See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/309548320

Frequency dependent mechanical modelling of poro-elastic materials for

sound transmission loss simulations

Article in Noise Control Engineering Journal · September 2016

DOI: 10.3397/1/376407

CITATIONS READS

4 391

2 authors:

Paolo Bonfiglio Francesco Pompoli

Materiacustica srl University of Ferrara
103 PUBLICATIONS 860 CITATIONS 79 PUBLICATIONS 1,037 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Paolo Bonfiglio on 30 October 2016.

The user has requested enhancement of the downloaded file.

Frequency-dependent mechanical modelling of poro-elastic materials
for sound transmission loss simulations
Paolo Bonfiglioa) and Francesco Pompolib)
(Received: 4 April 2016; Revised: 9 September 2016; Accepted: 9 September 2016)
Multilayer panels made of sheet steel, a poro-elastic layer and an impervious
mass are widely used in noise control for increasing sound insulation. A complete
description of the vibro-acoustical behavior of the poro-elastic material used in
these panels requires knowledge of mechanical parameters (i.e. complex modu-
lus and Poisson's ratio) in order to model the wave propagation through the elas-
tic structure constituting its skeleton. Although many of the materials used for
noise and vibration control have viscoelastic behavior, prediction models cur-
rently make use of the linear elasticity theory. As a consequence, acoustical quan-
tities, such as sound transmission loss, are predicted by using static or quasi-static
values of mechanical properties. In this research, an experimental time domain
approach in combination with an analytical model for linear viscoelasticity is uti-
lized for determining the frequency-dependent complex modulus and increas-
ing the accuracy of sound transmission loss simulations in practical applications.
© 2016 Institute of Noise Control Engineering.

Primary subject classification: 51.3; Secondary subject classification: 35

1 INTRODUCTION the reliability of this method once the input parameters

are correctly determined6.
Nowadays most of the activity of companies supply-
Within the context of PPP systems, the crucial issue
ing porous materials in automotive (but more generally
to be addressed is the determination of all solid and
in transportation) is based on numerical simulations in
fluid phase parameters describing wave propagation in
order to verify targets in terms of some acoustical indi-
poroelastic media using the well-established Biot's the-
cators. For example, the sound transmission loss of multi-
ory. Although the use of experimental measurement of
layer panels made of sheet steel, a poro-elastic core layer
fluid phase parameters (airflow resistivity, open porosity,
and an additional mass (hereafter indicated as “plate-
tortuosity, characteristic lengths) has been proven to be
porous-plate” PPP systems) is a widely system used to
accurate enough for most of the porous media used in
increase interior comfort as regards noise generated in the
several real applications6, considerable work still needs
engine compartment.
to be done in order to include all possible structural ef-
In the literature several researches have been pro-
fects, such as for example a relevant frequency depen-
posed to analyze the acoustical behavior of flat sand-
dency of the mechanical properties of core poro-elastic
wich structures with porous core1–3, also including the
media. From experimental tests and numerical simula-
effect of an exciting mean flow4 and multilayer panels
tions, it can be easily observed that the use of low fre-
having cylindrical symmetry5.
quency values of the complex modulus is not accurate
As outlined in the aforementioned references, sev-
enough to predict the performance of trimmed panels
eral numerical techniques can be used for determining
in sound transmission loss applications. Taking this as
the performance of such kind of trimmed panels. Among
a starting point, it is clear that a key point for robust
these, the use of the transfer matrix method allows a sim-
numerical simulations will require the determination of
ple implementation of the sound propagation of sound
mechanical properties (at least the real and imaginary
waves through each layer and several studies have proven
part of the complex modulus) in the entire frequency
range of interest (generally between 100 and 5000 Hz).
The experimental determination of the complex modu-
a)
Dipartimento di Ingegneria, Università degli Studi di Ferrara, lus of solids as a function of frequency can be performed
Via Saragat 1, 44122 Ferrara, ITALY; email: [email protected]. using several methods which may be classified as static,
b)
Dipartimento di Ingegneria, Università degli Studi di Ferrara, quasi-static and dynamic. They generally consist of mea-
Via Saragat 1, 44122 Ferrara, ITALY. surements made in a narrow range of frequencies (possibly

Noise Control Engr. J. 64 (5), September-October 2016 Published by INCE/USA in conjunction with KSNVE 627
performed at different temperatures). The choice of the relevant model constitutive equation with fractional
appropriate measurement technique is influenced by the order derivatives, the general form of the constitutive
geometry, the damping factor and the frequency range equation for the conventional viscoelastic models can
of interest. An extensive review of existing methods for be expressed as:
determining mechanical properties of materials used in
noise and vibration control has been presented by Jaouen d b1 d b2 d bn
sðt Þþ b1 sðt Þ þ b2 sð t Þ þ þ bn sðt ÞsðtÞ
et al.7. In this paper, different quasi-static8–10 and dy-
. . .
dt b1 dt b2 dt bn
namic11–18 techniques are discussed and compared. A d a1 d a2 d am
recent proposal has been suggested using a method to ¼ a0 eðtÞ þ a1 e ðt Þ þ a2 eð t Þ þ . . . þ am eðtÞ;
dt a1 dt a2 dt am
determine the values of the storage and loss moduli
(hereafter indicated as E1 [Pa] and E2 [Pa]) in a wide ð2Þ
frequency range (generally between 100 and 1500 Hz)
measuring the time domain accelerations and using a where 0 < a1 < a2 < . . . < am < 1 and 0 < b1 <
transfer matrix approach19 for wave propagations in b2 < . . . < bn < 1 are material constants. Starting from
linear elastic solids. Eqn. (2) Pritz20 suggested a five parameter model assum-
The aim of this research is to provide more reliable ing that all parameters are zero with the exception of a0,
simulations of multilayer systems (using a transfer ma- a1, a2, b1, a1, a2 and b1. Moreover in analogy with the
trix approach) where the mechanical parameters of the fractional Zener model, the author assumes a1 = b1 = b
poroelastic material are determined by the combined use and a2 = a , with a > b. With such assumptions and set-
of experimental data from the aforementioned time do- ting b1 = tb, a0 = G0, a1 = G1 tb and a2 = (G1 G0) ta, it
main method and a well-established standard linear follows that Eqn. (2) can be written as:
model of viscoelasticity20.
The paper is organized as follows. A description of db b d
b
sðt Þ þ tb sð t Þ ¼ G0 e ðt Þ þ G 0 t eðt Þ
the theoretical background and experimental set-up is dtb dtb
given in Secs. 2 and 3. Section 4 will then report the da
results for several tested configurations in terms of sound þ ðG1 G0 Þta eðt Þ: ð3Þ
dta
transmission loss of multilayer systems with a core with
a frequency-dependent complex modulus. Concluding In the previous equation, G0 is the static modulus of elas-
remarks will be made in the last section. ticity, G1 is related to the high-frequency limit value of
dynamic modulus and t is the relaxation time.
2 THEORETICAL BACKGROUND The complex modulus of the model can be derived
by transforming Eqn. (3) into the frequency domain since:
Consider the function e(t) as representative of some
cause (i.e. strain in solids) acting on a material and the  k
d

function s(t) the effect (i.e. the stress) resulting from this F eðt Þ ¼ ð joÞk F feðt Þg; ð4Þ
dt k
cause. Fung21 and Boltzman22 proposed that a variation
in strain, occurring at a certain instant of time, would where F denotes the Fourier transform operator. The deri-
produce an effect for a certain time later. According to vation results in the following expression for the complex
classic theory proposed by Volterra and Zachmanoglou23, shear modulus in frequency domain:
it follows that stress–strain deformation can be expressed
ð jotÞa
 
in terms of a relaxation function: ∗ 1
G ð joÞ ¼ G0 þ G1 : ð5Þ
G0 1 þ ð jotÞb
Zt
deðtÞ
sðt Þ ¼ C ðt tÞ dt: ð1Þ The complex modulus is derived once the Poisson's ratio n
dt
1 is known:
The function C(t), also known as memory function, is a E ∗ ð joÞ ¼ G∗ ð joÞ
2ð1 þ nÞ: ð6Þ
property of the material and is a function of the time de-
lay (t t) between cause and effect within the material. From a practical point of view, one can solve visco-
In particular C(t) is defined for (t t) ≥ 0 and is fixed at elasticity problems as elasticity problems with a complex
zero elsewhere. modulus that depends on frequency. This property is
Within this contest, the fractional derivative models known as the elastic/viscoelastic equivalence principle24.
have been proven to be efficient in describing the dy- Finally, by finding parameters G0, G1, t, a and b, it
namic behavior of real materials20. In fact replacing the is possible to calculate the frequency-dependent com-
integer order time derivatives of stress and strain in the plex modulus of a viscoelastic material.

628 Noise Control Engr. J. 64 (5), September-October 2016 Published by INCE/USA in conjunction with KSNVE
Table 1—Description of the tested materials. calculate in the frequency domain the complete set of
vibro-acoustical indicators (among them the top and bot-
Material Typology Thickness Density tom plate accelerations) from the transfer matrix model
(mm) (kg/m3) for a given frequency and a semi-infinite fluid termina-
M1 Polyurethane foam 10 100 tion. Once the Poisson's ratio is known in advance, by
M2 Polyurethane foam 10 25 minimizing the difference between measured signals and
calculating the time of flight between measured signals
and the envelope of the top mass accelerometer, it is pos-
sible to determine the complex modulus as function of
the frequency. The method here utilized has been proven
3 MATERIALS AND METHODS
to provide reliable results in the frequency range between
As a first stage of the research, the complex modu- 100 and 1500 Hz. In order to extend the frequency range
lus of two different poro-elastic media (described in of the measured complex modulus, experimental data
Table 1) was measured using the time domain approach was fitted using the five-parameter fractional model
described in Ref. 19. The entire measurement proce- described in Sec. 2. The fitting procedure is based on
dure can be summarized as follows. The tested material a bounded nonlinear best-fit scheme25 and was imple-
(having a cylindrical shape with a diameter of 45 mm mented in Matlab.
and here assumed to be homogeneous and isotropic) is The same materials were also tested using a quasi-
positioned between two plates (aluminum support plate, static procedure26 allowing for the determination of the
10 mm thick, and steel top plate, 1 mm thick), one of complex modulus in a limited range of frequency (gener-
which is excited by an electromagnetic shaker. Using a ally between 20 and 50 Hz).
sine burst at a given frequency as the excitation signal, Successively, the same poro-elastic materials were test-
the accelerometric response in the time domain is mea- ed when coupled to solid elastic layers. Experimental tests
sured at the top and bottom plates. From the tests, it is were carried out according to ISO 15186-1:2003 Stan-
possible to determine the time of flight between mea- dard27 in a laboratory for automotive applications. Rect-
sured signals and the envelope of the top mass accel- angular panels (size equal to 1.21  0.95 m2) were
erometer. Figure 1 depicts the measurement layout and mounted on a frame between two rooms (Fig. 2). A
the time domain response related to the accelerometers sound source (emitting a stationary white noise) was
at a given frequency. In order to determine the complex located within a highly reverberant room with a volume
modulus of a given material, the proposed methodology of 80 m3; the average (in time and space) sound pressure
requires the measurement layout to be simulated using a level Lp,s was measured using two rotating ½00 micro-
well-established transfer matrix approach6. In fact, ac- phones. The receiving room (with a volume of 74 m3)
cording to the formalism used in Ref. 6, it is possible to had highly absorbing lateral walls, ceiling and a reflective

Fig. 1—Description of the procedure.

Noise Control Engr. J. 64 (5), September-October 2016 Published by INCE/USA in conjunction with KSNVE 629
Fig. 2—Laboratory for sound transmission loss measurement (Adler Evo Company in
Villastellone, Turin). (a) Reverberant room for the measurement of the average sound
pressure level; (b) Receiving room for the measurement of the average sound intensity
level emitted from the multilayer system.

floor; inside this room and at a distance of 15 cm, the aver-

age sound intensity LI,r coming from the panel was mea- ½D2 ŠV2 ¼ F: ð8Þ
sured by a manual scanning procedure using a B&K [D2] being a square matrix obtained from a complete
3547 sound intensity probe. According to ISO Standard matrix D:
15186-1:2003 the sound transmission loss was calcu-
lated as: 2   3
If 1 Jf 1 ½T1 Š ½0Š ... ½0Š
TL ¼ Lp;s LI;r 6 ðdBÞ: ð7Þ 6 ½0Š ½I12 Š ½J12 Š½T2 Š ... ½0Š 7
6 7
D¼6 ½0Š ... ½I23 Š ½J23 Š½T
 3Š ½ 0Š 7;
The plate was fastened to a frame around the perim- 6   7
4 ½0Š ... ... I3f J3f 5
eter using a series of clips positioned 30 cm apart. The
½0Š ... ... ... 1 r0 c0
components were glued to each other using an adhesive
layer. Tests were carried out in the frequency range be- ð9Þ
tween 100 and 5000 Hz.
Finally, the transfer matrix method6 was used for the where its second column is eliminated and F repre-
numerical simulations of multilayer panels and correc- sents the vector obtained by multiplying the eliminated
tion was applied to take into account the finite size of second column by 1. In Eqn. (9) r0 and c0 represent
the panel28. According to the formalism used in Ref. 6, the air density and the sound speed, respectively. In
it is possible to calculate in the frequency domain the Eqn. (9), Iij and Jij are known coupling matrices be-
complete set of vibro-acoustical indicators V2 (pres- tween different layers and their exact expressions can
sures, velocities, stresses at both sides of the multi- be found in Ref. 6, while matrices T1, T2 and T3 are
layer structure) from the transfer matrix model for a referred to the first solid plate, core material and last
given frequency and a semi-infinite fluid termination solid plate, respectively. Matrices T1 and T3 depend on
by solving the following expression: mechanical and geometrical properties of solid plates.

Table 2—Geometrical and mechanical properties of multilayer components. Apex * indicates quasi-static
values.
Material Thickness Density E (Pa) n( ) ( ) Airflow Open Tortuosity Viscous Thermal
(mm) (kg/m3) resistivity porosity ( ) characteristic characteristic
(Pa*s/m2) ( ) length (mm) length (mm)
Steel plate 1.8 7850 2e11 0.33 0.05 – – – – –
Rubber layer 1.8 1800 2240e3 0.48 0.1 – – – – –
M1 12.7 / 25 100 32,000* 0.33* 0.37* 19,000 0.92 1.44 56 322
M2 10 25 36,446* 0.33* 0.15* 16,900 0.93 1.2 50 207

630 Noise Control Engr. J. 64 (5), September-October 2016 Published by INCE/USA in conjunction with KSNVE
 250000
Table 3—Tested configurations. E1 Time domain
200000 E1 5 parameter Pritz
Multilayer Steel plate Porous material Rubber layer E2 Time domain
E2 5 parameter Pritz

E1-E2 [Pa]
150000 E1 QS
A 1.8 mm M1 (12.7 mm) 1.8 mm E2 QS
B 1.8 mm M1 (12.7 mm) – 100000
C 1.8 mm M1 (25.0 mm) 1.8 mm
50000
D 1.8 mm M2 (10.0 mm) 1.8 mm
0
1 10 100 1000 10000
Frequency [Hz]
The porous materials were considered as poro-elastic
media and T2 is a matrix depending on both fluid and Fig. 4—Frequency-dependent complex
solid phases. Regarding the fluid phase, experimental modulus for material M2.
measurements of airflow resistivity29, open porosity30
and tortuosity31,32 were carried out. Viscous and thermal
characteristic lengths were determined using an inverse
Finally in the same graphs, the quasi-static values
procedure starting from minimization of surface imped-
of the same quantities show the consistency of the
ance as per the Johnson–Champoux–Allard model33.
measurement method although an underestimation of
Once Eqns. (8) and (9) are solved at each frequency
both real and imaginary part of the complex modulus
and angle of incidence, the diffuse field sound trans-
is observed. The same trend was observed in the experi-
mission loss can easily calculated and compared with
mental tests discussed in Ref. 19.
experimental tests.
The frequency-dependent complex moduli obtained
Several multilayer panels (made of steel plate, poro-
from the fitting procedure for poroelastic materials were
elastic material and rubber layer) were tested and com-
used to compare the experimental sound transmission
pared in this research. Tables 2 and 3 summarize the
loss with numerical simulations on multilayer trimmed
physical and mechanical properties of each component
flat panels.
and tested configuration.
Figures 5 to 8 show a comparison for all the multi-
layer systems described in Table 3 in terms of sound
4 RESULTS AND DISCUSSION transmission loss between the experimental values and
Figures 3 and 4 show the curves of storage and loss the numerical simulations obtained by using quasi-static
moduli for all the tested materials measured by using values (Table 2) and dynamic values (Figs. 3 and 4) for
the methodology described in Sec. 3. In order to reduce the storage and loss moduli. Regarding Poisson's ratio,
the inhomogeneity effect, three samples of each material quasi-static values were used for simulations, while the
were tested and the average value for each frequency was mechanical properties of steel plate and rubber layer were
shown along with the standard deviation. chosen from the literature.
In addition the fitted curves from five-parameter From the curves, two interesting aspects can be high-
model were also shown. Table 4 summarizes the values lighted: (i) the quasi-static values of mechanical prop-
of the five parameters obtained from the fitting proce- erties are not adequate to correctly simulate the sound
dure for both materials. transmission loss and huge differences can be found both
at the resonance frequency (typical of a plate-porous-
plate system) and at a frequency above 500 Hz where
200000 the difference between measurement and simulation could
180000
E1 Time domain be up by 16–23 dB; (ii) the use of frequency-dependent
160000 E1 5 parameter Pritz
140000 E2 Time domain
E2 5 parameter Pritz
E1-E2 [Pa]

120000 E1 QS
100000 E2 QS
80000
60000
Table 4—Values of the five parameters from the
40000 fitting procedure between viscoelasticity
20000 model and experimental data.
0
1 10 100 1000 10000
Frequency [Hz]
Material G0 G1 a b t
M1 4747 14,371 0.61 0.43 0.94
Fig. 3—Frequency-dependent complex M2 5502 18,332 0.76 0.59 1.02
modulus for material M1.

Noise Control Engr. J. 64 (5), September-October 2016 Published by INCE/USA in conjunction with KSNVE 631
 100 100
90 Test 90 Test
80 Quasi-static value 80 Quasi-static value
70 70
Viscoelastic model Viscoelastic model
60 60

TL [dB]
TL [dB]

50 50
40 40
30 30
20 20
10 10
0 0

100
125
160
200
250
315
400
500
630
800
1000
1250
1600
2000
2500
3150
4000
5000
1000
1250
1600
2000
2500
3150
4000
5000
100
125
160
200
250
315
400
500
630
800
Frequency [Hz] Frequency [Hz]

Fig. 5—Comparison for multilayer system A Fig. 7—Comparison for multilayer system C
in terms of sound transmission loss in terms of sound transmission loss
between experimental values and between experimental values and
numerical simulations obtained by numerical simulations obtained by
using quasi-static values and dynamic using quasi-static values and dynamic
values for mechanical properties. values for mechanical properties.

mechanical properties enhances the results of the simula- 100

tions in the entire frequency range except for system B 90 Test
where the effect of viscoelasticity can be neglected since 80 Quasi-static value
70
the poroelastic material is not coupled with the addi- Viscoelastic model
60
TL [dB]

tional rubber layer. 50

Table 5 summarizes the average of the absolute differ- 40
ences as regards the experimental data for all the tested 30
configurations and all the input data utilized for the 20
10
simulations. From the data it is possible to observe the 0
average of absolute differences with respect to experi-
100
125
160
200
250
315
400
500
630
800
1000
1250
1600
2000
2500
3150
4000
5000
mental data lower than 3 dB when frequency-dependent Frequency [Hz]
complex moduli are used, while absolute differences
could go up 8 dB when quasi-static values are utilized. Fig. 8—Comparison for multilayer system D
in terms of sound transmission loss
between experimental values and
numerical simulations obtained by
100
90 Test using quasi-static values and dynamic
80 Quasi-static value values for mechanical properties.
70
Viscoelastic model
60
TL [dB]

50
40
30 Table 5—Average of the absolute differences in re-
20
10
spect of the experimental data for all the
0 tested configurations and all the input data
100
125
160
200
250
315
400
500
630
800
1000
1250
1600
2000
2500
3150
4000
5000

utilized for the simulations.

Frequency [Hz]
Multilayer Mean (TLexp TLqs) Mean (TLexp TLfd)
Fig. 6—Comparison for multilayer system B (dB) (dB)
in terms of sound transmission loss
A 8.1 2.3
between experimental values and
B 1.6 1.5
numerical simulations obtained by C 7.7 1.7
using quasi-static values and dynamic D 8.1 2.2
values for mechanical properties.

632 Noise Control Engr. J. 64 (5), September-October 2016 Published by INCE/USA in conjunction with KSNVE
5 CONCLUSIONS 14. M. Etchessahar, “Caractérisation mécanique en basses fré-
quences des matériaux acoustiques (Low frequency mechanical
In this paper, a combined experimental and analyti- characterization of acoustical materials)”, PhD dissertation,
Université du Maine, Le Mans, France, (2002).
cal approach was proposed for the determination of the 15. L. Jaouen, B. Brouard, N. Atalla and C. Langlois, “A simplified
frequency-dependent complex modulus of poro-elastic numerical model for a plate backed by a thin foam layer in the
materials used in noise control applications. Data ob- low frequency range”, J. Sound Vibr., 280, 681–698, (2005).
16. J.F. Allard, M. Henry, L. Boeckx, P. Leclaire and W. Lauriks,
tained by means of the proposed methodology was used “Acoustical measurement of the shear modulus for thin porous
in combination with an analytical model for linear visco- layers”, J. Acoust. Soc. Am., 117, 1737–1743, (2005).
elasticity to simulate (using a transfer matrix scheme) the 17. L. Boeckx, P. Leclaire, P. Khurana, C. Glorieux, W. Lauriks and
sound transmission loss of the multilayer trimmed panel J.F. Allard, “Guided elastic waves in porous materials satu-
rated by air under lamb conditions”, J. Appl. Phys., 97,
for automotive applications and they allowed for a signif- 094911-094911-8, (2005).
icant increase in the accuracy in the calculation when 18. L. Boeckx, P. Leclaire, P. Khurana, C. Glorieux, W. Lauriks and
compared with numerical results obtained using quasi- J.F. Allard, “Investigation of the phase velocities of guided
acoustic waves in soft porous layers”, J. Acoust. Soc. Am., 117,
static values for mechanical properties. 545–554, (2005).
19. P. Bonfiglio and F. Pompoli, “Determination of the dynamic
6 REFERENCES complex modulus of viscoelastic materials using a time domain
approach”, Polymer Testing, 48, 89–96, (2015).
1. J.S. Bolton, N.-M. Shiau and Y.J. Kang, “Sound transmission 20. T. Pritz, “Five-parameter fractional derivative model for poly-
through multi-panel structures lined with elastic porous materials”, meric damping materials”, J. Sound Vibr., 265, 935–952, (2003).
J. Sound Vibr., 191(3), 317–347, (1996). 21. Y.C. Fung, Foundations of Solid Mechanics, Prentice Hall, Inc.,
2. J.M.P. António, A. Tadeu and L. Godinho, “Analytical evalua- New Jersey, (1965).
tion of the acoustic insulation provided by double infinite walls”, 22. L. Boltzmann, “Zur theorie der elastische nachwirkung (On the
J. Sound Vibr., 263(1), 113–129, (2003) theory of elastic effects)”, Analen der Physik, 7, 624, (1876).
3. O. Tanneau, J.B. Casimir and P. Lamary, “Optimization of multi- 23. E. Volterra and E.C. Zachmanoglou, Dynamic of Vibrations,
layered panels with poroelastic components for an acoustical Charles E. Merril Books, Inc., Columbus, Ohio, (1965).
transmission objective”, J. Acoust. Soc. Am., 120(3), 1227–1238, 24. J. Salençon, Viscoélasticité (Viscoelasticity), Presse des Ponts et
(2006). Chaussés, Paris, (1983).
4. Y. Liu and A. Sebastian, “Effects of external and gap mean 25. J.C. Lagarias, J.A. Reeds, M.H. Wright and P.E. Wright, “Con-
flows on sound transmission through a double-wall sandwich vergence properties of the Nelder–Mead simplex method in low
panel”, J. Sound Vibr., 344, 399–415, (2015). dimensions”, SIAM Journal of Optimization, 9(1), 112–147,
5. Y. Liu and C. He, “Analytical modelling of acoustic transmis- (1998).
sion across double-wall sandwich shells: Effect of an air gap 26. P. Bonfiglio, F. Pompoli and P. Shravage, “Quasi-static evalua-
flow”, Compos. Struct., 136, 149–161, (2016). tion of mechanical properties of poroelastic materials: Static and
6. J.F. Allard and N. Atalla, Propagation of Sound in Porous Media, dynamic strain dependence and in vacuum tests”, Acoustics 08,
John Wiley & Sons, United Kingdom, (2009). (2008).
7. L. Jaouen, A. Renault and M. Deverge, “Elastic and damping 27. ISO 15186-1:2003, “Acoustic — Measurement of sound insu-
characterizations of acoustical porous materials: Available ex- lation in buildings and of building elements using sound inten-
perimental methods and applications to a melamine foam”, sity”, International Organization for Standardization, (2003).
Applied Acoustics, 69(12), 1129–1140, (2008). 28. P. Bonfiglio, F. Pompoli and R. Lionti, “A reduced-order inte-
8. E. Mariez, S. Sahraoui and J.F. Allard, “Elastic constants of poly- gral formulation to account for the finite size effect of isotropic
urethane foam's skeleton for the Biot model”, InterNoise96, square panels using the transfer matrix method”, J. Acoust. Soc.
(1996). Am., 139, 1773, (2016).
9. C. Langlois, R. Panneton and N. Atalla, “Polynomial relations 29. ISO 9053:1991, “Acoustics — Materials for acoustical applica-
for quasi-static mechanical characterization of isotropic poro- tions — Determination of airflow resistance”, International
elastic materials”, J. Acoust. Soc. Am., 110, 3032–3040, (2001). Organization for Standardization, (1991).
10. M. Etchessahar, L. Benyahia, S. Sahraoui and J.F. Tassin, “Fre- 30. Y. Champoux, M.R. Stinson and G.A. Daigle, “Air-based sys-
quency dependence of elastic properties of acoustics foams”, tem for the measurement of porosity”, J. Acoust. Soc. Am.,
J. Acoust. Soc. Am., 117, 1114–1121, (2005). 89, 910–916, (1991).
11. T. Pritz, “Dynamic Young's modulus and loss factor of plastic 31. J.F. Allard, B. Castagnède, M. Henry and W. Lauriks, “Evalua-
foams for impact sound isolation”, J. Sound Vibr., 178, 315–322, tion of the tortuosity in acoustic porous materials saturated by
(1994). air”, Review of Scientific Instruments, 65, 7654–7755, (1994).
12. T. Pritz, “Frequency dependences of complex moduli and com- 32. P. Bonfiglio and F. Pompoli, “Frequency dependent tortuosity
plex Poisson's ratio of real solid materials”, J. Sound Vibr., 214, measurement by means of ultrasonic tests”, ICSV14, (2007).
83–104, (1998). 33. P. Bonfiglio and F. Pompoli, “Inversion problems for determin-
13. ASTM E756:1998, “Standard test method for measuring vibra- ing physical parameters of porous materials: Overview and
tion damping properties of materials”, American Society for comparison between different methods”, Acta Acustica united
Testing and Materials, (1998). with Acustica, 99(3), 341–351, (2013).

Noise Control Engr. J. 64 (5), September-October 2016 Published by INCE/USA in conjunction with KSNVE 633

View publication stats