylinder surface, as well as to an interaction
coated upstream cylinder and that obtained for the bare–bare refer- between the eddies and the small jets emanating from the pores. In
ence configuration also show a dependence on the Reynolds number: agreement with the numerical study by Liu et al. [51] it is believed
For the porous material with high air flow resistivity (foam 1 with that the first mechanism, which is also known as hydrodynamic
r  64;500 Pa ⋅ s∕m2 ) the Strouhal number is close to that of the absorption, plays a major role regarding the noise reduction contrib-
reference configuration over the whole range of Reynolds numbers, uted by a porous coating of the downstream cylinder
whereas the Strouhal numbers obtained for the foams with the lowest Figure 5 then shows the sound pressure levels and Strouhal num-
air flow resistivities (foam 3 with r  4000 Pa ⋅ s∕m2 and foam 4 bers of the first spectral peak obtained for the configurations with a
with r  700 Pa ⋅ s∕m2 ) are notably lower, with a difference in the modified downstream cylinder compared to the bare–bare reference
order of 0.05 compared to the reference. Now, the Strouhal number configuration. Now, Fig. 5a reveals the above-mentioned observation
obtained for the configuration with a porous coating with medium air that the peak level measured for the configuration with foam 2 is
considerably lower than that of the remaining configurations at the
flow resistivity (foam 2 with r  9800 Pa ⋅ s∕m2 ) is similar to that of
two highest Reynolds numbers only. It should be noted here that
the reference configuration at low to moderate Reynolds numbers up
previous experiments with cylinders coated with this material yielded
to approximately 50,000, but then drops to values similar to those
such inconclusive results as well: In a study with porous coated
obtained for the configurations with porous materials with low air
cylinders with an outer diameter D  30 mm and inner diameters
flow resistivities. This reveals that in the Reynolds number range
d between 10 and 20 mm, leading to coating thicknesses from 5 to
between approximately 50,000 to 65,000 a flow transition exists near
10 mm, this material did not lead to a suppression of the vortex
the upstream cylinder coated with foam 2.
shedding tone [48]. In a different study on wall-mounted, porous
coated finite cylinders with a much smaller outer diameter of D 
2. Modification of the Downstream Cylinder
18 mm and an inner diameter of d  10 mm, the same material lead
In a second step, only the downstream cylinder of the tandem to a complete suppression of the vortex shedding tonal peak [49].
configuration was modified with porous material, while the upstream Thus, the effect observed here at the two highest Reynolds numbers
cylinder remained bare. The results are shown in Fig. 4, again for the may also be related to the aspect ratio. Further experiments, such as
maximum Reynolds number of 1.27 ⋅ 105 . Here, a very interesting particle image velocimetry measurements, or numerical simulations
observation can be made: While most of the porous coatings only would help to further analyze the cause of the vortex shedding peak
lead to a slight reduction of the frequency of the tonal peak with suppression. At the lower Reynolds numbers, all configurations with
similar peak levels compared to the reference configuration, the a porous coated downstream cylinder lead to peak levels that are
results obtained with foam 2 are different. Both the tonal peak level similar to those measured for the reference configuration.
as well as the broadband noise obtained for this material are notably Interestingly, the decrease in peak level for the configuration with
reduced compared to the reference. This noise reduction was also foam 2 does not affect the peak Strouhal number, which is shown in
observed at microphones A and C of the present study, so it can be Fig. 5b, as all configurations with a porous coated downstream
cylinder lead to Strouhal numbers only slightly below that of the

Different from [47], the term “low air flow resistivity” here refers to values reference configuration. Thus, the overall effect of a porous coating
r < 5 kPa ⋅ s∕m2 , “medium” means 5 kPa ⋅ s∕m2 < r < 10 kPa ⋅ s∕m2 , and applied to the downstream cylinder on the overall noise is much
“high” means r > 10 kPa ⋅ s∕m2 . smaller than the effect of a porous coating on the upstream cylinder.
4096 GEYER

a) b)
Fig. 5 Maximum sound pressure level (a) and Strouhal number (b) of the vortex shedding peak obtained for the tandem cylinder configurations with
microphone B (see Fig. 1b) with only the downstream cylinder modified ( bare–bare, bare–foam 1, bare–foam 2, bare–foam 3,
bare–foam 4).
Downloaded by INDIAN INST OF TECHN. BOMBAY on September 12, 2022 | http://arc.aiaa.org | DOI: 10.2514/1.J061180

This agrees well with the findings from Liu et al. [51], who state that a modification of only the upstream cylinder (see Fig. 3a), for
the effect of a coating of the downstream cylinder (an absorption of which the level differences, although notable, were still below
impingement energy and a stabilization of the downstream wake) is those observed in Fig. 7a. Again, this agrees well with the findings
considerably smaller than the stabilizing wake effect of a coating of from Liu et al. [51].
the upstream cylinder. The resulting Strouhal numbers are shown in Fig. 7b as a function
of Reynolds number. Basically, the results are similar to those
3. Modification of Both Cylinders observed for the configurations where only the upstream cylinder
Finally, the effect of a modification of both cylinders with porous was modified (see Fig. 3b for comparison): The porous coated
material on the far-field noise was investigated. The resulting sound cylinders all result in peak Strouhal numbers notably below that of
pressure level spectra are shown in Fig. 6 for a Reynolds number of the reference configuration over the whole range of Reynolds num-
1.27 ⋅ 105 for all porous coatings except foam 1, which was not used bers (although it should be noted that foam 4 yields slightly higher
because the overall performance of cylinders coated with this par- Strouhal numbers than foam 3 in the present case, and the opposite
ticular material is very similar to the bare reference cylinder. Basi- was true for the modification of only the upstream cylinder). This
cally, it can be seen that all porous materials again lead to a notable confirms that the effect of the coating of the downstream cylinder,
shift of the tonal peak to lower frequencies. This shift is slightly larger which is mainly the absorption of turbulent kinetic energy from the
for foam 2 and foam 3, for which the Strouhal number is 0.132, than inflow and a stabilization of the wake of the downstream cylinder, is
for foam 4, for which it is 0.140, compared to the value of 0.162 for much smaller than the stabilizing effect of a porous coating of the
the bare–bare reference configuration. The peak level also decreases upstream cylinder on the flow between the cylinders.
for all porous coatings compared to the reference configuration. Now,
the physical mechanisms responsible for this noise reduction are a B. Flow Measurement Results
combination of those that affect the upstream cylinder (and, sub- To provide insight into the phenomena leading to the observed
sequently, the wake of the upstream cylinder) and those that affect the shift of the vortex shedding peak of the tandem cylinder configuration
downstream cylinder. due to the porous coatings, results from flow measurements with both
The effect of a modification of both the upstream and the down- cylinders in place will be analyzed. This will mostly be done for the
stream cylinder with the same porous material on the level and the configurations where only the upstream cylinder is modified, as these
Strouhal number of the first tonal peak is shown in Fig. 7. It can be cases were found to show a noticeable shift of the vortex shedding
seen that all the porous coatings, when applied to both the upstream peak to lower frequencies. Additional results of flow measurements
and the downstream cylinder, lead to a reduction of the tonal peak in the wake of the single cylinders can be found in [74].
between roughly 10 dB at low Reynolds numbers to about 5 dB at Figure 8 presents mean velocities measured along a streamwise
high Reynolds numbers. Thus, the impact of a porous coating on line at y  0 between the two cylinders. As described in Sec. II.D, the
both cylinders is even more effective at reducing the tonal peak than measured velocity contains both the streamwise (x) component as

Fig. 6 Sound pressure level spectra obtained for the tandem cylinder configurations with microphone B (see Fig. 1b) with both the upstream and the
downstream cylinder modified, at Re  1.27 ⋅ 105 ( bare–bare, foam 2–foam 2, foam 3–foam 3, foam 4–foam 4, back-
ground noise).
GEYER 4097

a) b)
Fig. 7 Maximum sound pressure level (a) and Strouhal number (b) of the vortex shedding peak obtained for the tandem cylinder configurations with
microphone B (see Fig. 1b) with both the upstream and the downstream cylinder modified ( bare–bare, foam 2–foam 2, foam 3–foam 3,
foam 4–foam 4).
Downloaded by INDIAN INST OF TECHN. BOMBAY on September 12, 2022 | http://arc.aiaa.org | DOI: 10.2514/1.J061180

Fig. 8 Mean velocities measured along a streamwise line at y  0 between the tandem cylinders with only the upstream cylinder modified, at
Re  1.27 ⋅ 105 , upstream cylinder located at x∕D  0, downstream cylinder at x∕D  3.7 ( bare–bare, foam 1–bare, foam 2–bare,
foam 3–bare, foam 4–bare).

well as the lateral (y) component of the velocity field. It is visible that cylinder is bare (bare–foam 4), the resulting curve is very similar to
there are basically two groups of curves: The first one contains the the one measured for the reference tandem configuration consisting
result observed for the bare–bare reference configuration and the of two bare cylinders. The main difference is that the decrease of the
configuration where the upstream cylinder is coated with foam 1, mean velocity between x∕D  1.3 and x∕D ≈ 3.2 is stronger for the
the material with the highest air flow resistivity of the current study. configuration with the porous coated downstream cylinder. When
For these cases, the mean velocity rapidly increases from about 0.3U0 both cylinders are coated with foam 4, the general slope of the curve is
directly downstream of the upstream cylinder (at x∕D ≈ 0.5) to the closer to that observed when just the upstream cylinder is coated, as it
maximum value of 0.6 at x∕D ≈ 1.3. Then, the value slightly steadily increases with x∕D. However, the mean velocities are much
decreases to U∕U0 ≈ 0.55 at the measurement location closest to lower for this configuration, as it increases from about 0.03U0 at
the downstream cylinder. The curves obtained for the second group, x∕D ≈ 0.5 to 0.15U0 at x∕D ≈ 3.2. Overall, the results confirm that
which contains the configurations where the upstream cylinder is the mean velocity between the tandem cylinders is governed by the
coated with porous materials of low and medium air flow resistivity, upstream cylinder. When it is bare, the curve resembles that of the
are notably different, with some differences for the individual porous bare–bare reference configuration, nearly independent from a poten-
coatings. The mean velocity steadily increases from a very low value tial coating of the downstream cylinder. When the upstream cylinder
of approximately 0.03U0 until it reaches a plateau of U∕U0 ≈ 0.45 to is coated, the mean velocity is decreased over a large part of the wake.
0.5 for foams 2 and 3 and U∕U0 ≈ 0.38 for foam 4. Then the mean This agrees with the general trend observed for the porous coated
velocity increases again to a value of approximately 0.62U0 for single cylinders, as reported in [48]. It is also in qualitative agreement
foams 2 and 3 and 0.5U0 for foam 4 just upstream of the downstream with plots of the instantaneous vorticity field for porous coated
bare cylinder. This shows that the effect of a porous coating of the upstream cylinders compared to bare upstream cylinders provided
upstream cylinder with a material with low and medium air flow by Liu et al. [51].
resistivity leads to a much stronger wake deficit in a large region The results shown in Figs. 8 and 9 again confirm that a porous
downstream of the cylinder. However, in close proximity to the surface coating applied to the upstream cylinder has a strong effect on the
of the downstream cylinder the mean velocities at y  0 are very flowfield. The resulting reduction of the mean velocity can be
similar and virtually independent of the porous coating. Thus, the assumed to be an important reason for the observed shift of the vortex
spectral content of the turbulent flow has to be an important contributor shedding peak to lower frequencies. However, the fact that the addi-
to the observed shift of the peak of the far-field noise as well. tional modification of the downstream cylinder (foam 4–foam 4 in
The effect of a porous coating of only the downstream cylinder and Fig. 9) leads to a very strong decrease of the mean velocity, but only a
of both cylinders on the mean velocity between the tandem cylinders much smaller change of the peak Strouhal number (compare Fig. 2
will be exemplarily analyzed using only foam 4 (which has the lowest and Fig. 6) is evidence that the reduced mean velocity is not the only
air flow resistivity of the current study) as porous material. The results cause of the observed shift of the peak Strouhal number. Thus, the
are shown in Fig. 9. Basically, it is visible that when the upstream spectral content of the turbulent flow between the two cylinders needs
4098 GEYER

Fig. 9 Mean velocities measured along a streamwise line at y  0 between tandem cylinder configurations modified with foam 4, at Re  1.27 ⋅ 105 ,
Downloaded by INDIAN INST OF TECHN. BOMBAY on September 12, 2022 | http://arc.aiaa.org | DOI: 10.2514/1.J061180

upstream cylinder located at x∕D  0, downstream cylinder at x∕D  3.7 ( bare–bare, foam 4–bare, bare–foam 4, foam 4–foam 4).

to be analyzed as well. In addition, it is quite possible that porous compared to the bare–bare reference configuration. Over a large
coatings also behave quite unique with respect to the vortex shedding part of the streamwise distance between the two cylinders in the
tonal noise when subjected to a turbulent inflow, which is the case for region starting immediately downstream of the upstream cylinder
the configurations with a porous coated downstream cylinder. This (x∕D  0.5), no distinct tonal peaks are visible. This changes when
needs to be investigated in future studies. The fact that the porous the probe approaches the downstream cylinder around x∕D  0.25,
coated cylinders have a less distinct solid–fluid interface than a bare where again two peaks can be seen. However, these peaks now appear
cylinder, which effectively leads to an effective diameter smaller than at lower Strouhal numbers of approximately 0.125 and 0.25, which
the geometrical diameter, may also affect the vortex shedding in agrees well with the Strouhal numbers obtained from the correspond-
these cases. ing acoustic measurements as shown in Fig. 2. As for the bare–bare
To investigate the spectral content of the flowfield between the two reference configuration, the second peak at Sr ≈ 0.25 is much more
cylinders, corresponding spectra of the velocity fluctuations mea- pronounced. Thus, the analysis of these spectra reveals that a coating
sured along a streamwise line at y  0 between both cylinders will of the upstream cylinder with a porous material with low air flow
additionally be analyzed. They are shown in Fig. 10 as a function of resistivity weakens velocity fluctuation peaks in a large region down-
streamwise distance x∕D and Strouhal number Sr based on outer stream of the upstream cylinder, while essentially the same shift of the
cylinder diameter. For the bare–bare reference configuration, this plot tonal peaks to lower Strouhal numbers compared to the bare–bare
reveals two peaks: the first, much weaker peak appears around Sr  configuration as in the acoustic spectra can be seen in close proximity
0.17 and the second, stronger peak appears at twice this Strouhal to the downstream bare cylinder. In addition, the fact that the char-
number. In agreement with theory [70] it can be concluded that the acteristic tandem cylinder spectral signature is especially distinct
first peak is caused by the fluctuating lift force acting on the cylinder very close to the downstream cylinder indicates that the downstream
and the second by the fluctuating drag force. Comparing this result cylinder is the dominant location of the tonal noise source. This has
with results for the single bare cylinders in [74] reveals some inter- also been stated in prior tandem cylinder studies (see, e.g., [1,4).
esting differences, which can be attributed to the presence of the In a next step, turbulence spectra will be analyzed that were not
downstream cylinder: In the single-cylinder configuration, all cylin- measured at y∕D  0, but off-axis at y∕D  −0.5 between the two
ders generate a much stronger first peak and a much weaker second cylinders. These measurements were not taken along a streamwise
peak. For the bare–bare tandem cylinder configuration, the second line, but only at four selected locations (given are the x and y
peak at Sr ≈ 0.35 is much stronger than the first, revealing that the coordinates normalized with the outer diameter D, respectively):
addition of the downstream cylinder seems to weaken lift fluctuations P1  0.5; −0.5, P2  1.5; −0.5, P3  2.5; −0.5, and P4 
and/or to promote drag fluctuations. For the configuration where the 3.0; −0.5, with the axis of the upstream cylinder being located at
upstream cylinder is coated with the porous material with the highest (0.0, 0.0) and that of the downstream cylinder at (3.7, 0.0). For these
air flow resistivity (foam 1), the result is similar to the bare–bare measurements, which again were only performed for configurations
reference case, although the first peak is barely visible and the second with a porous coated upstream cylinder, the measurement duration
peak is much sharper. The results for the configurations where the was increased to 60 s. Figure 11 shows the resulting velocity spectra
upstream cylinder is coated with a porous material of low to medium as a function of Strouhal number. Basically, the results confirm the
air flow resistivity (foams 2–4), the results look much different findings from the measurements at y  0. The spectra obtained for

Fig. 10 Power spectral densities of the velocity fluctuations measured along a streamwise line at y  0 between the tandem cylinders with only the
upstream cylinder modified, at Re  1.27 ⋅ 105 (normalized with Φ0  1 m2 s−1 , x∕D  0 corresponds to the axis of the upstream cylinder and x∕D  3.7
to the axis of the downstream cylinder).
GEYER 4099

a) b) c) d)
Downloaded by INDIAN INST OF TECHN. BOMBAY on September 12, 2022 | http://arc.aiaa.org | DOI: 10.2514/1.J061180

Fig. 11 Power spectral densities of the velocity fluctuations measured at various positions in the shear layer of the upstream cylinder with both cylinders
in place, but only the upstream cylinder modified, at Re  1.27 ⋅ 105 ( bare–bare, foam 1–bare, foam 2–bare, foam 3–bare, foam
4–bare, normalized with Φ0  1 m2 s−1 : a) P 1  0.5D; − 0.5D, b) P 2  1.5D; − 0.5D, c) P 3  2.5D; − 0.5D, and d) P 4  3.0D; − 0.5D).

the bare–bare reference configuration and for the foam 1–bare con- the acoustic far-field. Additional hot-wire measurements were per-
figuration are similar in level, referring to both the broadband level formed between the two cylinders in an attempt to help understanding
and the level of the two peaks at Strouhal numbers of approximately the effect of the porous coating on the flowfield.
0.17 and 0.35 due to the fluctuating lift force and drag force, respec- Depending on which of the two cylinders is modified, it was found
tively. With decreasing distance to the downstream cylinder (from that the porous coatings lead to differences in the amplitude and the
Figs. 11a–11d), the level of the first peak decreases for the reference width of the tonal peak and even more notable differences in the
configuration, which is especially visible at the point closest to corresponding peak frequency. In agreement with the numerical
the downstream cylinder. For the remaining configurations (foam results from Liu et al. [51] it was observed that a modification of
2–bare, foam 3–bare, and foam 4–bare), the broadband level the upstream cylinder has a much stronger effect on the tonal noise
increases notably from the location closest to the upstream cylinder generation than a modification of the downstream cylinder. When the
at (x  0.5D, y  −0.5D) (Fig. 11a) to the location closest to the upstream cylinder is coated with a porous material with low to
downstream cylinder at (x  3.0D, y  −0.5D) (Fig. 11d). This is medium air flow resistivity, a notable shift of the tonal peak to lower
consistent with the trend observed in Fig. 8. The vortex shedding frequencies can be observed. This effect is not caused by a tripping of
peaks are barely visible at the first two locations, but increase with the shear layer to turbulence due to the increased surface roughness of
decreasing distance from the downstream cylinder. Consistent with the porous coating compared to the bare cylinder, since this would
the acoustic results and the velocity spectra obtained at y  0, how- lead to a shift of the peak to higher frequencies. The shift depends on
ever, the peaks can again be observed at notably lower Strouhal both the properties of the porous materials as well as on the Reynolds
numbers of approximately 0.125 (due to lift fluctuations) and 0.25 number. If only the downstream cylinder is modified, the shift is very
(due to drag fluctuations). Both the measurements at y  0 (Fig. 10) small. A modification of both cylinders even increases the trend
as well as the off-axis measurements (Fig. 11) show that the observed observed for the case of a modified upstream cylinder, as the shift
shift of the Strouhal number due to a porous coating, although it is is even more pronounced.
less distinct in close proximity to the upstream cylinder, is not a The analysis of the mean velocity and power spectral density of the
transitional effect that takes place between the two cylinders, but it velocity fluctuations between the cylinders revealed the complex
rather affects the whole flowfield around both cylinders. nature of the flowfield and showed notable differences between the
bare–bare reference configuration and the configurations for which
the upstream cylinder was coated with a porous material with low to
IV. Conclusions medium air flow resistivity. The reference configuration only shows a
The aerodynamic noise generated by a so-called tandem cylinder relatively short wake of the upstream cylinder, and hence a high mean
configuration in a flow, which serves as a simplified model for velocity over a large region between the two cylinders. The broad-
components of the landing gear of airplanes, is a classical problem band level of the velocity fluctuations measured between the two bare
in aeroacoustics. Since the use of porous coatings to reduce the cylinders is consistently high. The configurations with an upstream
aerodynamic noise from single cylinders in a crossflow has been cylinder modified with a material with low to medium air flow
shown to be very effective, the use of porous materials applied to a resistivity show a strongly reduced mean velocity over a large part
tandem cylinder configuration seems an obvious extension of that of the wake from the upstream cylinder with a very low broadband
approach. However, the only investigations on the effect of porous level of the corresponding velocity fluctuation spectra. Both the
coatings on the noise generated by such a tandem cylinder configu- broadband level and the tonal level increase strongly with decreasing
ration so far are the detailed numerical studies by Liu et al. [50,51]. To distance from the downstream cylinder. Interestingly, the shift of the
the knowledge of the authors, the current work represents the first frequency of the spectral peak due to lift fluctuations can already be
experimental study in which porous materials are applied to circular seen for the cases with porous coated upstream cylinders at locations
cylinders in a tandem configuration. very close to the upstream cylinder, even if the peaks are very weak
The porous coated cylinders in the present study consist of a only. Thus, the shift of the spectral peak is no transitional phenome-
nonporous inner cylinder with a diameter of 20 mm, which is covered non, but is visible at all locations between the cylinders. The notable
by a 10 mm thick porous layer, resulting in an outer diameter D of increase of the peak levels when the probe approaches the down-
40 mm. Four porous materials were used, which are characterized by stream cylinder leads to the conclusion that the spatial location of the
their air flow resistivity and their porosity. In addition, a configuration tonal noise sources is the downstream cylinder.
using two bare cylinders served as a reference. The tandem cylinder Regarding the noise reducing effect of the porous coatings, the
configuration contains two cylinders with a streamwise spacing of present study confirmed the findings from the numerical investiga-
3.7D, thus reproducing a well-known benchmark case [4]. Acoustic tions by Liu et al. [50,51]. A coating of the upstream cylinder has the
measurements were performed using single microphones located in highest impact, which is due to the stabilizing effect of the porous
4100 GEYER

material on the wake and the notable reduction of the mean velocity Conference (30th AIAA Aeroacoustics Conference), AIAA Paper
downstream of the upstream cylinder. This agrees well with previous 2009-3157, 2009.
investigations of the flowfield in the wake of single cylinders [48]. https://doi.org/10.2514/6.2009-3157
The second mechanism is the absorption of turbulent kinetic energy [7] Lockard, D., “Summary of the Tandem Cylinder Solutions from the
Benchmark Problems for Airframe Noise Computations-I Workshop,”
by the porous coating of the downstream cylinder. This has also been 49th AIAA Aerospace Sciences Meeting Including the New Horizons
observed in studies on the reduction of turbulence interaction noise Forum and Aerospace Exposition, AIAA Paper 2011-353, 2011.
through airfoils with porous leading edges [45,46,73]. A third cause https://doi.org/10.2514/6.2011-353
for the reduction of both the tonal noise and the broadband noise is the [8] Bres, G. A., Freed, D., Wessels, M., Noelting, S., and Pérot, F.,
stabilizing effect that the porous coating of the downstream cylinder “Flow and Noise Predictions for the Tandem Cylinder Aeroacoustic
has on the flow in the wake. All these effects have been found in Benchmark,” Physics of Fluids, Vol. 24, No. 3, 2012, Paper 036101.
previous studies to be more pronounced for porous materials with low https://doi.org/10.1063/1.3685102
airflow resistivity and high porosity (see, e.g., [40,48,73). This is [9] Uzun, A., and Hussaini, M. Y., “An Application of Delayed Detached
because, for typical open porous materials such as foams, a low Eddy Simulation to Tandem Cylinder Flow Field Prediction,” Com-
airflow resistivity and a high porosity are consistent with large pores, puters & Fluids, Vol. 60, 2012, pp. 71–85.
https://doi.org/10.1016/j.compfluid.2012.02.029
which will be able to absorb larger and more turbulent structures than
[10] Hutcheson, F. V., Brooks, T. F., Lockard, D. P., Choudhari, M. M., and
materials with smaller pores and, on average, more fluid will be able Stead, D. J., “Acoustics and Surface Pressure Measurements from
to enter the material. Hence, in the current study, materials with low Tandem Cylinder Configurations,” 20th AIAA/CEAS Aeroacoustics
Downloaded by INDIAN INST OF TECHN. BOMBAY on September 12, 2022 | http://arc.aiaa.org | DOI: 10.2514/1.J061180

airflow resistivity and high porosity were also found to be more Conference, AIAA Paper 2014-2762, 2014.
effective. https://doi.org/10.2514/6.2014-2762
Regarding the observed shift of the vortex shedding peak toward [11] Weinmann, M., Sandberg, R., and Doolan, C., “Tandem Cylinder
lower frequencies by the porous coatings, several mechanisms can be Flow and Noise Predictions Using a Hybrid RANS/LES Approach,”
identified. The first is the reduced mean velocity in the wake of the International Journal of Heat and Fluid Flow, Vol. 50, 2014, pp. 263–
upstream cylinder when that cylinder is coated with a material with 278.
low to medium air flow resistivity. Another possible reason is that the https://doi.org/10.1016/j.ijheatfluidflow.2014.08.011
porous coating of the downstream cylinder also leads to a shift of [12] Maryami, R., Showkat Ali, S. A., Azarpeyvand, M., Dehghan, A., and
the vortex shedding peak when the inflow is turbulent (as it is due to Afshari, A., “Experimental Study of the Unsteady Aerodynamic Load-
the presence of the upstream cylinder). This can be assumed to again ing for a Tandem Cylinder Configuration,” 25th AIAA/CEAS Aeroa-
coustics Conference, AIAA Paper 2019-2742, 2019.
be caused by the absorption of turbulence within the porous coating
https://doi.org/10.2514/6.2019-2742
as well as by the penetration of the porous coating with fluid. The [13] Lockard, D. P., and Lilley, G. M., “The Airframe Noise Reduction
effect of a turbulent inflow on the tonal noise generation by porous Challenge,” National Aeronautics and Space Administration, NASA/
coated single cylinders needs to be analyzed in future measurements TM–2004–213013, 2004.
in order to provide more details on this phenomenon. Finally, the [14] Dobrzynski, W., “Almost 40 Years of Airframe Noise Research:
virtual reduction of the diameter of the downstream porous coated What Did We Achieve?” Journal of Aircraft, Vol. 47, No. 2, 2010,
cylinder due to fluid entering the porous material may also play a role pp. 353–367.
regarding the observed shift of the vortex shedding peak. https://doi.org/10.2514/1.44457
[15] Papaioannou, G., Yue, D., Triantafyllou, M., and Karniadakis, G., “On
the Effect of Spacing on the Vortex-Induced Vibrations of Two Tandem
Acknowledgments Cylinders,” Journal of Fluids and Structures, Vol. 24, No. 6, 2008,
pp. 833–854.
The author thanks Baran Calisci (Technical Acoustics Group, https://doi.org/10.1016/j.jfluidstructs.2007.11.006
Brandenburg University of Technology Cottbus–Senftenberg, Cott- [16] Nishiyama, H., Ota, T., and Matsuno, T., “Heat Transfer and Flow Around
bus, Germany) for his help with the acoustic measurements, and Elias Elliptic Cylinders in Tandem Arrangement,” JSME International Journal.
Arcondoulis and Yu Liu (Department of Mechanics and Aerospace Ser. 2, Fluids Engineering, Heat Transfer, Power, Combustion, Thermo-
Engineering, Southern University of Science and Technology, physical Properties, Vol. 31, No. 3, 1988, pp. 410–419.
Shenzhen, People’s Republic of China) for the helpful discussions. https://doi.org/10.1299/jsmeb1988.31.3_410
[17] Luo, S., and Teng, T., “Aerodynamic Forces on a Square Section
Cylinder That Is Downstream to an Identical Cylinder,” Aeronautical
References Journal, Vol. 94, No. 936, 1990, pp. 203–212.
https://doi.org/10.1017/S0001924000022880
[1] Jenkins, L., Khorrami, M., Choudhari, M., and McGinley, C., “Charac-
[18] Yen, S.-C., San, K., and Chuang, T., “Interactions of Tandem Square
terization of Unsteady Flow Structures Around Tandem Cylinders for
Cylinders at Low Reynolds Numbers,” Experimental Thermal and Fluid
Component Interaction Studies in Airframe Noise,” 11th AIAA/CEAS
Science, Vol. 32, No. 4, 2008, pp. 927–938.
Aeroacoustics Conference, AIAA Paper 2005-2812, 2005.
https://doi.org/10.1016/j.expthermflusci.2007.07.001
https://doi.org/10.2514/6.2005-2812
[19] Knacke, T., and Thiele, F., “Prediction of Broadband Noise from
[2] Khorrami, M. R., Choudhari, M. M., Lockard, D. P., Jenkins, L. N., and
Two Square Cylinders in Tandem Arrangement Using a Combined
McGinley, C. B., “Unsteady Flowfield Around Tandem Cylinders as
Prototype Component Interaction in Airframe Noise,” AIAA Journal, DDES/FWH Approach,” Turbulence and Interactions, Springer, Berlin,
Vol. 45, No. 8, 2007, pp. 1930–1941. 2014, pp. 123–131.
https://doi.org/10.2514/1.23690 https://doi.org/10.1007/978-3-662-43489-5_15
[3] Wu, J., Welch, L., Welsh, M., Sheridan, J., and Walker, G., “Spanwise [20] Dawi, A. H., and Akkermans, R. A., “Direct and Integral Noise Com-
Wake Structures of a Circular Cylinder and Two Circular Cylinders in putation of Two Square Cylinders in Tandem Arrangement,” Journal of
Tandem,” Experimental Thermal and Fluid Science, Vol. 9, No. 3, 1994, Sound and Vibration, Vol. 436, 2018, pp. 138–154.
pp. 299–308. https://doi.org/10.1016/j.jsv.2018.09.008
https://doi.org/10.1016/0894-1777(94)90032-9 [21] Akbari, M., and Price, S., “Numerical Investigation of Flow Patterns for
[4] Lockard, D., Khorrami, M., Choudhari, M., Hutcheson, F., Brooks, Staggered Cylinder Pairs in Cross-Flow,” Journal of Fluids and Struc-
T. F., and Stead, D., “Tandem Cylinder Noise Predictions,” 13th tures, Vol. 20, No. 4, 2005, pp. 533–554.
AIAA/CEAS Aeroacoustics Conference (28th AIAA Aeroacoustics https://doi.org/10.1016/j.jfluidstructs.2005.02.005
Conference), AIAA Paper 2007-3450, 2007. [22] Sumner, D., Richards, M., and Akosile, O., “Two Staggered Circular
https://doi.org/10.2514/6.2007-3450 Cylinders of Equal Diameter in Cross-Flow,” Journal of Fluids and
[5] Neuhart, D., Jenkins, L., Choudhari, M., and Khorrami, M., “Measure- Structures, Vol. 20, No. 2, 2005, pp. 255–276.
ments of the Flowfield Interaction Between Tandem Cylinders,” 15th https://doi.org/10.1016/j.jfluidstructs.2004.10.006
AIAA/CEAS Aeroacoustics Conference (30th AIAA Aeroacoustics [23] Schewe, G., and Jacobs, M., “Experiments on the Flow Around
Conference), AIAA Paper 2009-3275, 2009. Two Tandem Circular Cylinders from Sub-Up to Transcritical
https://doi.org/10.2514/6.2009-3275 Reynolds Numbers,” Journal of Fluids and Structures, Vol. 88, 2019,
[6] Doolan, C., “Flow and Noise Simulation of the NASATandem Cylinder pp. 148–166.
Experiment Using OpenFOAM,” 15th AIAA/CEAS Aeroacoustics https://doi.org/10.1016/j.jfluidstructs.2019.05.001
GEYER 4101

[24] Wilkins, S. J., and Hall, J. W., “Experimental Investigation of a Tandem Vol. 80, 2019, Paper 108469.
Cylinder System with a Yawed Upstream Cylinder,” Journal of Pressure https://doi.org/10.1016/j.ijheatfluidflow.2019.108469
Vessel Technology, Vol. 136, No. 1, 2014. [43] Rubio Carpio, A., Martínez, R. M., Avallone, F., Ragni, D., Snellen,
https://doi.org/10.1115/1.4025612 M., and van der Zwaag, S., “Experimental Characterization of the
[25] Wilkins, S. J., Hogan, J. D., and Hall, J. W., “Vortex Shedding in a Turbulent Boundary Layer over a Porous Trailing Edge for Noise
Tandem Circular Cylinder System with a Yawed Downstream Cylin- Abatement,” Journal of Sound and Vibration, Vol. 443, 2019,
der,” Journal of Fluids Engineering, Vol. 135, No. 7, 2013. pp. 537–558.
https://doi.org/10.1115/1.4023949 https://doi.org/10.1016/j.jsv.2018.12.010
[26] Kozlov, A. V., and Thomas, F. O., “Plasma Flow Control of Cylinders [44] Ananthan, V. B., Bernicke, P., Akkermans, R., Hu, T., and Liu, P.,
in a Tandem Configuration,” AIAA Journal, Vol. 49, No. 10, 2011, “Effect of Porous Material on Trailing Edge Sound Sources of a Lift-
pp. 2183–2193. ing Airfoil by Zonal Overset-LES,” Journal of Sound and Vibration,
https://doi.org/10.2514/1.J050976 Vol. 480, 2020, Paper 115386.
[27] Eltaweel, A., Wang, M., Kim, D., Thomas, F. O., and Kozlov, A. V., https://doi.org/10.1016/j.jsv.2020.115386
“Numerical Investigation of Tandem-Cylinder Noise Reduction Using [45] Zamponi, R., Satcunanathan, S., Moreau, S., Ragni, D., Meinke, M.,
Plasma-Based Flow Control,” Journal of Fluid Mechanics, Vol. 756, Schröder, W., and Schram, C., “On the Role of Turbulence Distortion on
2014, pp. 422–451. Leading-Edge Noise Reduction by means of Porosity,” Journal of
https://doi.org/10.1017/jfm.2014.420 Sound and Vibration, Vol. 485, 2020, Paper 115561.
[28] Li, L., Liu, P., Xing, Y., and Guo, H., “Experimental Investigation on the https://doi.org/10.1016/j.jsv.2020.115561
Noise Reduction Method of Helical Cables for a Circular Cylinder and [46] Ocker, C., Geyer, T. F., Czwielong, F., Krömer, F., Pannert, W., Merkel,
Downloaded by INDIAN INST OF TECHN. BOMBAY on September 12, 2022 | http://arc.aiaa.org | DOI: 10.2514/1.J061180

Tandem Cylinders,” Applied Acoustics, Vol. 152, 2019, pp. 79–87. M., and Becker, S., “Permeable Leading Edges for Airfoil and Fan Noise
https://doi.org/10.1016/j.apacoust.2019.03.027 Reduction in Disturbed Inflow,” AIAA Journal, Vol. 59, No. 12, 2021,
[29] Rabiee, A. H., and Esmaeili, M., “Simultaneous Vortex- and Wake- pp. 1–18.
Induced Vibration Suppression of Tandem-Arranged Circular Cylinders https://doi.org/10.2514/1.J060396
Using Active Feedback Control System,” Journal of Sound and Vibra- [47] Geyer, T. F., and Sarradj, E., “Circular Cylinders with Soft Porous Cover
tion, Vol. 469, 2020, Paper 115131. for Flow Noise Reduction,” Experiments in Fluids, Vol. 57, No. 3, 2016,
https://doi.org/10.1016/j.jsv.2019.115131 p. 30.
[30] Sumner, D., “Two Circular Cylinders in Cross-Flow: A Review,” https://doi.org/10.1007/s00348-016-2119-7
Journal of Fluids and Structures, Vol. 26, No. 6, 2010, pp. 849–899. [48] Geyer, T. F., “Experimental Evaluation of Cylinder Vortex Shedding
https://doi.org/10.1016/j.jfluidstructs.2010.07.001 Noise Reduction Using Porous Material,” Experiments in Fluids,
[31] Zhou, Y., and Alam, M. M., “Wake of Two Interacting Circular Cylin- Vol. 61, No. 7, 2020, pp. 1–21.
ders: A Review,” International Journal of Heat and Fluid Flow, Vol. 62, https://doi.org/10.1007/s00348-020-02972-0
2016, pp. 510–537. [49] Geyer, T. F., “Vortex Shedding Noise from Finite, Wall-Mounted,
https://doi.org/10.1016/j.ijheatfluidflow.2016.08.008 Circular Cylinders Modified with Porous Material,” AIAA Journal,
[32] Sueki, T., Takaishi, T., Ikeda, M., and Arai, N., “Application of Porous Vol. 58, No. 5, 2020, pp. 2014–2028.
Material to Reduce Aerodynamic Sound from Bluff Bodies,” Fluid https://doi.org/10.2514/1.J058877
Dynamics Research, Vol. 42, No. 1, 2010, Paper 015004. [50] Liu, H., Azarpeyvand, M., Wei, J., and Qu, Z., “Tandem Cylinder
https://doi.org/10.1088/0169-5983/42/1/015004 Aerodynamic Sound Control Using Porous Coating,” Journal of Sound
[33] Liu, H., Wei, J., and Qu, Z., “Prediction of Aerodynamic Noise Reduc- and Vibration, Vol. 334, 2015, pp. 190–201.
tion by Using Open-Cell Metal Foam,” Journal of Sound and Vibration, https://doi.org/10.1016/j.jsv.2014.09.013
Vol. 331, No. 7, 2012, pp. 1483–1497. [51] Liu, H., Wang, Y., Wei, J., and Qu, Z., “The Importance of Controlling
https://doi.org/10.1016/j.jsv.2011.11.016 the Upstream Body Wake in Tandem Cylinders System for Noise
[34] Arcondoulis, E. J., Liu, Y., Li, Z., Yang, Y., and Wang, Y., “Structured Reduction,” Proceedings of the Institution of Mechanical Engineers,
Porous Material Design for Passive Flow and Noise Control of Cylin- Part G: Journal of Aerospace Engineering, Vol. 232, No. 3, 2018,
ders in Uniform Flow,” Materials, Vol. 12, No. 18, 2019, p. 2905. pp. 517–531.
https://doi.org/10.3390/ma12182905 https://doi.org/10.1177/0954410016682271
[35] Hu, Z., Liu, H., Chen, N., Hu, J., and Tong, F., “Vortex Shedding Noise [52] Ffowcs Williams, J., and Hawkings, D. L., “Sound Generation by
and Flow Mode Analysis of Cylinder with Full/Partial Porous Coating,” Turbulence and Surfaces in Arbitrary Motion,” Philosophical Trans-
Aerospace Science and Technology, Vol. 106, 2020, Paper 106154. actions for the Royal Society of London. Series A, Mathematical and
https://doi.org/10.1016/j.ast.2020.106154 Physical Sciences, Vol. 264, No. 1151, 1969, pp. 321–342.
[36] Wen, K., Arcondoulis, E. J., Li, Z., and Liu, Y., “Structure Resolved https://doi.org/10.1098/rsta.1969.0031
Simulations of Flow Around Porous Coated Cylinders Based on a [53] Sarradj, E., Fritzsche, C., Geyer, T. F., and Giesler, J., “Acoustic and
Simplified Pore-Scale Model,” Aerospace Science and Technology, Aerodynamic Design and Characterization of a Small-Scale Aeroacous-
Vol. 119, 2021, Paper 107181. tic Wind Tunnel,” Applied Acoustics, Vol. 70, No. 8, 2009, pp. 1073–
https://doi.org/10.1016/j.ast.2021.107181 1080.
[37] Geyer, T. F., Sarradj, E., and Fritzsche, C., “Measurement of the Noise https://doi.org/10.1016/j.apacoust.2009.02.009
Generation at the Trailing Edge of Porous Airfoils,” Experiments in [54] Porteous, R., Moreau, D. J., and Doolan, C. J., “A Review of Flow-
Fluids, Vol. 48, No. 2, 2010, pp. 291–308. Induced Noise from Finite Wall-Mounted Cylinders,” Journal of Fluids
https://doi.org/10.1007/s00348-009-0739-x and Structures, Vol. 51, 2014, pp. 240–254.
[38] Roger, M., Schram, C., and De Santana, L., “Reduction of Airfoil https://doi.org/10.1016/j.jfluidstructs.2014.08.012
Turbulence-Impingement Noise by means of Leading-Edge Serrations [55] Richter, A., and Naudascher, E., “Fluctuating Forces on a Rigid Circular
and/or Porous Material,” 19th AIAA/CEAS Aeroacoustics Conference, Cylinder in Confined Flow,” Journal of Fluid Mechanics, Vol. 78, No. 3,
AIAA Paper 2013-2108, 2013. 1976, pp. 561–576.
https://doi.org/10.2514/6.2013-2108 https://doi.org/10.1017/S0022112076002607
[39] Schulze, J., and Sesterhenn, J., “Optimal Distribution of Porous Media [56] Zdravkovich, M. M., Flow Around Circular Cylinders: Volume 1:
to Reduce Trailing Edge Noise,” Computers & Fluids, Vol. 78, 2013, Fundamentals, Vol. 1, Oxford Univ. Press, Oxford, 1997, Chap. 1.
pp. 41–53. [57] Acoustics: Materials for Acoustical Applications–Determination of
https://doi.org/10.1016/j.compfluid.2011.12.022 Airflow Resistance, ISO 9053:1991, International Organization for
[40] Sarradj, E., and Geyer, T. F., “Symbolic Regression Modeling of Noise Standardization, Geneva, 1991.
Generation at Porous Airfoils,” Journal of Sound and Vibration, [58] Alam, M. M., “The Aerodynamics of a Cylinder Submerged in the
Vol. 333, No. 14, 2014, pp. 3189–3202. Wake of Another,” Journal of Fluids and Structures, Vol. 51, 2014,
https://doi.org/10.1016/j.jsv.2014.02.037 pp. 393–400.
[41] Ali, S. A. S., Azarpeyvand, M., and Da Silva, C. R. I., “Trailing-Edge https://doi.org/10.1016/j.jfluidstructs.2014.08.003
Flow and Noise Control Using Porous Treatments,” Journal of Fluid [59] Phillips, O. M., “The Intensity of Aeolian Tones,” Journal of Fluid
Mechanics, Vol. 850, 2018, pp. 83–119. Mechanics, Vol. 1, No. 6, 1956, pp. 607–624.
https://doi.org/10.1017/jfm.2018.430 https://doi.org/10.1017/S0022112056000408
[42] Bernicke, P., Akkermans, R., Ananthan, V. B., Ewert, R., Dierke, J., and [60] Iglesias, E. L., Thompson, D., and Smith, M., “Experimental Study of
Rossian, L., “A Zonal Noise Prediction Method for Trailing-Edge Noise the Aerodynamic Noise Radiated by Cylinders with Different Cross-
with a Porous Model,” International Journal of Heat and Fluid Flow, Sections and Yaw Angles,” Journal of Sound and Vibration, Vol. 361,
4102 GEYER

2016, pp. 108–129. Short, Modified Periodograms,” IEEE Transactions on Audio and
https://doi.org/10.1016/j.jsv.2015.09.044 Electroacoustics, Vol. 15, No. 2, 1967, pp. 70–73.
[61] Amiet, R., “Refraction of Sound by a Shear Layer,” Journal of Sound https://doi.org/10.1109/TAU.1967.1161901
and Vibration, Vol. 58, No. 4, 1978, pp. 467–482. [69] Bendat, J. S., and Piersol, A. G., Random Data: Analysis and Measure-
https://doi.org/10.1016/0022-460X(78)90353-X ment Procedures, Wiley, New York, 2010, Chap. 8.
[62] Morfey, C., and Joseph, P., “Shear Layer Refraction Corrections for Off- https://doi.org/10.1002/9781118032428
Axis Sources in a Jet Flow,” Journal of Sound and Vibration, Vol. 239, [70] Etkin, B., Korbacher, G., and Keefe, R. T., “Acoustic Radiation from a
No. 4, 2001, pp. 819–848. Stationary Cylinder in a Fluid Stream (Aeolian Tones),” Journal of the
https://doi.org/10.1006/jsvi.2000.3218 Acoustical Society of America, Vol. 29, No. 1, 1957, pp. 30–36.
[63] Porteous, R., Geyer, T. F., Moreau, D. J., and Doolan, C. J., “A https://doi.org/10.1121/1.1908673
Correction Method for Acoustic Source Localisation in Convex [71] Keefe, R. T., “Investigation of the Fluctuating Forces Acting on a
Shear Layer Geometries,” Applied Acoustics, Vol. 130, 2018, Stationary Circular Cylinder in a Subsonic Stream and of the Associated
pp. 128–132. Sound Field,” Journal of the Acoustical Society of America, Vol. 34,
https://doi.org/10.1016/j.apacoust.2017.09.020 No. 11, 1962, pp. 1711–1714.
[64] Wang, L., Chen, R., You, Y., Wu, W., and Qiu, R., “A Unified Correction https://doi.org/10.1121/1.1909102
Method for the Acoustic Refraction (UCMAR) Caused by a Three [72] Fey, U., König, M., and Eckelmann, H., “A New Strouhal–Reynolds-
Dimensional Shear Layer,” Acta Acustica United with Acustica, Number Relationship for the Circular Cylinder in the Range,” Physics of
Vol. 105, No. 5, 2019, pp. 732–742. Fluids, Vol. 10, No. 7, 1998, pp. 1547–1549.
https://doi.org/10.3813/AAA.919353 https://doi.org/10.1063/1.869675
Downloaded by INDIAN INST OF TECHN. BOMBAY on September 12, 2022 | http://arc.aiaa.org | DOI: 10.2514/1.J061180

[65] Sarradj, E., “A Fast Ray Casting Method for Sound Refraction at Shear [73] Geyer, T. F., Sarradj, E., Giesler, J., and Hobracht, M., “Experimental
Layers,” International Journal of Aeroacoustics, Vol. 16, Nos. 1–2, Assessment of the Noise Generated at the Leading Edge of Porous
2017, pp. 65–77. Airfoils Using Microphone Array Techniques,” 17th AIAA/CEAS Aero-
https://doi.org/10.1177/1475472X16680463 acoustics Conference (32nd AIAA Aeroacoustics Conference), AIAA
[66] Bennaceur, I., Mincu, D.-C., Mary, I., Terracol, M., Larchevêque, L., Paper 2011-2713, 2011.
and Dupont, P., “Numerical Simulation of Acoustic Scattering by a https://doi.org/10.2514/6.2011-2713
Plane Turbulent Shear Layer: Spectral Broadening Study,” Computers [74] Geyer, T. F., Arcondoulis, E., and Liu, Y., “Experimental Investigation
& Fluids, Vol. 138, 2016, pp. 83–98. of Noise Generation by Porous Coated Tandem Cylinder Configura-
https://doi.org/10.1016/j.compfluid.2016.08.012 tions,” AIAA AVIATION 2021 FORUM, AIAA Paper 2021-2266,
[67] Jiao, J., “Aeroacoustic Wind Tunnel Correction Based on Numerical 2021.
Simulation,” Ph.D. Thesis, Technical Univ. Braunschweig, Braun- https://doi.org/10.2514/6.2021-2266
schweig, Germany, 2017.
[68] Welch, P., “The Use of Fast Fourier Transform for the Estimation D. Papamoschou
of Power Spectra: A Method Based on Time Averaging over Associate Editor

You might also like