Applied Acoustics 115 (2017) 131138

Contents lists available at ScienceDirect

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

Predicting the sound absorption of natural materials: Best-fit inverse

laws for the acoustic impedance and the propagation constant
Umberto Berardi a,, Gino Iannace b
a
Department of Architectural Science, Ryerson University, 350 Victoria Street, Toronto, Ontario M5B 2K3, Canada
b
Dipartimento di Architettura e Disegno Industriale, Seconda Universita di Napoli, Italy

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

Article history: Natural materials are becoming a valid option for sound absorption treatments. In particular, among
Received 16 June 2016 them, natural fibers have received increasing attention given their good thermal insulation properties,
Received in revised form 7 August 2016 lack of harmful effects on health, and availability in large quantities. This paper discusses an inverse
Accepted 10 August 2016
method to predict the acoustical properties of nine natural fibers. Six vegetative fibers: kenaf, wood,
Available online 6 September 2016
hemp, coconut, straw, and cane; one animal fiber, sheep wool; recycled cardboard; and granular cork
are investigated. The absorption coefficient and the flow resistance for samples of different thickness
Keywords:
have been measured. Moving from the Delany-Bazley model, this study compares the impedance tube
Sustainable materials
Natural fibers
results with the theoretically predicted ones. Then, using a least-square fit procedure based on the
Airflow resistance Nelder-Mead method, the coefficients that best predict both the acoustic impedance and the propagation
Sound absorption constant laws are calculated. The inverse approach used in this paper allows to determine different phys-
Best-fit methods ical parameters and to obtain formulas to include the investigated natural fibers in software modelling
for room acoustics applications.
2016 Elsevier Ltd. All rights reserved.

1. Introduction
Sound absorption in room acoustics is generally obtained with
porous synthetic materials, such as rock wool, glass wool or polye-
ster, which are expensive to produce and have several environ-
mental implications. These issues have increased the attention
towards natural materials [13]. These are generally defined
according to the natural or renewable sources for their constituent
materials, the low level of environmental pollution emitted during
their production, or the low embodied energy [4,5]. In particular,
natural fibers have received increasing attention for acoustic uses
[68]. They are competitive materials thanks to their low density,
good mechanical properties, easy processing, high stability, mini-
mal health impacts, high quantity availability, low price, and
reduced environmental impacts for their production [4].
Based on their microscopic configurations, porous absorbing
materials have been classified as cellular, fibrous, and granular
[5]. In particular, fibrous materials consist of a series of tunnel-
like openings that are formed by interstices in material fibers. Fiber
materials are generally classified as natural or artificial. Natural
fibers can be vegetable (e.g. kenaf, hemp, and wood), animal (e.g.
wool and fur felt) or mineral (asbestos); whereas synthetic fibers
Corresponding author.
E-mail address: uberardi@ryerson.ca (U. Berardi).
can be mineral (e.g. fiberglass, mineral wool, and glass wool) or
polymer (polyester). Moreover, vegetable fibers are categorized
into: stalk or wood fiber (e.g. straw of wheat, rice, softwood or
hardwood); bast fiber or skin fiber (e.g. flax, jute, kenaf, industrial
hemp, ramie, rattan, and soybean); leaf fiber (e.g. sisal, palm, and
agave); seed fiber (e.g. cotton and kapok); and fruit fiber (e.g. coco-
nut). Microscopic analysis of natural fibers reveals that natural
fibers present irregular shape and are aggregated in structures
much less regular than synthetic porous materials [6]. The irregu-
lar structure is the main limit of theoretical models to predict the
sound absorption behavior of natural materials.
The materials studied in this paper are mainly vegetable fibers,
which were compressed in order to create sound absorbing sam-
ples. Although natural fibers are often commercialized in panels
or blocks created through the use of a binder, in this paper unpro-
cessed natural raw fibers are considered. Material stability was
obtained only by compression, without the use of any binder.
This paper follows some recent studies of the authors about
measurements of the sound absorption of natural fibers [6]. In par-
ticular, this paper elaborates the laboratory measurements to the
ability to model the acoustic behavior of the fibers. The scope is
to obtain formulas that could be used as predictive tools for the
acoustic behavior of natural fibers. In this way, the paper will pro-
vide a way to overcome the empirical approach generally required
while assessing inhomogeneous natural fibers.

http://dx.doi.org/10.1016/j.apacoust.2016.08.012
0003-682X/ 2016 Elsevier Ltd. All rights reserved.
132 U. Berardi, G. Iannace / Applied Acoustics 115 (2017) 131138
The paper is structured in the following way: Section 2 presents
a review of the models for the sound absorption of porous materi-
als; Section 3 presents the laboratory measurements; Section 4
briefly presents the natural fibers object of the investigation;
Section 5 reports the results of the best-fit approach used to fit
theoretical models to laboratory results; and Section 6 reports
some concluding remarks.
2. Sound absorption modelling of porous materials
Useful parameters to compare the sound-absorbing characteris-
tics of fibers are their diameter, their length, and the regularity of
their shape. Electronic microscopy techniques have shown that the
diameter of natural fibers tends to be larger than in synthetic fibers
and much more irregular [6]. Arenas and Crocker reported the
average diameter for some natural fibers: 833 lm for cotton,
21 mm for kenaf, 1638 mm for wood, and 22 lm for hemp [5].
Animal fibers may be even thicker, as the fiber diameter of the
wool equals to 37 mm when used as a raw material and 63 mm
when used as a wool batt6. Given the not constant diameter of
natural fiber, theoretical sound absorbing models fail from being
accurate when they are applied to natural fibers.
A porous material exposed to incidental sound waves allows the
air molecules within the pores to vibrate and, by doing it, to trans-
form some sound energy into thermal and viscous heat. At low fre-
quencies, these energy losses are isothermal and limited (resulting
in low sound absorption at low frequencies), while at high frequen-
cies, they are adiabatic and are generally more important. In
fibrous materials, most of the sound energy is absorbed by its
scattering among the fibers and by their vibration.
Several models for predicting the sound absorption mecha-
nisms of porous (synthetic) materials exist, and the one to use in
each case is generally selected based on the distinctive absorption
mechanism and the type of porosity of the analyzed material.
Existing sound absorbing models aim to describe the character-
istic wave impedance and the characteristic sound propagation
constant using some basic physical properties of the materials,
such as the porosity, the tortuosity, and the airflow resistance.
Models requiring more than five material properties have been
proposed over the years [6,9,10]. However, since the direct mea-
sure of material properties (such as the porosity or the tortuosity)
is often difficult to perform and inaccurate in the results, indirect
or inverse methods, where these parameters are derived from
the analysis of some material behaviors, are often preferred.
Delany and Bazley proposed a simple model for fibrous absor-
bent materials only employing the non-acoustical parameter of
the airflow resistivity for predicting the acoustical characteristics
and many other parameters [9]. This model considers a porous
layer as a bulk material with the rigid frame media of the material,
so that the flow resistivity is sufficient to determine the character-
istic wave impedance zc and the characteristic propagation con-
stant (or complex wave number) kc, according to the following
equations:



 
q f c2 q f c4
zc q 0 c 1 c 1 0  jc3 0 1 i
r r



 
q f c6 q f c8
kc x=c c5 0  jc7 0 2 eight coefficients ci. The optimization process was performed using
r r
where q0c is the characteristic impedance (Pa s/m), x the angular
frequency (x = 2pf), c the sound speed (m/s), r the flow resistivity
(Rayl/m), f the frequency (Hz), and ci (with i = 1. . .8) are eight
numerical coefficients. Using a best-fitting approach to a large
amount of experimental data of mineral fibrous porous absorbers,
Delany and Bazley defined the eight ci coefficients necessary to
define both the characteristic impedance and the propagation con-
stant. The final relationships proposed by Delany and Bazley are [9]:

0:623
0:66 !
qf qf
zc q0 c 1 0:078 0  j0:074 0 3
r r

0:7
0:595 !
qf qf
kc x=c 0:0987 0  j0:189 0 4
r r
Known the formulas (3) and (4), it is then possible to obtain
several acoustic and non-acoustic properties. For example, the
normal-incidence sound absorption coefficient is easily obtained
once the thickness and the flow resistivity of the material are
known.
Delany and Bazley formulas (3) and (4) were obtained over a
well-defined frequency range (10 < f/r < 1000) and with a porosity
of the material close to 1. In fact, at both lower and higher values of
f/r, DelanyBazley clarified that other power-law relations should
be expected. Moreover, the DelanyBazley ci coefficients reported
in formulas (3) and (4) were obtained from sound absorption mea-
surements of mineral fibers with a diameter between 1 and 10 lm.
Given these limits, it is not surprising that, although the simplicity
of the DelanyBazley model which is also incorporated in some
standards, this model has shown several limits when applied to
thicker fibers such as the natural ones [6,7,10].
Many authors, including Qunli, Miki, Mechel, Wang et al., have
provided other ci coefficients for the formulas (1) and (2) by
directly measuring the characteristics wave impedance and propa-
gation constants and subsequently applying curve-fitting proce-
dures [1012]. In 1990, Miki provided modified coefficients to
those proposed by Delany and Bazley for porous materials [13].
Dunn and Davern modified these coefficients for a foam material,
and similarly, Garai and Pompoli by experimenting polyester fibers
with diameter ranging from 20 to 50 lm, proposed other
coefficients that fit better for textile fibers than the coefficients
originally proposed by DelanyBazley [10]. However, since
measuring the propagation constants is a difficult and inaccurate
process, an alternative common way is to measure the normal
incidence sound absorption coefficient in the frequency domain
for a known flow resistivity material and then to determine the
eight coefficients ci using best-fit approaches [11].
A best-fit approach to fit experimental results to formulas (1)
and (2) requires to adopt an iterative numerical method to obtain
the eight coefficients that best describe the measured acoustic
behavior of the samples [14]. In this paper, the minimization of a
quadratic error function is performed. The error function is the
squared difference between the sound absorption coefficient
measured for a material sample at the i-th frequency (an,i), and
the corresponding estimated value ( a b n;i ) using Eqs. (1) and (2).
The minimization of the error function required to set to zero the
following error expression, so to obtain the values which best
approximate the ci coefficients:
@e XN
@ab
2 an;j  a
b n;j n;j 0 i 1; . . . 0:8 5
@Ai @A
i1
A Matlab computer code was implemented to minimize the
nonlinear Eq. (5) and to obtain the corresponding values of the
the Nelder-Mead simplex method [14]. The optimization method
whose result will be reported in Section 4 showed to be accurate
and quick, giving small errors of the prediction in just a few
seconds. Moreover, although different values were set as initial
input values of Eqs. (1) and (2), the optimization always converged
to the same ci coefficients values. Once the characteristic impe-
dance was found, and the formula (1) and (2) for each material
 U. Berardi, G. Iannace / Applied Acoustics 115 (2017) 131138 133

were expressed, it was possible to obtain the sound absorption [21]. Between the two procedures for measuring the airflow resis-
coefficient (a) of the material from the surface impedance (zs) tance, the continuous flow procedure and the alternate flow one,
using the formulas (6)(8): this last was preferred since it allows to perform measurements
over a higher number of cycles, averaging the results, and reducing
a 1  jRj2 6 the uncertainty of the measured value. Moreover, the alternate
method was considered more reliable in terms of the adiabatic
zs  q0 c compression of the air inside the volume since the equipment
R 7
zs q0 c was easier to close hermetically. As suggested in the ISO 9053 stan-
dard [21], ten measurements using the equipment shown in Fig. 1
zs jzc cotkc d 8 were performed for four different air velocities (0.5, 1, 2, and 4 m/s)
for each sample in order to determine the average airflow resis-
where R is the sound pressure reflection coefficient and d is the
tance and its standard deviation.
thickness of the sampled material.
Table 1 reports the values of the air flow resistivity together
For sake of completeness it is important to remember that
with their standard deviation for each of the selected materials.
beside to best fitting empirical studies regarding sound absorption
The measurements confirmed that the uncertainty associated with
measurements, in the last two decades, a significant literature of
the airflow resistivity measurement is generally high due to den-
theories and models has tried to understand the interaction of por-
sity change, presence of non-laminar air flow (also as a conse-
ous materials with sound fields. For example, Biot established an
quence of the use of different air movement speeds), low
easy theoretical explanation for saturated porous materials which
pressure differences on the two sides of the sample, and inaccuracy
is still considered enough accurate [15].
in thickness determination [2224].
While Delany and Bazley model is unable to predict the frame
The high standard deviation values of the airflow resistance
resonance of a porous material, rigid frame methods were intro-
measurements is common in the characterization of natural mate-
duced by Johnson and Allard to predict the acoustic characteristics
rial properties with direct methods. This was one of the reasons
of porous materials while considering the frame as motionless
that justified the adoption in this paper of an inverse method based
[16,17]. In particular, in order to improve the explanation of sound
on the lowest possible number of material properties.
propagation in porous materials, Allard [17] described a porous
layer as a mixture of air and an elastic frame, which could be mod-
elled through some non-acoustical parameters, such as the bulk 4. Natural fibers
density, the fiber density, the Prandtl number of the air movement,
the thermal characteristic length, and some elastic coefficients. The For sake of brevity, the present section briefly describes the nat-
Johnson-Champoux-Allard model that has resulted from the previ- ural fibers investigated in this paper, and shown in Fig. 2. A more
ous interpretation of the frame nature is nowadays used to inter- complete description of the tested materials and of the fiber pro-
pret the behavior of porous materials and has proved particularly cessing can be found in [6]. Table 1 reports the sound absorption
good in the low frequency region. Although previously described coefficients in one-third octave for each fiber.
models have been recently revised to consider the existence of
materials with different pore sizes [18], their validity is limited
for the interpretation of the behavior of natural materials, which
have highly irregular fibers, pores, and structures.

3. Acoustic measurements

Sound absorption measurements were performed using the

impedance tube method according to the ISO 10534-2 standard
[19]. The Kundts tube had an internal diameter of 10 cm, a length
of 56 cm, and mounted two 00 microphones, at a distance of 5 cm
for measurements above 250 Hz and of 10 cm for measurements
below 250 Hz. In fact, the lower frequency of validity for a Kundts
tube should respect the condition S > 0.05 c/f, where S is spacing
between the microphones, c is speed of sound, and f the lower fre-
quency; this means that a 10 cm distance between the micro-
phones allows accurate data below 250 Hz. Phase mismatch was
checked by inverting the microphone positions. To reduce the pos-
sible gaps among the sample and the container, extreme care was
taken in sealing the sample border without creating local compres-
sion between the tube and the samples. The effect of the irregular-
ities in the samples, and in particular at the edges of them, was
taken into consideration by repeating the tests with four different
samples for each material, and by repeating the measurements
four times for each sample.
Together with the sound absorption measurements, also the
airflow resistance was measured. Many authors have proposed
empirical formulas for the determination of the airflow resistance
of fibrous materials using values such as the bulk density of the
material and the diameter of the fibers [20]. However, given the
lack of a theoretical agreement, a direct measurement of the air- Fig. 1. Kundts tube for the absorption coefficient measure (above) and airflow
flow resistance was done according to the ISO 9053 standard resistance measurement system at alternate airflow (below).
134 U. Berardi, G. Iannace / Applied Acoustics 115 (2017) 131138

Table 1
Main properties of the materials studied in this paper and their acoustic absorption measurements in one-third octave bands.

Material Density Thickness Flow resistivity Frequency [Hz]

3
[kg/m ] [m] [Rayl/m] 125 160 200 250 315 400 500 630 800 1000 1250 1600 2000
Kenaf Light 50 0.06 2700 (290) 0.09 0.09 0.09 0.14 0.21 0.25 0.31 0.44 0.56 0.67 0.80 0.84 0.87
Dense 100 0.04 3500 (240) 0.08 0.12 0.16 0.16 0.20 0.30 0.34 0.41 0.54 0.73 0.83 0.93 0.93
0.06 3500 (240) 0.10 0.17 0.2 0.25 0.36 0.53 0.63 0.73 0.8 0.94 0.96 0.94 0.93
Wood Fibers 100 0.06 1600 (300) 0.20 0.26 0.29 0.31 0.57 0.75 0.79 0.86 0.75 0.64 0.71 0.91 0.91
Mineralized 260 0.03 1800 (450) 0.05 0.07 0.08 0.15 0.06 0.10 0.07 0.03 0.06 0.16 0.17 0.35 0.45
Hemp 50 0.03 1400 (170) 0.01 0.03 0.05 0.11 0.16 0.18 0.21 0.32 0.40 0.47 0.61 0.67 0.75
Coconut 60 0.05 1500 (200) 0.10 0.19 0.26 0.16 0.16 0.29 0.34 0.36 0.47 0.68 0.77 0.79 0.77
0.10 1500 (200) 0.25 0.36 0.29 0.32 0.58 0.78 0.82 0.89 0.87 0.76 0.81 0.92 0.97
Straw 60 0.10 410 (120) 0.21 0.26 0.24 0.25 0.36 0.50 0.63 0.77 0.88 0.96 0.98 0.95 0.92
Cane Mixed 400 0.04 850 (130) 0.05 0.10 0.12 0.15 0.09 0.17 0.22 0.22 0.27 0.49 0.63 0.59 0.57
Only wooden 470 0.04 1000 (125) 0.01 0.04 0.05 0.06 0.08 0.10 0.10 0.14 0.25 0.49 0.64 0.48 0.42
Only bark 145 0.04 800 (40) 0.10 0.11 0.13 0.16 0.12 0.22 0.28 0.30 0.45 0.72 0.75 0.60 0.62
Cardboard 140 0.115 250 (50) 0.10 0.13 0.19 0.19 039 0.53 0.47 0.45 0.56 0.50 0.55 0.61 0.75
Sheep wool 40 0.04 2100 (150) 0.10 0.13 0.11 0.13 0.17 0.26 0.33 0.43 0.55 0.71 0.85 0.93 0.95
Cork 100 0.03 1000 (150) 0.01 0.01 0.02 0.02 0.03 0.05 0.02 0.07 0.14 0.21 0.50 0.91 0.78

Kenaf Mineralized wood Hemp

Coconut Straw Cane

Cardboard Sheep wool Cork

Fig. 2. Photos of the different fibers tested in this paper.

 U. Berardi, G. Iannace / Applied Acoustics 115 (2017) 131138 135

4.1. Kenaf 4.5. Straw

The kenaf fiber is obtained from the stem of the plant, and is
widespread in Africa, Asia, and India. Kenaf panels are commer-
cially available as semi-rigid panels of different density and thick-
ness. The measured sound absorption coefficients were low at low
frequency and high at medium and high frequencies (Table 1). The
tests showed that doubling the density of the material has compa-
rable effects to increase its thickness.
4.2. Wood
The wood fiber is largely available as a waste of wood works.
The dried raw material was cut and weakened. The addition of bin-
der is common in many processed wood products in order to give
more shape stability, but it results in a significant increase of the
airflow resistivity, and lower sound absorption [24]. The results
of the acoustic measurements (Fig. 2) show high sound absorption
values above 500 Hz.
4.3. Hemp
The hemp fiber derives from the textile hemp whose low qual-
ity fibers cannot be used for textile applications. The plant grows in
temperate zones, and it is generally available in large quantities.
The hemp fiber has good thermal and acoustic properties [25].
Results in Table 1 show that the values of the absorption coefficient
at medium and high frequencies were discrete, whereas at low fre-
quency, the absorption was negligible and generally lower from
those reported in other studies [25].
The large availability of straw resulting from agricultural sub-
product has often led to consider this material for building applica-
tions. The thermal resistance performance of straw have resulted
in many cases to adopt this fiber as the main material in building
envelopes. The sound absorption coefficients of straw were evalu-
ated in other studies [27], with results below the ones found in the
present analysis, where straw showed sound absorption coeffi-
cients above 0.80 consistently above 630 Hz.
4.6. Cane
The cane or reed is a widespread plant that grows near water
courses. It has a very fast growing process, and for this reason, it
is frequently cut, making the raw material widely available. Giant
reeds usually reach 6 m in height and 23 cm in diameter, with
3060 cm long leaves. Canes were crushed in order to have smaller
pieces. Three types of shredded material were considered: solely
wooden parts (average length 4 cm, width 1 cm, and thickness
3 mm), mixed composed of wooden parts and the cortex with
mixed dimensions (the bark comes from the outer coating of the
giant reeds), and only bark parts. The absorption coefficient for
the bark showed a low value at low frequencies for the 4 cm thick
sample, and only the test for 8 cm thick sample resulted in more
significant sound absorption. The test for the wooden part showed
a strong resonance, which occurred at the frequency of 1250 Hz for
the 8 cm thick samples, and 630 Hz for the 4 cm thick ones.
4.7. Cardboard

4.4. Cocunut Cardboard is produced by recycled papers, and it is a light and

The palm plant is typical in tropical regions, and its fruit fiber,
the coconut, has generally been used in applications where robust-
ness is demanded. Coconut fiber is obtained from the mesocarp,
the thick fibrous layer that covers the shell of the nut. The coconut
fiber has a mean diameter of 250 lm, being a big fiber if compared
with typical mineral or other natural fibers [8,26]. The raw coconut
fiber showed a good sound absorption coefficient at both low and
medium frequencies, although the acoustic performance was more
remarkable for thicker samples. The sound absorption results were
comparable with those by Fouladi et al. [8].
cheap material, largely available worldwide [28]. For the sound
absorbing tests, the material was mounted so that the veins were
parallel to the impedance tube. This allowed the sound waves to
propagate through the material, although this assembly resulted
in particularly low value of the airflow resistivity.
Results of the sound absorption coefficient of the cardboard
samples show that the material has a good absorption coefficient
only at medium and high frequencies. In fact, the low airflow resis-
tivity even for a thick sample shows that the cardboard cannot be
considered rigid and porous, while the internal surface of the card-
board cavities plays a significant role for the sound absorption.

Table 2
Best fitting coefficients calculated for the different natural fibers studied in this paper.

Material Density Thickness C1 C2 C3 C4 C5 C6 C7 C8

Kenaf Light 50 0.06 0.0238 1.0683 0.1115 0.6167 0.1858 0.5462 0.1568 0.4173
Dense 100 0.04 1.2405 0.0370 0.4082 0.3840 0.8666 0.6454 3.6462 20.7484
0.06 0.0406 1.5843 0.0345 1.7689 0.2014 1.0106 0.2966 0.9598
Wood Fibers 100 0.06 0.0770 2.4825 0.8795 0.6510 0.4505 1.3031 0.0965 0.8289
Mineralized 260 0.03 0.4709 0.3157 0.7472 0.3304 0.9262 0.1364 0.0386 6.0367
Hemp 50 0.03 0.0615 1.2986 0.0022 1.6450 0.3105 0.0658 0.1497 2.0666
Coconut 60 0.05 0.2067 0.1153 0.1339 0.3849 0.2756 0.3789 0.3718 2.4672
0.10 0.1055 0.8093 0.1798 0.4844 0.2339 0.6318 0.9610 0.2537
Straw 60 0.06 0.0518 3.9867 1.3083 1.5634 1.1650 1.0008 0.3775 1.6028
Cane Mixed 400 0.04 0.5418 3.4170 0.0042 7.0730 0.2629 0.3907 1.4495 1.4566
Only wooden 470 0.04 0.2341 2.4582 0.0006 2.3153 0.1333 0.7513 0.5179 1.1595
Only bark 145 0.04 0.0088 4.2802 0.3209 0.9552 0.5619 0.9230 0.7159 0.8210
Cardboard 140 0.115 0.4879 0.7849 0.0485 1.9727 0.1816 0.5399 0.4635 1.2578
Sheep wool 40 0.04 0.4741 1.2224 0.4913 0.3186 0.8614 0.2924 0.1730 5.8785
Cork 100 0.03 0.1511 6.3056 0.1621 2.3860 0.2376 1.7423 0.4727 1.0326
136 U. Berardi, G. Iannace / Applied Acoustics 115 (2017) 131138

4.8. Sheep wool
Since ancient times, the sheep wool has been used for cloth-
ing thanks to its excellent thermal insulation properties [25,29].
The wool is generally composed of many different amino acids,
which form long chains and contribute to the fiber resilience.
As it grows from the sheep skin, wool naturally groups into
staples which contain thousands of fibers of different kinds. In
case of fire, the wool is self-extinguishing and does not emit
toxic substances. The sheep wool is an excellent sound-
absorbing material, thanks to micro-cavities of which it is com-
posed [29,30]. The value of the absorption coefficient resulted
high at medium and high frequencies, with a fairly homogenous
behavior. In fact, the sound absorption values obtained in this
study were generally higher than those reported in other studies
[25,30], especially at high frequency.

Fig. 3. Best-fit curves with quadratic errors for the different fibers tested in this paper.
 U. Berardi, G. Iannace / Applied Acoustics 115 (2017) 131138
Fig. 3 (continued)
4.9. Cork
The cork oak, an evergreen oak species reaching a height of
20 m, is the raw material for the production of cork. Cork oaks
grow mainly in Portugal, Spain, Tunisia, Morocco, Italy, France,
and Algeria, with the European cork industry producing the vast
majority of cork worldwide [6]. The cork is not a fiber, since it is
composed of spherical light granules containing air. This property
allows the panel to be light, elastic, and thermal insulating. Sound
absorption results showed that the absorption coefficient at low
and medium frequencies is negligible, probably as a consequence
of the large size of the cork grains. On the contrary, results up to
0.9 were obtained above 1600 Hz.
5. The best-fit modelling to the measured results
In order to apply the methodology presented in Section 3, it was
first necessary to calculate the normal sound absorption at the dif-
ferent frequency bands. Given the uncertainty in the low frequency
measurements with the Kundts tube, the calculations limited to
the sound absorption values in the one-third octave bands from
125 Hz to 2000 Hz. Then, using the inverse method described in
Section 2, which is based on the minimization of the errors
between measured values and predicted ones, the eight best-fit
coefficients for determining the acoustic impedance and the prop-
agation constant for each material studied in this paper were
found.
Table 2 reports the obtained coefficients, while the results of the
best-fit curves are shown in Fig. 3. Looking at the coefficients in
Table 2, it emerges a randomness of their values. In fact, the results
prevent generalize any trend for natural fibers, and also for mate-
rial with similar airflow resistivity, the best-fitting coefficients
were different as a consequence of the different sound absorption
data obtained in the impedance tube measurements, and in the
optimization procedure.
The resonances and the details of acoustic response were gener-
ally well depicted in the modelling predictions shown in Fig. 3. This
figure also shows that the quadratic errors obtained for the differ-
ent materials was always negligible, although 11 one-third octave
bands data were adopted. This proves that the coefficients reported
in Table 2 well represent the ones needed for predicting both the
acoustic impedance and the propagation constant of the several
natural materials with high accuracy.
6. Conclusions
Measurements carried out on samples of natural fibers have
shown that these materials have good sound absorption coeffi-
cients, especially at medium and high frequencies.
137
High standard deviation values of the airflow resistance mea-
surements confirmed the inhomogeneity of natural fibers and the
difficulties in creating models for predicting the behavior of these.
The high variability of the results that were obtained suggested to
limit the measure of material properties with direct methods, and
suggested to consider models able to predict the material behavior
with the lowest possible number of variables. The study then used
an inverse optimization method to calculate the coefficients that
best describe the acoustic impedance and the propagation constant
for the different natural fibers. Using an optimization technique it
was possible to define laws that resulted in negligible errors
between the measured and predicted values for the different fibers.
This means that the adoption of 11 one-third octave bands data
allowed a good test for the validity of the defined theoretical equa-
tions. Finally, this paper provides a series of coefficients for defin-
ing the laws that model the acoustic behavior of different natural
fibers. The availability of simple models to predict their behaviors
and the consequent possibility to implement these models in sim-
ulation software were considered necessary steps to promote a
broader diffusion of natural fibers for sound absorption applica-
tions. Hopefully, acoustic software will now be able to include nat-
ural materials within their typical libraries.
Acknowledgements
The first author is grateful to Jess Carbajo and Jaime Ramis
from the University of Alicante for their support in the optimiza-
tion of the MATLAB code used for the analysis reported in this
paper.
References
[1] Gl P, Gourdon E, Arnaud L. Acoustical properties of materials made of
vegetable particles with several scales of porosity. Appl Acoust
2011;72:24959.
[2] Asdrubali F, Schiavoni S, Horoshenkov KV. A review of sustainable materials for
acoustic applications. Build Acoust 2012;19(4):283312.
[3] Iannace G, Maffei L, Trematerra P. On the use of green materials for the
acoustic correction of classrooms. In: Proceedings of EuroNoise, Prague. p.
8994.
[4] Zabalza Bribia I, Capilla AV, Aranda Usn A. Life cycle assessment of building
materials: comparative analysis of energy and environmental impacts and
evaluation of the eco-efficiency improvement potential. Build Environ
2011;46:113340.
[5] Arenas JP, Crocker MJ. Recent trends in porous sound absorbing materials for
noise control. Sound Vib 2010;44(7):127.
[6] Berardi U, Iannace G. Acoustic characterization of natural fibers for sound
absorption applications. Build Environ 2015;94(2):84052.
[7] Fatima S, Mohanty AR. Acoustical and fire-retardant properties of jute
composite materials. Appl Acoust 2011;72:10814.
[8] Fouladi MH, Ayub M, Nor MJM. Analysis of coir fiber acoustical characteristics.
Appl Acoust 2011;72:3542.
[9] Delany ME, Bazley EN. Acoustical properties of fibrous absorbent materials.
Appl Acoust 1970;3(2):10516.
138 U. Berardi, G. Iannace / Applied Acoustics 115 (2017) 131138
[10] Garai M, Pompoli F. A simple empirical model of polyester fibre materials for
acoustical applications. Appl Acoust 2005;66(12):138398.
[11] Arenas JP, Rebolledo J, del Rey R, Alba J. Sound absorption properties of
unbleached cellulose loose-fill insulation material. Bioresources 2014;9
(4):622740.
[12] Bonfiglio P, Pompoli F. Inversion problems for determining physical
parameters of porous materials: overview and comparison between different
methods. Acta Acust United Acustica 2013;99(3):34151.
[13] Miki Y. Acoustical properties of porous materials modifications of Delany-
Bazley models. J Acoust Soc Jpn 1990;11(1):1924.
[14] Berardi U, Iannace G. Determination through an inverse method of the acoustic
impedance and the propagation constant for some natural fibers. In:
Proceedings of InterNoise, San Francisco.
[15] Biot MA. Generalized theory of acoustic propagation in porous dissipative
media. J Acoust Soc Am 1962;34:125464.
[16] Attenborough K. Acoustical characteristics of porous materials. Phys Rep
1982;82(3):179227.
[17] Allard JF. Propagation of sound in porous media. Elsevier Appl Sci 1993.
[18] Pispola G, Horoshenkov KV, Khan A. Comparison of two modeling approaches
for highly heterogeneous porous media. J Acoust Soc Am 2007;121(2):9616.
[19] ISO 10534-2. Acoustics determination of sound absorption coefficient and
impedance in impedance tubes Part 2: transfer-function method; 1998.
[20] Bies DA, Hansen CH. Flow resistance information for acoustical design. Appl
Acoust 1980;13(5):35791.
[21] ISO 9053. Acoustics materials for acoustical applications determination of
airflow resistance; 1991.
[22] Del Rey RJ, Alba J, Arenas JP, Ramis J. Evaluation of two alternative procedures
for measuring airflow resistance of sound absorbing materials. Arch Acoust
2013;38(4):54754.
[23] Ramis J, del Rey R, Alba J, Godinho L, Carbajo J. A model for acoustic absorbent
materials derived from coconut fiber. Mater Constr 2014;64(313):e008.
[24] Wassilieff C. Sound absorption of wood-based materials. Appl Acoust 1996;48
(4):33956.
[25] Oldham DJ, Egan C, Cookson R. Sustainable acoustic absorbers from the
biomass. Appl Acoust 2011;72:35063.
[26] Khedari J, Charoenvai S, Hirunlabh J, Teekasap S. New low-cost insulation
particleboards from mixture of durian peel and coconut coir. Build Environ
2004;39:5965.
[27] McGinnes C, Kleiner M, Xiang N. An environmental and economical solution
using straw. J Acoust Soc Am 2005;118:1869.
[28] Asdrubali F, Pisello AL, DAlessandro F, Bianchi F, Fabiani C, Cornicchia M, et al.
Experimental and numerical characterization of innovative cardboard based
panels: thermal and acoustic performance analysis and life cycle assessment.
Build Environ 2016;92:14559.
[29] Ballagh KO. Acoustical properties of wool. Appl Acoust 1996;48(2):10120.
[30] Berardi U, Iannace G. Characterization of sheep wool panels for room acoustic
applications. In: Proceedings of ICA 2016, Buenos Aires, Argentina.