Applied Acoustics 114 (2016) 10–17

Contents lists available at ScienceDirect

Applied Acoustics
journal homepage: www.elsevier.com/locate/apacoust

Technical note

A note on a circular panel sound absorber with an elastic boundary

condition
Jorge P. Arenas ⇑, Felipe Ugarte
Institute of Acoustics, Univ. Austral de Chile, PO Box 567, Valdivia, Chile

a r t i c l e i n f o a b s t r a c t

Article history: Panel sound absorbers are typically used to absorb low-frequency noise in concert halls, auditoriums,
Received 22 February 2016 recording studios, and other architectural applications. These systems are composed of flexible panels
Received in revised form 29 June 2016 mounted over an air space that can be either partly or completely filled with a porous material. In this
Accepted 5 July 2016
paper, a theoretical model is derived for predicting the sound-absorption coefficient of a cylindrical
low-frequency absorber made of a circular plate. The theory takes into account the mass, bending stiff-
ness, damping loss and the elastic boundary condition of the circular plate. The effects of the stiffness of
Keywords:
an air-back cavity and of partially adding a porous material into the cavity are also considered. It is
Panel absorber
Low-frequency absorption
observed that the low-frequency resonances of such a system are dependent upon the clamping condi-
Elastic plate tion, the width of the air-back cavity, and mechanical properties of the plate. There is good agreement
Noise control between theoretical and experimental results.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction usually quite complex because the mechanical behavior depends

on the way of clamping, the physical properties of the plate, and
In many architectural applications there is a need to absorb the air volume and filling.
low-frequency noise, which is difficult to achieve by using porous
sound-absorbing materials. Panel sound absorbers are typically
1.1. Previous studies
used to absorb low-frequency sound in concert halls, auditoriums,
recording studios and other architectural applications. Basically,
Ford and McCormick [1] were the first to develop a theory for
these systems are composed of flexible panels made of wood,
square panel absorbers including the effect of the plate’s bending
metal, gypsum or plastic board, mounted over an air space that
waves. This model has been used by other authors for analyzing
can be either partly or completely filled with a porous material.
low-frequency absorber systems. Ford and McCormick employed
Thus, several flexural resonance modes can be excited by an inci-
the virtual work principle to derive the equation of motion of a thin
dent sound wave, and maximum absorption takes place at the low-
and homogenous square plate with clamped boundary conditions
est natural frequency of the coupled panel-cavity system.
that is subjected to plane wave excitation.
If a sufficiently large plate is fastened over its entire surface area
Fuchs [2] discussed the acoustic behavior of sound absorbers
in front of an air cavity and to an absorbing layer located in front of
made of square plates and established design recommendations.
a rigid wall, the plate as a whole is allowed to oscillate in a piston-
In his book, he showed the sound absorption of the panel for differ-
like manner on the equivalent spring formed by the layers. Then,
ent damping conditions of the system. The maximum sound-
the resonance frequency of the plate absorber can be determined
absorption coefficient is shifted to low frequencies as the surface
from the mass of the plate and the stiffness of the equivalent spring
mass of the panel and the thickness of the air cavity are increased.
formed by the elastic layers in the cavity (in this case, the plate’s
Evidently, the increase in the air cavity thickness is usually limited
bending stiffness is not important). It is well-known that the heav-
by the available space in the room, and the increase of the mass can
ier the plate and the softer the layer, the lower the resonance fre-
add to the costs of the solution. Fasold and Veres [3] have stated
quency. However, if the plate is clamped it will perform bending
that the air cavity thickness does not have to be too small or too
oscillations. In this case, computation of its eigenfrequencies is
large, compared with the wavelength of the sound wave that is
intended to be absorbed.
⇑ Corresponding author. The application of a vibrating plate to absorb low-frequency
E-mail address: jparenas@uach.cl (J.P. Arenas). noise was presented by Wang et al. [4]. They employed this

http://dx.doi.org/10.1016/j.apacoust.2016.07.002
0003-682X/Ó 2016 Elsevier Ltd. All rights reserved.
 J.P. Arenas, F. Ugarte / Applied Acoustics 114 (2016) 10–17
physical principle to add a completely clamped plate with a sealed
air-back cavity to the sidewall of a duct to reduce the transmitted
sound. The analysis was developed for homogeneous and nonho-
mogeneous plates; better sound-absorption values were obtained
for the latter.
Fuchs [2] reviewed other configurations of sound absorbers,
such as panels completely filled with sound-absorbing material
and multilayer absorbers. Other authors have developed models
that included perforated and microperforated plates as vibrating
panels. These plates are an inexpensive and effective option to
absorb midfrequency noise. The perforations are reduced to sub-
millimeter sizes to release sufficient acoustic resistance values
and sufficiently low-mass to provide a relatively broadband sound
absorption. Thus, the use of microperforated panels increases the
effective frequency range while sacrificing sound absorption at
lower frequencies.
Lee and Swenson [5] analyzed low-frequency sound absorbers
both theoretically and experimentally using one or two square per-
forated panels with an air-back cavity. During the experimental
work they noticed the presence of absorption peaks at very low
frequencies that were associated with the panel vibration. In the
work of Frommhold et al. [6], a thin vibrating panel on a light hon-
eycomb structure was studied. This system formed small air vol-
umes, which acted as Helmholtz resonators. They used the
expression of Ford and McCormick [1] to include the effect of
mechanical vibration of the plate in the theoretical model to esti-
mate the sound absorption of the system. It was concluded that
the system had two resonances, one associated with the Helmholtz
resonator and the other with the stiffness of the plate.
The case of an infinite elastic panel with an air-back cavity has
been analyzed by Sakagami et al. [7,8] for an arbitrary angle of inci-
dence. Their results indicated that the contribution of the air cavity
is dominant and that the contribution to the sound absorption of
the panel’s damping loss factor is negligible. Obviously, the effect
of structural resonance of the panel was not included because this
effect appears only when the panel is finite and reflections occur at
the boundaries. Subsequently, Sakagami et al. [9] included the
effect of the vibration of the microperforated panel in a simplified
manner, considering the surface density of the plate for inclusion
into the impedance equations of the equivalent circuit. Thus, the
reactance of the panel’s mass was added in parallel to the impe-
dance of the microperforations. The results showed that a very
light panel reduced the amplitude of the absorption peak and
slightly increased the resonance frequency. However, for panels
having a surface mass greater than 2 kg/m2, the difference was
very small and the maximum value of absorption was almost the
same as in the case of a immobile panel.
Another study analyzed the case of a microperforated panel
with an air-back cavity [10], considering the effect of the structural
vibration of the panel. The theory involved a modal analysis of the
plate’s classical equation coupled with the acoustic wave equation.
They observed that if the frequency of the incident sound is less
than the first structural resonance, then the structural vibration
degrades the sound absorption and, conversely, a frequency
greater than the first resonance causes an improvement.
In later work, Lee et al. [11] also studied theoretically the case of
a microperforated panel absorber, considering the effects of the
microperforation, the resonance of the panel and the Helmholtz
resonator. The analysis included an equivalent circuit: Lee and col-
leagues considered that the impedances of the three effects are
connected in parallel and then connected in series to the air cavity.
They observed the presence of two characteristic peaks in the
curve of the sound-absorption coefficient of the system.
Tayong et al. [12] analyzed the microperforated system with an
air-back cavity in the case of very high sound pressure levels inci-
dent on the panel where the nonlinear behavior should be consid-
11
ered. Their results suggested the existence of an absorption
maximum that depends upon the flow velocity in the perforations,
which was experimentally confirmed. Interestingly, they used the
model of Ford and McCormick [1] to incorporate the effect of the
structural vibration of the panel, even though their samples were
circular and the boundary conditions were approximately
clamped.
Sakagami and Morimoto [13] studied the characteristics of a
sound-absorb-ing panel coupled with an air-back cavity, which
had a nonuniform depth. This case can be found in some particular
architectural applications. The theoretical results indicated that
the absorption peak can be wider than in the case of a cavity of uni-
form depth.
In general, a panel absorber is used to cover a relatively large
area in a room. Thus, characterization of its sound-absorbing prop-
erties can be made in a reverberation chamber. However, for
design purposes, measurements using an impedance tube are often
employed to obtain the normal-incident sound-absorption coeffi-
cient. Thus, small samples of panel absorbers having circular areas
are tested in the tube. Obviously, in such a case the shape geometry
and boundary conditions of the circular plate will be important for
determining both the eigenfrequencies and mode shapes of the
system and, consequently, the acoustic behavior of the sound
absorber.
It seems that the case of a sound absorber made of a circular
plate has been considered only by Hiraizumi et al. [14] in a much
forgotten paper. In their work, the problem is analyzed for a sys-
tem composed of a clamped circular plate, an air cavity and even-
tually a sound-absorbing material partially filling the cavity. The
theoretical analysis was based on solving the classical wave equa-
tion in cylindrical coordinates, considering each layer as an iso-
lated medium. Unlike Ford and McCormick, they did not use
electrical analogies in the derivation.
1.2. Statement of the problem
In the formulation presented by Hiraizumi et al. [14] they con-
sidered a plate with a clamped rim and they did not consider the
general case of an elastic boundary condition. Thus, the main
objective of this note is to present a sound absorption formula
for a circular panel sound absorber that includes the effect of the
panel’ structural vibration when subjected to a general elastic
boundary condition. The absorber is composed of a circular plate
and backing air layer, which is partially filled with a porous
sound-absorbing material. Then, when the system is excited by air-
borne sound, the flexural oscillation of the plate on the spring
formed by the air cavity and the sound attenuation inside the por-
ous material are the main sound-absorption mechanisms. The next
section presents the derivation of the theoretical equations. The
following sections describe the experimental setup and the results.
Concluding remarks are presented in the final section.
2. Theory
The low-frequency sound absorber is represented in Fig. 1. A
plane sound wave of amplitude p0 coming from the top strikes
an impervious layer formed by a thin plate with radius a and thick-
ness h. This sound wave induces a vibration that is propagated
through the air cavity of thickness d1 and through a sound-
absorbing layer of thickness d2 , which is resting in front of a rigid
backing surface with ideal infinite acoustic impedance. Complex
exponential convention ejxt has been used to represent harmonic
time dependence for all the theory presented in this paper.
Because the study focuses on the problem of a circular sample
placed inside of an impedance tube, the analysis has to be done
12 J.P. Arenas, F. Ugarte / Applied Acoustics 114 (2016) 10–17

Ra
p0 X /i ðrÞ 0 /i ðrÞr dr
WðrÞ ¼   : ð7Þ
D i k4  k4 R a /2 ðrÞr dr
i p 0 i

The spatial-average displacement amplitude per mode is found

by averaging on the surface area of the plate, using the definition
given by Ford and McCormick [1] and Lee et al. [10] as
ZZ R a 2
1 2p0 /i ðrÞr dr
hWii ¼ W i ðrÞdS ¼  0 
R : ð8Þ
S S Da2 k4i  k4p a /2i ðrÞr dr
0

Inserting Eq. (2) into Eq. (8) and solving the integrals, it follows
that

Fig. 1. The circular panel absorber with an elastic boundary condition. 2p0 ½J 1 ðki Þ þ C i I1 ðki Þ2
hWii ¼ ; ð9Þ
qS li ðx2i  x2 Þ
where ki ¼ ki a is the i-th eigenvalue, x2i ¼ Dki =qS ; x2 ¼ Dkp =qS ,
4 4
for a vibrating circular plate with an elastic edge support. The
attenuation of sound waves due to the propagation of energy and
through the air cavity and the porous material is discussed in
k2i n 2 o
Section 2.3. li ¼ J 0 ðki Þ þ J21 ðki Þ þ C 2i ½I20 ðki Þ  I21 ðki Þ
2
2.1. Average modal plate impedance þ ki C i ½J 0 ðki ÞI1 ðki Þ þ J 1 ðki ÞI0 ðki Þ: ð10Þ
The average velocity amplitude for harmonic excitation can be
Since the plate is excited by the action of a uniformly dis- pffiffiffiffiffiffiffi
expressed as hui ¼ jxhWi, where j ¼ 1. Since the specific acous-
tributed pressure only axisymmetric modes need to be considered.
tic impedance of the absorber at mode i-th is Z pi ¼ p0 =huii , then
In classical plate theory, the equation that describes the axisym-
 x 2 
metric vibrations of a thin circular uniform plate subjected to a dis- qS li x i
Z pi ¼ j 1 : ð11Þ
tributed harmonic force f ðrÞ (referring to the unit area), of 2½J 1 ðki Þ þ C i I1 ðki Þ2 x
frequency x, is [15]
Now, defining Bi ¼ k4i Ai , where
r4 WðrÞ  k4p WðrÞ ¼ f ðrÞ=D ¼ FðrÞ; ð1Þ
li
Ai ¼ ; ð12Þ
where WðrÞ is the transverse displacement of the plate, kp is the 2½J 1 ðki Þ þ C i I1 ðki Þ2
structural wavenumber (kp ¼ qS x2 =D), D ¼ Eh =12ð1  m2 Þ is the
4 3
and including the hysteretic damping of the plate as D ¼ Dð1 þ jgi Þ,
plate’s bending stiffness, and qS ; E, and m are the plate’s mass per where gi is the damping loss factor of the plate at mode i-th, Eq. (11)
unit area, Young’s modulus, and Poisson’s ratio, respectively. is written as
The i-th eigenfunction is given by [15]
Bi Dgi Bi D
/i ðrÞ ¼ Ui ½J 0 ðki rÞ þ C i I0 ðki rÞ; ð2Þ Z pi ¼ þ jqS Ai x  ; ð13Þ
xa4 xa4
where Ui and C i are real constants, J 0 is the Bessel function of zero which has the same form as the equation presented for a square
order of the first kind, I0 is the hyperbolic or modified Bessel func- plate in the paper by Ford and McCormick [1].
tion of zero order of the first kind, and ki is related to the angular
frequency of mode i as ki ¼ qS x2i =D. 2.2. Elastic boundary condition
4

Now, WðrÞ and FðrÞ in Eq. (1) can be expanded into the series
[16] In practice, it is well-known that the plate edge becomes some-
X X thing between a classical simply supported edge and a clamped
WðrÞ ¼ ci /i ðrÞ and FðrÞ ¼ i /i ðrÞ; ð3Þ edge. Then, the edge conditions of the circular plate can be formu-
i i lated at r ¼ a by using rotational and translational springs. If the
where edge of the circular plate is supported on rotational and transla-
Z Z tional springs all around as shown in Fig. 1, the boundary condi-
a a
2 2 tions at r ¼ a can be stated as follows [15]
ci ¼ WðrÞ/i ðrÞr dr and i ¼ FðrÞ/i ðrÞr dr; ð4Þ
va2 0 va2 0 " !#
@2W 1 @W 1 @ 2 W @W
and D þm þ 2 ¼ K R ; ð14Þ
@r 2 r @r r @h2 @r
ZZ
1
v¼ /2i r dr dh; ð5Þ and
S S
" !#
where S ¼ pa2 ; 0 6 r 6 a and 0 6 h 6 2p. @ 1  m @ 1 @ 2 W 1 @W
D ðr2 WÞ þ  ¼ K T W; ð15Þ
Multiplying Eq. (1) by /i , integrating the resulting equation as @r r @h r @r@h r 2 @h
Ra
ðÞr dr, and using Eqs. (3) and (4) we obtain
0 where K R is the distributed stiffness, i.e. moment/unit length,
Z
2 X /i ðrÞ a opposing the edge rotation, and where K T is the distributed stiff-
WðrÞ ¼ 2 FðrÞ/i ðrÞrdr: ð6Þ
va i k4i  k4p 0
ness, i.e. force/unit length, opposing the translational motion in
direction W. It is convenient to define the boundary conditions
Replacing FðrÞ ¼ f ðrÞ=D into Eq. (6), considering that the dis- through the dimensionless constants T ¼ a3 K T =D and R ¼ aK R =D.
tributed force per unit area is the constant sound pressure p0 on Now, inserting Eq. (2) with WðrÞ ¼ /ðrÞ into Eqs. (14) and (15),
the plate’ surface, and using Eq. (5), Eq. (6) is written as we derive the following equations
 J.P. Arenas, F. Ugarte / Applied Acoustics 114 (2016) 10–17 13

k3i J 1 ðki Þ þ TJ0 ðki Þ ki J 0 ðki Þ þ ðR þ m  1ÞJ 1 ðki Þ

Ci ¼ ¼ ; ð16Þ
ð1  TÞI0 ðki Þ þ k3i I1 ðki Þ ki I0 ðki Þ þ ðR þ m  1ÞI1 ðki Þ

which is a particular case of the frequency equation derived by

Rdzanek et al. [17]. Eq. (16) can be numerically solved to obtain
the values of ki and the corresponding values of C i .
The values for constants Ai and Bi were determined from the
corresponding values for ki and C i computed using the MATLAB
routine developed by Zagrai and Donskoy [18] for a circular plate
with elastic supports. The values for i ¼ 1; 2 and 3 are given in
Table 1.

2.3. Effect of the air cavity and infill material

The effects of the air layer and the presence of a porous absorb-
Fig. 2. An electro-acoustical equivalent circuit for a low-frequency panel absorber
ing layer in the cavity are considered using the impedance transla- made of a flexible plate mounted over an air space that is partly filled with a porous
tion theorem [19]. Thus, the total impedance of the absorber is sound-absorbing material.
P 1
Z ¼ Z p þ Z CF , where Z 1
p ¼ Z pi and Z CF is the impedance due to
the cavity-foam combination given by
3. Experiment
Z 20
þ Z1 Z2
Z CF ¼ ; ð17Þ 3.1. Methodology
Z1 þ Z2
An experiment was performed for testing the theory presented
where
above. A rigid impedance tube of inner diameter 100 mm was
adapted to implement a circular plate sound absorber at one end
Z 1 ¼ jZ 0 cotðkd1 Þ; ð18Þ
and a loudspeaker at the opposite end (see Fig. 3). The standard-
ized two-microphone technique was used to measure the
Z 2 ¼ jZ c cotðCd2 Þ; ð19Þ normal-incidence sound-absorption coefficient of the absorber
[20]. The loudspeaker was driven with white noise signal that var-
Z 0 is the characteristic impedance of the air (q0 c), q0 is the density ied from 0 to 2 kHz. The signals were generated and processed by a
of air, c is the speed of sound, k is the free field wave number (x=c), two-channel real-time fast Fourier transform analyzer, which was
Z c is the characteristic impedance of the foam, and C is the propa- connected to a computer for data analysis. Because of the random
gation constant of the foam. Z c and C can be determined by either characteristics of the noise signal, a total of 200 averages were
experimenting or using a model [19]. Eq. (19) is valid when consid- taken for each measurement. The transfer functions were mea-
ering that the air inside the porous medium can be replaced on the sured between the signals of two 1/4-in. condenser microphones
macroscopic scale by an equivalent fluid. at increments of 1 Hz. In light of the diameter of the standing wave
Therefore, the total acoustic behavior of the sound absorber can tube and the center-to-center spacing between two microphones
be represented by the electro-acoustic equivalent circuit shown in (102 mm), the measurements according to the ISO standard were
Fig. 2. assumed valid in the frequency range 169–1500 Hz. The distance
It should also be noted that, at very low frequencies (kd1  1), from the plate to the center of the nearest microphone used to
Z 1 can be replaced by jq0 c2 =xd1 . Therefore, for a plate-cavity measure sound absorption was 230 mm.
arrangement (without the foam), the frequencies for maximum Measurements were performed on two thin aluminum (Young’s
sound absorption are given at the eigenfrequencies, in Hz, modulus 70.8 GPa and Poisson ratio 0.33) plates used for the con-
struction of the absorbers. The nominal plate thicknesses were 0.4
rffiffiffiffiffiffiffiffisffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi and 0.8 mm with a mass density of 2626 and 2652 kg/m3, respec-
c q0 1 Dd Bi
fi ¼ þ ; ð20Þ
2p qS d Ai q0 c2 a4 Ai tively. In order to approach the condition of a clamped boundary,
the plate was compressed between two square steel frames mea-
suring 220 mm wide and 10 mm thick. Both frames had a circular
where d is the thickness of the air-space layer behind the plate. Eq.
perforation of 100 mm diameter at their center to form a
(20) has the same form as the one presented for a square plate in
restrained circular plate of radius 50 mm. The opposing faces of
the book by Fuchs [2]. It has to be noted that Eq. (20) is not valid
each frame were milled flat and straight. Bolts were used to assem-
for very large plates because very large plates do not vibrate in
ble the frame pieces together and to fasten the frame-plate combi-
the assumed bending mode [6].
nation to a flanged end of the impedance tube.
Finally, the normal-incidence sound-absorption coefficient can
A removable sample holder was made of transparent rigid poly-
be calculated from a ¼ 1  jRj2 , where R is the complex reflection carbonate tube 25 cm in length with a 100 mm inner diameter. The
coefficient obtained from the total surface impedance of the absor- sample holder had a moving steel piston sealed with an O-ring so
ber R ¼ ðZ  Z 0 Þ=ðZ þ Z 0 Þ.

Table 1
Constants occurring in calculating the modal plate’s impedance according to Eq. (13).

Edge A1 A2 A3 B1 B2 B3
C 1.8830 5.9137 12.4604 0:0197  104 0:9354  104 9:893  105
S-S 1.4126 8.2256 19.8807 0:0035  104 0:7282  104 1:0942  105
14
Fig. 3. Schematics of the experimental setup.
the length of the air cavity inside the holder could be varied
through a plunger.
The impedance tube was adapted to have a third reference 1/4-
in. microphone passing through the steel frame and located very
close to the plate surface to measure the sound pressure at the
plate surface (see Fig. 3). The sample holder was designed to have
independent flanges to connect to the steel frames so the holder
could be removed without modifying the boundary condition of
the plate. Care was taken to ensure a good sealing of all the ele-
ments of the experimental setup.
With the sample holder removed, the vibration response of the
plate to an incident broadband sound pressure was measured at its
center using a noncontact laser vibrometer to overcome possible
errors caused by mounting and mechanical effects in a common
accelerometer. Then, the modal damping loss was determined
using the 3-dB method at each resonance frequency [21]. In addi-
tion, the responses of the plate were used to obtain the measured
values of rotational and translational stiffness to characterize the
boundary condition. Exact values of K T and K R (i.e. T and R) were
indirectly determined by curve-fitting the equations presented in
Section 2 and the corresponding experimental vibration response
data.
A recycled cylindrical open-celled polyurethane sound-
absorbing material (see Table 2) was cut using a custom-made cir-
cular cutting tool to fit the diameter of the sample holder. Then the
material was carefully placed inside the cavity and a perfect fit
ensured. There was no particular reason to choose this material
as infill other than that recycled polyurethane foam was exten-
sively characterized in previous studies [22,23]. To compute the
2
Normalized impedance
−2
−4
−6
−8
Sound absorption coefficient
0.8
0.6
0.4
0.2
0
0 500
Fig. 4. Acoustic properties of the recycled polyurethane foam: (a) Normalized impedance, — resistance; - - - - reactance; (b) normal-incidence sound-absorption coefficient.
J.P. Arenas, F. Ugarte / Applied Acoustics 114 (2016) 10–17
Table 2
Recycled polyurethane foam characteristics.
Thickness Radius Mass density Flow resistivity Porosity
(mm) (mm) (kg/m3) (N s/m4)
48.5 50 66 7825 0.96
values of the characteristic impedance and the propagation con-
stant of the recycled foam, the semi-empirical approach presented
by del Rey et al. [22] was used (see Fig. 4).
Experiments were conducted for both plates; the air cavity
depth behind the plate (10, 20, and 50 mm) was varied by moving
the plunger along the longitudinal axis of the impedance tube. A
speed of sound c ¼ 343 m/s was assumed during all the
experiments.
3.2. Results
For the plate that measured 0.4 mm thick, the measured values
of damping loss factor were 0.051 and 0.0093 at the first and sec-
ond axisymmetric mode (243 Hz and 1187 Hz), respectively. For
the plate that measured 0.8 mm thick, only one resonance was
measured (619 Hz) because the second mode lay out of the fre-
quency range studied. In this case the measured damping loss
was 0.037.
As expected, the ideal clamped boundary condition was not
attained. For the plate that was 0.4 mm thick, the measured values
of T and R were 1  1050 and 1.4, respectively. In the case of the
plate that was 0.8 mm thick, the values of T and R were 1  1050
and 5.8, respectively. These values indicate that the translational
stiffness of the plate was high enough to approach a clamped edge
while the rotational stiffness was low, resulting in something
between a classical simply supported edge and a clamped edge.
Fig. 5 shows the comparison between experimental and theo-
retical results for the sound-absorption coefficient as a function
of frequency for the absorber made of the plate that was 0.4 mm
thick. The experimental results are similar in shape to those pre-
sented by other authors [1,10,11,14]. We noticed two maximum
values of absorption near 300 Hz and 1260 Hz. We also observed
that the theory predicted the experimental results very well. A
(a)
(b)
1000 1500 2000 2500 3000
Frequency, Hz
 J.P. Arenas, F. Ugarte / Applied Acoustics 114 (2016) 10–17 15

1 1
(a) (a)
Sound absorption coefficient

Sound absorption coefficient

0.8 0.8

0.6
0.6

0.4
0.4

0.2
0.2

0
200 400 600 800 1000 1200 1400
0
Frequency, Hz 200 400 600 800 1000 1200 1400
Frequency, Hz
1
(b)
Sound absorption coefficient

1
0.8 (b)

Sound absorption coefficient

0.8
0.6

0.6
0.4

0.4
0.2

0.2
0
200 400 600 800 1000 1200 1400
Frequency, Hz 0
1 200 400 600 800 1000 1200 1400
(c) Frequency, Hz
Sound absorption coefficient

0.8
1
(c)
Sound absorption coefficient

0.6
0.8

0.4
0.6

0.2
0.4

0
200 400 600 800 1000 1200 1400 0.2
Frequency, Hz

Fig. 5. Results of normal-incidence sound-absorption coefficient of a panel 0

200 400 600 800 1000 1200 1400
absorber made of an aluminum circular plate with a diameter of 100 mm and a
thickness of 0.4 mm, with a layer of recycled foam of a depth of 48.5 mm, for Frequency, Hz
different air cavity thicknesses: (a) 10 mm, (b) 20 mm, (c) 50 mm. — theoretical;
- - - - experimental. Fig. 6. Results of normal-incidence sound-absorption coefficient of a panel
absorber made of an aluminum circular plate with a diameter of 100 mm and
thickness of 0.8 mm, with a layer of recycled foam of a depth of 48.5 mm, for
different air cavity thicknesses: (a) 10 mm, (b) 20 mm, (c) 50 mm. — theoretical;
small peak near 1 kHz may have been caused by the presence of an
- - - - experimental.
asymmetric mode (a theoretical asymmetric mode of the circular
plate is at 1082 Hz) due to some imperfection in the experiment.
Fig. 6 shows the corresponding results for the sound absorber
Table 3
made of the plate that was 0.8 mm thick. As expected, doubling Frequencies of the circular panel absorber at which maximum sound absorption is
of the thickness of the plate increased the resonance frequencies T E
attained. f i is the theoretical value and f i is the measured value.
of the plate. Therefore, only one peak of sound absorption was T E
h (mm) i d1 (mm) f i (Hz) f i (Hz) Error (%)
observed at around 660 Hz, in the frequency range where plane
wave assumption is valid. There was good agreement between 0.4 1 10 312 309 0.97
the predicted and measured results. Figs. 5 and 6 show that the 20 306 305 0.33
50 292 294 0.68
sound-absorption peak at the first resonance is reduced in magni- 2 10 1260 1258 0.16
tude with the larger air gap width. The effect in the second reso- 20 1259 1260 0.08
nance is less evident. 50 1257 1264 0.55
We note that the sharpness of the lowest peak in the sound 0.8 1 10 663 665 0.30
absorption in Fig. 5 is wider than the sound absorption peak in 20 662 664 0.30
Fig. 6 because the sharpness increases with an increase in the 50 657 661 0.61
16 J.P. Arenas, F. Ugarte / Applied Acoustics 114 (2016) 10–17

1 occurs. This may be because the theoretical models have always

d =20 mm
0.9 1 assumed a perfectly clamped boundary condition, which in prac-
Sound absorption coefficient

0.8 tice is difficult to obtain. Here, because of the more general elastic
boundary condition, equations were derived to describe the
0.7
normal-incidence sound absorption of a circular panel absorber.
0.6
Therefore, the effect of the flexural vibration of the plate on the
0.5 sound-absorbing properties can be taken into account. Experimen-
0.4
d1=50 mm d1=10 mm tal values of the sound-absorption coefficient on a circular alu-
0.3 minum panel absorber with recycled foam partly filling an air
0.2 cavity behind the plate agreed very well with the theoretical pre-
0.1
dictions. The theory may be used to design a unit of low-
frequency absorption for conditioning rooms. An advantage of this
0
500 550 600 650 700 750 800 850 900 950 system is that if the plate is sufficiently stiff it can have more than
Frequency, Hz one resonant frequency and consequently more than one absorp-
tion band. The theory may also be used for tuning the absorber
Fig. 7. Results of normal-incidence sound-absorption coefficient of a panel in order to improve its performance. Further work is planned to
absorber made of an aluminum circular plate with a diameter of 100 mm and
include the effect of mechanical losses at the edges of the plate
thickness of 0.8 mm, with and without sound-absorbing material, for different air
cavity thicknesses d1 . — with foam; - - - - without foam. by writing the rotational and translational stiffness as complex val-
ues which may be useful to model the fixation of the panel onto the
cavity by an elastic adhesive.
Table 4
Frequencies of a panel absorber made of an aluminum circular plate with a diameter
of 100 mm and thickness of 0.8 mm, without sound-absorbing material in the cavity, Acknowledgments
T E
for different air cavity thicknesses d1 . f i is the theoretical value and f i is the measured
value. The work presented in this paper was supported by Conicyt-
i d1 (mm) T E Error (%) Fondecyt under grant No 1110605.
fi (Hz) fi (Hz)

1 10 738 751 1.7

20 693
50 668
thickness of panel’s material. This fact was also observed by Hirai-
zumi et al. [14].
The values of frequencies at which maximum sound absorp-
tions were obtained are presented in Table 3. The errors in the pre-
diction of these frequencies were less than 1% in all cases.
Increasing the depth of the air cavity produced a slight decrease
of the fundamental resonance of the panel absorber. The effect
was less significant at the second resonance frequency. It is
observed that the magnitudes and frequencies of the surface
acoustic impedance of the panel absorber and corresponding
sound-absorption coefficient peaks may be significantly affected
by the rigidity of the clamping condition. Further experiments
showed that increasing the thickness of the porous material did
not significantly affect the results of sound absorption, at least
for the size restrictions of the sample holder.
Fig. 7 shows the results of normal-incidence sound-absorption
coefficient of the panel absorber made of the plate that was
0.8 mm thick without the foam in the cavity. The values of fre-
quencies at which maximum sound absorptions were obtained in
this case are presented in Table 4. The errors in the prediction of
these frequencies were less than 2% in all cases. As observed by
Hiraizumi et al. [14], we notice that the magnitudes and frequen-
cies of the sound-absorption coefficient peaks are influenced by
the sound absorbing layer. The presence of the foam reduces the
value of the frequency at which maximum sound absorption is
attained. In addition, the porous material increases the maximum
value of sound absorption coefficient and it also broadens the res-
onant peak. These effects are more noticeable for large air cavity
thicknesses.
4. Concluding remarks
It has been observed in previous work on panel absorbers that
the theory does not exactly predict some of the experimental val-
ues of the frequencies at which maximum sound absorption
700 1.0 References
671 0.61
[1] Ford RD, McCormick MA. Panel sound absorbers. J Sound Vib 1969;10:411–23.
http://dx.doi.org/10.1016/0022-460X(69)90219-3.
[2] Fuchs H. Applied acoustics: concepts, absorbers, and silencers for acoustical
comfort and noise control. Berlin: Springer; 2013. http://dx.doi.org/10.1007/
978-3-642-29367-2.
[3] Fasold W, Veres E. Schallschutz + Raumakustik in der Praxis. Berlin: Bauwesen;
2003.
[4] Wang C, Han J, Huang L. Optimization of a clamped plate silencer. J Acoust Soc
Am 2007;121:949–60. http://dx.doi.org/10.1121/1.2427126.
[5] Lee J, Swenson GW. Compact sound absorbers for low frequencies. Noise
Control Eng J 1992;38:109–17. http://dx.doi.org/10.3397/1.2827811.
[6] Frommhold W, Fuchs HV, Sheng S. Acoustic performance of membrane
absorbers. J Sound Vib 1994;170:621–36. http://dx.doi.org/10.1006/
jsvi.1994.1091.
[7] Sakagami K, Takahashi D, Gen H, Morimoto M. Acoustic properties of an
infinite elastic plate with a back cavity. Acustica 1993;78:288–95.
[8] Sakagami K, Kiyama M, Morimoto M, Takahashi D. Sound absorption of a
cavity-backed membrane: a step towards design method for membrane-type
absorbers. Appl Acoust 1996;49:237–47. http://dx.doi.org/10.1016/S0003-
682X(96)00025-4.
[9] Sakagami K, Morimoto M, Yairi M. A note on the effect of vibration of a
microperforated panel on its sound absorption characteristics. Acoust Sci Tech
2005;26:204–7. http://dx.doi.org/10.1250/ast.26.204.
[10] Lee YY, Lee EWM, Ng CF. Sound absorption of a finite flexible micro-perforated
panel backed by an air cavity. J Sound Vib 2005;287:227–43. http://dx.doi.org/
10.1016/j.jsv.2004.11.024.
[11] Lee YY, Sun HY, Guo X. Effects of the panel and Helmholtz resonances on a
micro-perforated absorber. Int J Appl Math Mech 2005;4:49–54.
[12] Tayong R, Dupont T, Leclaire P. On the variations of acoustic absorption peak
with particle velocity in micro-perforated panels at high level of excitation. J
Acoust Soc Am 2010;127:2875–82. http://dx.doi.org/10.1121/1.3372714.
[13] Sakagami K, Morimoto M. Sound absorption characteristics of a panel abrober
with an air-backed cavity of non-uniform depth. Memoirs Grad Sch Sci Tech
Kobe Univ 2007;25-A:25–31.
[14] Hiraizumi M, Takahashi K, Sone T, Nimura T. Study on the sound absorption by
panel vibration—analysis in the model of circular panel with clamped
boundary. J Acoust Soc Jpn 1974;30:276–84.
[15] Rao SS. Vibration of continuous systems. New York: John Wiley and Sons;
2007.
[16] Malecki I. Physical foundations of technical acoustics. New York: Pergamon
Press; 1969.
[17] Rdzanek WP, Rdzanek WJ, Engel Z, Szemela K. The modal low frequency noise
of an elastically supported circular plate. Int J Occ Saf Erg 2007;13:147–57.
http://dx.doi.org/10.1080/10803548.2007.11076718.
[18] Zagrai A, Donskoy D. A soft table for the natural frequencies and modal
parameters of uniform circular plates with elastic edge support. J Sound Vib
2005;287:343–51. http://dx.doi.org/10.1016/j.jsv.2005.01.021.
[19] Allard JF, Atalla N. Propagation of sound in porous media: modelling sound
absorbing materials. 2nd ed. Chichester: John Wiley and Sons; 2009.
 J.P. Arenas, F. Ugarte / Applied Acoustics 114 (2016) 10–17 17

[20] ISO 10534-2. Acoustics – determination of sound absorption coefficient and [22] del Rey R, Alba J, Arenas JP, Sanchis VJ. An empirical modelling of porous sound
impedance in impedances tubes – Part 2: Transfer-function absorbing materials made of recycled foam. Appl Acoust 2012;73:604–9.
method. Geneva: International Organization for Standardization; 1998. http://dx.doi.org/10.1016/j.apacoust.2011.12.009.
[21] ASTM E 756. Standard test method for measuring vibration-damping [23] Arenas JP, Darmendrail L. Measuring sound absorption properties of porous
properties of materials. West Conshohocken: ASTM International; 2010. materials using a calibrated volume velocity source. Meas Sci Technol
http://dx.doi.org/10.1520/E0756-05R10. 2013;24:105005. http://dx.doi.org/10.1088/0957-0233/24/10/105005.