On variation of absorption factor due to measurement method and correction factors for conversion between methods
Anna Färma) Ragnar Glavb)
Acoustics department, Scania CV AB, Södertälje, Sweden
KTH Royal Institute of Technology, School of Engineering Science,
Department of Aeronautical and Vehicle Engineering, Centre for ECO2 Vehicle Design, MWL 100 44 Stockholm, Sweden
Susann Boijc)
KTH Royal Institute of Technology, School of Engineering Science,
Department of Aeronautical and Vehicle Engineering, Centre for ECO2 Vehicle Design, MWL 100 44 Stockholm, Sweden
Sound absorbing materials are used in many applications to reduce sound, and their sound absorbing characteristics are most often determined experimentally since theoretical determination is difficult. Sound absorption factors are used in material specifications as well as input to numerical simulations.
Several methods for experimental determination of the absorption factor exist, two of them standardized and frequently used. It is commonly known that the absorption factor obtained by these two methods differs as different sound fields are prescribed by the standards. However, the size of the differences has not been so well described. Due to this difference, the choice of method is critical in order to avoid errors in simulations and specifications of material properties.
Experimental determination of absorption factors for three commonly used absorbers was performed, resulting in significant differences between the two methods. Correction factors to compensate the absorption factor determined at one acoustic state and used in another are given. Theory verifying the differences is also presented.
a) email: anna.farm@scania.com
b) email: ragnar.glav@scania.com
c) email: sboij@kth.se
1 INTRODUCTION
Porous material is used in many applications to reduce sound. The absorption factor of the material determines its quality and is most often determined experimentally. Experimental determination of the absorption factor is performed by either of two standardized methods, each based on a specific acoustic state. The obtained value on the absorption factor will depend on the acoustic state, either normal incidence or diffuse field excitation. Thus, systematic differences are to be expected in absorption for a specific material determined by the two different methods1. No thorough investigation has to the authors’ knowledge been made to determine the size of the differences.
This paper discusses the significance of these differences and also suggests correction factors between methods. It is shown that the choice of measurement method is indeed crucial.
Theory confirming differences obtained experimentally is provided and the results are discussed.
2 MATERIAL
The absorption factor is studied for isotropic porous foams with open pores, without any covering films or layers of different porosity. These are rather simple materials where viscous losses in the medium are the dominating characteristic. Structural motion of the material is often negligible.
The flow resistivity, σ, is a parameter often used to describe its acoustic properties. Typical values on commercially used absorbers are 1.0 k – 100 kNs/m4. The thickness is typically about a few centimetres.
3 EXPERIMENTAL METHODS
Two standardized methods, one based on normally and one based on randomly incident sound waves are used to experimentally determine the absorption factor of the materials. The methods also differ considerably concerning required measurement facilities and amount of sample material needed. Obviously these two issues often largely influence the choice of method in practice.
3.1 Normal incidence
The standing wave tube measurement2 is based on plane waves normally incident on the measurement sample. The pressure is measured at two positions, and the impedance of the surface is obtained using the wave decomposition technique. The absorption factor for normal incidence, αn, is then obtained from the surface impedance. The method is based on plane waves in the tube, why tubes of different diameters have to be used to cover a wider frequency range.
The method solely requires the impedance tube and relatively small measurements samples;
no dedicated acoustic measurement room is needed. The mounting in the tube will alter the stiffness of the sample, hence, for materials where the structural motion is significant compared to the viscous losses, the mounting affects the obtained absorption. In such cases, different results are obtained for different tube diameters. This implies that the method is preferably used for materials where viscous losses are dominant.
3.2 Th incident Sabines tested m determin reverber material tends to 4
Ex thicknes Table 1
Th 1/3-octa αn, whil wave tu line). T 5000 Hz
Fig. 1
Diffu he reverbera t on the ma
’ formula o material, th
ned. The m rant field. T l is needed, o overestima EXPE xperimental
ss 20 mm w – Flow resi
he absorptio ave bands ar
le the dashe ube measure The measure
z.
1 – Absorp thickness
se field exc ation room m
aterial rand on the mea he absorptio method requ
The method typically 1 ate the absor ERIMENT
determinat with flow res istivity of th
on factor for re given in ed lines rep ements, sam ements in t
ption factor d s 20 mm. So
citation measuremen domly from
sured rever on factor f ires a room d is prefera 10 m2. Due
rption factor AL RESUL tion of the sistivity acco he three teste
Material A: Grey B: Light C: Black r materials A
Figure 1. Th present abso mples of two the reverbe
determined f olid line rep
line a
nt3 is based m all angles rberation tim for a diffus m that meets ably used b to absorptio r.
LTS absorption ording to Ta ed materials Flow [k Grey
A–C determ he solid line orption for o diameters eration room
from two m resent abso at diffuse fie
d on a diffus , from norm me in the r se field ex s the standa below 5 kH on at the ed
factor is m able 1.
s.
w resistivity kNs/m4]
3.5 8.0 40 mined from t
es represent diffuse fiel
are used; 2 m are valid
measurement orption at no eld excitatio
se sound fie mal to graz room both w
citation, αd
ard regardin Hz. A large dges of the m
made for th
y
the two mea t absorption d excitation 29 (thin line d in the fre
t methods fo ormal incide on, αd.
eld. Sound zing angle.
with and w
d, of the m ng the prope e area of th material, th
hree porous
asurement m n at normal i
n, αd. In the e) and 100 m equency ran
or materials ence, αn,, and
waves are By using without the material is
erties of a he sample his method
foams of
methods in incidence, e standing mm (thick nge 315 –
s A-C of d dashed
As expected the absorption factor determined at normal incidence and at diffuse field excitation respectively differs significantly for all tested materials. The absorption at normal incidence for two of the materials, A and B, show the anticipated characteristics of a porous absorber with viscous losses due to fluid motion inside the pores. The third material C, however, differs from that pattern.
For material C, a clear peak in the absorption factor at normal incidence indicates that the structural motion in the material has significant impact on the resulting absorption. The flow resistivity for this foam is high which by definition means that the material resists fluid motion, instead forcing the structure to move. Structural motion in material C is further verified when results from the wide and narrow tubes are compared (thin and thick solid black line) since the resonance occurs at a higher frequency for the narrow tube where the sample is smaller and hence stiffer due to the mounting at the edges.
For the two materials A and B, the relation between the absorption factors obtained by the two methods show the same pattern for both materials. The absorption for diffuse field excitation is larger than the absorption for normal incidence except at the highest frequency bands measured. For even higher frequencies it is expected that the absorption for diffuse field excitation is smaller than for normal incidence1. At normal incidence the absorption increases slowly with frequency up to a certain value where it stabilizes. For diffuse field excitation, the absorption increases faster with frequency up to about the same value of the absorption factor as for normal incidence. This behaviour is to be expected since the apparent thickness of the absorber is larger for sound incident at oblique angle of incidence, increasing the absorption at low frequencies. Since the absorption factor is rather small in the low frequency region, this difference in absorption is significant in a wide frequency range. Relative differences as large as 100 % are seen and this is indeed substantial.
For the kind of material studied, where viscous losses due to fluid motion inside the pores are dominating and the structural motion is negligible, correction on absorption factors between methods would be of great use. In Figure 2 below, the correction factor K, according to Eq. (1) is given for materials A and B. Numerical values of K are given in Table 2.
(1)
Fig. 2 –
Table 2 diffuse f
Thr i) the pr amount interacti differen i) T This is t reflected increase
– Measured diffuse fie – Measured field excitat
ree mechani ropagation d
of energy ing with the nce in absorp The propag the shortest d wave. Fo ed by a facto
d correction eld excitation
d correction ion on 20 m 1/3
isms contrib distance in
entering t e surface, di ption factor gation distan possible di or other ang
or 1/cosϴ, l
factor K be n on 20 mm n factor K b mm thick ma 3-octave ban
[Hz]
315 400 500 630 800 1000 1250 1600 2000 2500 3150 4000 5000 butes to the
the materia the material ifferent mec obtained fr nce at norm istance a wa gles of inci eading to in
etween abso m thick mater
etween abso terials of dif nd A: Grey
KA 0.06 0.12 0.19 0.22 0.23 0.24 0.32 0.27 0.21 0.15 0.09 -0.03 -0.31 e total amou al, ii) the p
l. Dependin chanisms w rom the two al incidence ave can trav idence, ϴ, t ncreased abs
orption facto rials of diffe orption fact ifferent flow y B: Light KB 0.02 0.09 0.22 0.30 0.32 0.31 0.38 0.36 0.30 0.19 0.14 0.04 -0.24 unt of visco article velo ng on the will dominate
methods co e equals twi vel inside the
the propaga sorption.
or for norma ferent flow re tor for norm w resistivity,
t grey 2 9 2 0 2 1 8 6 0 9 4 4
4
ous losses r ocity in the frequency e. This is th ompared in t ice the thick e material w ation distan
al incidence esistivity, σ.
mal incidenc σ.
related to ab material, an and the so he explanati this paper.
kness of the with an incid nce in the m
e and for .
e and
bsorption;
nd iii) the ound field
ion for the e material.
dent and a material is
ii) High particle velocity gives high absorption. The particle velocity is zero at the backing rigid wall, hence small amounts of sound is absorbed there. Away from the rigid wall the velocity increases and is maximized one fourth of a wave length from the wall. The amount of absorption is hence coupled to the material thickness relative to the wavelength.
iii) When the angle of incidence increases the amount of sound energy reflected directly at the surface increases leading to smaller amount of sound energy entering the material. This phenomenon can e.g. be observed at the interface between air and an infinitely thick absorber (i.e. no rigid backing and no reflected wave inside the absorber). In that case the absorption is determined solely from the reflection at the interface, since all energy that enters the material is absorbed.
5 THEORETICAL RESULTS
There are many different material models that can be used to determine the sound absorption of porous materials. These models have various levels of complexity, ranging from semi-empirical Equivalent Fluid Models, EFM, (such as Delany and Bazley4), limp models neglecting motion in the structure5 down to models capturing the microstructure of the material such as Biot6 and Attenborough7. The latter models require assumptions on the microstructure properties and knowledge of material parameters which in many cases are hard to estimate.
Here, an EFM suggested by Delany and Bazley is used to calculate the absorption via the wave decomposition technique and the transfer matrix method (TMM)8 on a 20 mm thick material fronting a rigid wall. The absorption factor is calculated for angles of incidence ranging from normal to grazing incidence, from which the random incidence absorption is calculated as
0 2 sin 2 . (2) The absorption factor for normal incidence corresponds to the absorption determined in the standing wave tube, αn = α(0), whereas the randomly incident absorption approximately corresponds to the absorption for diffuse field excitation, αr ≈ αd.
For the flow resistivity according to materials A and B in this paper, the absorption factor at normal and random incidence is calculated. Since structural motion occur in material C, the behaviour of the material is not captured properly by the simplified model used, why the absorption is not calculated for material C.
Fig. 3 –
The normal frequenc behavio calculat the surfa 6
The two met Abs experim factor o awarene determin
For and the these m except i large an findings fields, v The aligned determin absorpti correctio
– Calculate wall at no incidence, e calculated incidence i cy range stu our as the r
tions thus ve face, that wa
SUM e absorption thods, howe sorption fac mentally by t obtained fr ess of the va
ne the absor r two of the difference materials. Th in the highe ngles of inci s are also re verifying the e method fo with the ex ned from e ion factor is
ons on the a
d absorptio ormal and ra
, α(0),and d d absorption is seen to b udied. The relation betw
erify the dif as shown by
MARY n factor of a ever, resultin ctors for thre two method rom the tw
alidity of th rption in an e materials, between ab he absorptio
est frequenc idence. Inde eproduced b e phenomen or determin xpected sou
ach method s determine absorption f
n factor for andom incid dashed line a n for materi
be smaller relation be ween the ex fferences in y the experim
a material is ng in differe ee porous m ds; for norm wo methods he methods
appropriate the structu bsorption fr on was high cy studied, a
eed, the siz by theoretic non.
ing the abs und field in d gives the
d by a meth factor have
r material A dence. Solid at random in
als A and B than for ra etween α(0)
xperimental n absorption
ments.
s convention ent absorptio materials wit
ally inciden s differs s
and the inte e way.
ural motion rom the two her at diffus
as the appa ze of the dif cal calculatio
orption fact the applica absorption hod with an to be made
and B of th d line repres ncidence, α B are shown andom incid and αr in t lly determin
, due to typ
nally determ on factor.
th different nt waves and
ignificantly ended applic
had neglig o methods se field exc arent thickne
fferences wa ons of abso tor should t ation of the for that sp nother acous
in order to
hickness 20 m sent absorpt
r.
n in Figure dence for b the calculat ned αn and pe of sound
mined experi flow resisti d for diffuse y indicating cation of th ible influen showed the itation than ess of the a as larger tha orption facto to as large e material. T
ecific acous stic state th
avoid erron
mm in front tion at norm 3. The abs both materi tions shows αd in Figu field interac
imentally by ivity were d e field excit g the impo he material i nce on the a
e same beha n at normal absorber inc an anticipat ors at differ extent as po The absorpt
stic state on an in the ap neous predic
t of a rigid mal
sorption at ials in the the same ure 1. The cting with
y either of determined ation. The ortance of
in order to absorption aviour for
incidence creases for ted. These rent sound ossible be tion factor nly. If the pplication, ctions. For
this purpose, correction factors between diffuse field excitation and normal incidence are given for the two porous absorbers investigated in this work.
Theoretical determination of the correction factor from the thickness and flow resistivity of the material would be a useful tool in the future why further investigation on this are to be made.
The relation between the absorption factors for different acoustic states could also be measured for this purpose.
7 ACKNOWLEDGEMENTS
This research was supported by the Center for ECO2 Vehicle Design, and Scania CV AB provided material for experiments.
8 REFERENCES
1. Färm A., Boij S. and Glav R. “On Sound Absorbing Characteristics and suitable Measurement methods”. Proceedings of the 7th International Styrian Noise, Vibration and Harshness Congress, 2012.
2. ISO 354:2003. Acoustics -- Measurement of sound absorption in a reverberation room.
3. International standard ISO 10534-1:1996. Acoustics - Determination of sound absorption coefficient and impedance in impedance tubes -- Part 1: Method using standing wave ratio 4. Delany M. E. and Bazley, E. N., “Acoustic properties of fibrous absorbent materials”,
Applied Acoustics, 3: 105-116, 1970.
5. Corcos, G.M., “The Structure of the Turbulent Pressure Field in Boundary Layer Flows”, J.
of Fluid Mechanics, 18, 1964.
6. Biot, M. A. ”Generalized theory of acoustic propagation in porous dissipative media”, Journal of the Acoustical Society of America, 34 (9): 1254 - 1264, 1962.
7. Attenborough K., “ Acoustical characterization of porous materials” Physics Reports, 82 (3):
179-227, 1982.
8. Sastry J.S., Munjal M.L., “A transfer matrix approach for evaluation of the response of a multi-layer infinite plate to a two-dimensional pressure excitation”, Journal of Sound and Vibration, 182 (1):109–128, 1995.