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

net/publication/256132749

Transfer matrix modeling and numerical verification of locally reacting

acoustic mosaics

Conference Paper · January 2012

CITATIONS READS

0 168

5 authors, including:

R. Panneton Kévin Verdière

Mecanum Inc, and Université de Sherbrooke Mecanum Inc
181 PUBLICATIONS   3,569 CITATIONS    17 PUBLICATIONS   112 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

PhD on "Characterisation of Macroscopically Inhomogeneous Porous Materials" View project

New designs of sound absorbing materials View project

All content following this page was uploaded by R. Panneton on 25 January 2016.

The user has requested enhancement of the downloaded file.

 Transfer matrix modeling
and numerical verification
DRIVE of locally reacting acoustic
mosaic
Intro
Method Raymond Panneton1
Results Kévin Verdière1
Saïd Elkoun1
Conclusion
Thomas Dupont2
Philippe Leclaire2

1 GAUS, Univ. de Sherbrooke (Qc, Canada)

2 DRIVE, ISAT, Univ. de Bourgogne (Nevers, France)

1
 Conventional sound
absorbers (CSA)
Multilayers
DRIVE

Single layers
Intro
Method
Results
Conclusion

The common point is…

2
 Modeling conventional
sound absorbers (CSA)
DRIVE Single layer – Transfer Matrix Method (TMM)

 p a b p
Intro v =  c a v 
  front    back
Method
Results Multilayer – TMM in series
 p  A B  p
Conclusion v =   
  front  C D   v back

 A B  a b a b a b a b a b
C D =  c a   c a   c a   c a   c a 
   1  2  3  4  5

3
 Transmission through a series of layers (multilayer)
Transfer matrix in series 1
p−
2 3 4

DRIVE
 p1 
  p +
 p2 
 
 p3 
 
 p4 
 
…
 v1   v2   v3   v4 

Intro 0 x1 x
Method p ( x ) Ae − jk1x + Be jk1x
=
 
k1 : Complex wave number of medium 1
Results A − jk1x B jk1x
v ( x)
= e − e Z1 : Characteriztic impedance of medium 1

Z1 Z1
Conclusion
( x1 ) Ae− jk1x1 + Be jk1x1
 
p1= p ( 0 )= A + B =
p2 p=
A B 
1
( x1 )  e− jk1x1 −  e jk1x1

= ( 0)
v1 v= ( A − B) =
v2 v=
Z1 Z1
Z

 cos k1h1 jZ1 sin k1h1 

 p1     p2   p1   p2 
 = 1     = T1  

 v1   j  sin k1h1 cos k1h1   v2   v1   v2 
 Z1 
 Transmission through a series of layers (multilayer)
Transfer matrix in series 1
p−
2 3 4

DRIVE
 p1 
  p +
 p2 
 
 p3 
 
 p4 
 
…
 v1   v2   v3   v4 

Intro 0 x1 x2 x3 x
Method  p1   p2   p2   p3   p3   p4 
  = T1     = T2     = T3  
Results  v1   v2   v2   v3   v3   v4 

Conclusion
 p2   p4  Transfer matrix combined
  = T1T2 T3  
 v2   v4  in series

 p2   p4 
  = TG 
 v2   v4 
 Non-conventional sound
absorbers (NCSA)
DRIVE
Patchwork Dead-end surface porosity
Intro
Method
Results
Conclusion
Grooved surface
The common point is…

6
Source : J. Ducourneau et al.
Journal of Sound and Vibration 329 (2010) 2276–2290
 Modeling non-conventional
sound absorbers (NCSA)
DRIVE Analytical, Semi-analytical, and FE models
• Non-propagative static model – Thin patchworks(kl<<1 )
[Atalla, Panneton, and Allard, ACUSTICA - Acta acustica 83, 891-896 (1997)]
Intro • Finite element method – Patchworks
[Atalla, Panneton, and Allard, ACUSTICA - Acta acustica 83, 891-896 (1997)]
Method
• Wave based method (indirect Treff method) - Patchworks
Results [Lanoye, Vermerir, Lauriks, Sgard, Desmet, JASA 123, 793-802 (2008)]
• Waveguide method – Grooved surface
Conclusion [J. Ducourneau et al. , Journal of Sound and Vibration 329, 2276–2290 (2010) ]
• Finite element method – Parallel MPP absorbers
[Wang and Huang, J. Acoust. Soc. Am. 130, 208-218 (2011)]
• Wave theory, Bruijn’s theory – Grooved surface
[Wang, Leistner and Li, Applied Acoustics 79, 960-968 (2012)]

7
 Objective of the research
DRIVE
Extend the Transfer Matrix Method
(TMM) to non-conventional sound
Intro absorbers having surface
Method hetorogeneities.

Results For this presentation, we limit to

Conclusion locally reacting acoustic mosaic

8
 Transfer Matrix Method
Acoustic Mosaic Periodic Elementary Patch
DRIVE (PEP)

Intro
Method
Results
Conclusion Ex.: PEP with N =3 cells

Assumptions
• The walls of a cell are impervious (locally reacting)
• Each cell is filled with a sound absorber(single or multilayer)
• Sound absorbers are modeled as rigid/limp equivalent fluid
• PEP excited by a normally incidence plane wave
• Wavelength much larger than the PEP size
• The surface ratio of a cell in a PEP is :
ri = Si S PEP 9
 Transfer Matrix Method
PEP with N cells
DRIVE  P1   P1′ 
  1  
U1  U1′ 
P  P2   P ′   P′ 
Intro     2  2  
U  U 2  U 2′  U ′

…
Method  PN   PN′ 
  N  
Results U N  U N′ 

Conclusion  Pi  Ti ,11 Ti ,12   Pi′ 

 =   Transfer matrix
U T T −
 i   i ,21 i ,22   i 
U ′

U i   yi ,11 yi ,12   Pi 
 =   Admittance matrix
U i′  yi ,21 yi ,22   Pi′
yi ,11 =
Ti ,22 / Ti ,12 ; yi ,21 =
−1/ Ti ,12 ; yi ,22 =
Ti ,11 / Ti ,12 ;
(Ti,12Ti,21 − Ti,11Ti,22 ) / Ti,12 =
yi ,12 = yi ,21 ) 10
−1/ Ti ,12 (with |T|=1 for reciprocity , yi ,12 =
 Transfer Matrix Method
PEP with N cells
DRIVE  P1   P1′ 
  1  
U1  U1′ 
P  P2   P ′   P′ 
Intro     2  2  
U  U 2  U 2′  U ′

…
Method  PN   PN′ 
  N  
Results U N  U N′ 

Conclusion  Pi  Ti ,11 Ti ,12   Pi′ 

 =   Transfer matrix
U T T −
 i   i ,21 i ,22   i 
U ′

U i   yi ,11 yi ,12   Pi 
 =   Admittance matrix
U i′  yi ,21 yi ,22   Pi′
yi ,11 =
Ti ,22 / Ti ,12 ; yi ,21 =
−1/ Ti ,12 ; yi ,22 =
Ti ,11 / Ti ,12 ;
(Ti,12Ti,21 − Ti,11Ti,22 ) / Ti,12 =
yi ,12 = yi ,21 ) 11
−1/ Ti ,12 (with |T|=1 for reciprocity , yi ,12 =
 Transfer Matrix Method
 P1   P1′ 
Continuity conditions  
U1 
1  
U1′ 
DRIVE P  P2   P ′   P′ 
    2  2  
U  rUi i  Continuity of upstream flow U  U 2  U 2′  U ′
=∑
      Continuity of downstream flow

…
 
U ′ rU ′
 i i  PN   PN′ 
Intro   N  
U N  U N′ 

Method  P   Pi  Continuity of upstream pressure PEP

   = 
   Pi′
P ′ Continuity of downstream pressure for open-end cells
Results
Conclusion rU
i i  yi ,11 yi ,12   Pi 
  = ri   
rU ′
i i  yi ,21 yi ,22   Pi′

U    yi ,11 yi ,12    P  Admittance matrix

  =  ∑ ri      combined in
U ′   yi ,21 yi ,22    P′ parallel 12
 Transfer Matrix Method
From admittance to transfer matrix…
DRIVE

Intro
Method
Results
Conclusion

13
 Transfer Matrix Method
Particular case of cells with closed end
DRIVE
U  rU  Continuity of upstream flow
    =∑ i i 
U ′ rU ′ Continuity of downstream flow
i i
Intro Pi = P Continuity of upstream pressure

Method
Pj′ = P′ Continuity of downstream pressure for open-end cells
Results y
Pk′ = − k ,21 P Downstream pressure for cells closed to end
Conclusion yk ,22

Boundary condition backing a cell

j Cell with open end (subscript j)

k Cell with closed end (subscript k)

14
 Transfer Matrix Method
General case of open and closed backing…
DRIVE

Intro
Method
Results
Conclusion

15
 Transfer Matrix Method
Adding a backing layer (ex.: air cavity backing)
DRIVE parallel
1
Intro P  P′   P′′ 
  2    
U  U ′ U ′′

…
Method series
N
Results
PEP Cavity
Conclusion
 P′   P ′′ 
  = Tcavity   Transfer matrix for air-cavity
 
U ′  −U ′′ 
 cos k0 d jρ0 c0 sin k0 d 
Tcavity =  1 
 j ρ c sin k0 d cos k0 d 
 0 0  16
 Transfer Matrix Method
Adding a backing layer (ex.: air cavity backing)
DRIVE parallel
1
Intro P  P′   P′′ 
  2    
U  U ′ U ′′

…
Method series
N
Results
PEP Cavity
Conclusion
P  P ′′ 
  = TGlobal  
 
U  −U ′′  Combination of cells
in parallel and
TGlobal = TPEP Tcavity in series

17
 Transfer Matrix Method
Acoustical indicators
DRIVE
A B
TGlobal = 
Intro  C D 
Method Sound absorption coefficient
Results 2
A − ρ cC
Conclusion α= 1−
A + ρ cC

Sound transmission loss

B
= 20log A +
nSTL + ρ cC + D − 6 dB
ρc
18
 Finite Element Method
Objective : Validate the developed transfer matrix
DRIVE against FE simulations of impedance
tube measurements
Intro
Method
Virtual sample
Results
Conclusion
 Finite Element Method
FE model of the PEP only
DRIVE

Intro
Method Periodic
Results conditions

Conclusion Cell 1

1
Cell 2

Hardwall
condition

1
PEP Pairs of periodic surface
Cell 3 condition on the air domain
in contact with PEP
 Finite Element Method
FE model of the PEP in the tube
DRIVE

Intro
Method
Results
Micro 3 plane
Conclusion on termination
Air (anechoic or
cavity hardwall)
PEP
Pµ3

Micro 2 plane
Pµ2
Micro 1 plane
Pµ1
 Finite Element Method
Acoustical indicators
DRIVE
Reflection coefficient

Intro H12 e jks − 1 2 jkL End

R= − jks
e Air
Method 1 − H12 e cavity
PEP

Results Pµ 2
H12 = Micro 2 plane
Conclusion Pµ1
Micro 1 plane
Sound absorption coefficient
2
α = 1− R
Sound transmission loss

Pµ1
nSTL = 20log Pµ3 Pinc Pinc = − jkX µ1 jkX µ1
e + Re
 Description of sample 1
Honeycomb filled with porous materials
DRIVE

PEP contains: 3 cells

Intro PEP thickness: 15 mm
Cell width: 12 mm
Method Wall thickness: 0.5 mm
Results
Conclusion
Material 1 2 3
Open porosity 0.941 0.883 0.765
Flow resistivity (N.s.m-4) 6 260 34 403 80 697
Tortuosity 1.35 1.77 1.88
Viscous characteristic length (µm) 125 343 28
Thermal characteristic length (µm) 302 360 355

23
 Description of sample 1
Porous material modeling
DRIVE • JCA model
• Rigid frame
• Equivalent fluid model
Intro • Visco-inertial effects
• Johnson et al. (1987) - φ, σ, α∞, Λ
Method • Thermal effects
• Champoux and Allard (1991) - φ, Λ′
Results
Conclusion Transfer matrix for ith material

 cos keq l jZ eq sin keq l 

 
Ti =  1
j sin keq l cos keq l 
 Z eq 
 i
ρeq
keq = ω Z eq= ρeq K eq 24
K eq
 Results on sample 1
DRIVE

Intro
Method
Results
Conclusion

25
FE results obtained with COMSOL
 Results on sample 1
DRIVE

Intro
Method
Results
Conclusion

26
FE results obtained with COMSOL
 Results on sample 1
DRIVE

Intro
Method
Results
Conclusion

27
FE results obtained with COMSOL
 Description of sample 2
Air or Solid
Patchwork in square impedance tube1
DRIVE

Configuration: Series and Parallel

Intro Tube size : 0.6 x 0.6 m2
Cutoff frequency: 570 Hz
Method End condition: Rigid end
1st layer thickness: 50 mm foam
Results 2nd layer thickness: 50 mm
Conclusion
Material Melamine
foam
Open porosity 0.98
Flow resistivity (N.s.m-4) 10 000
Tortuosity 1.1
Viscous characteristic length (µm) 100
Thermal characteristic length (µm) 150
28
1Lanoye, Vermerir, Lauriks, Sgard, Desmet, JASA 123, 793-802 (2008)
 Description of sample 2
Air or Solid
Special remark
DRIVE

One assumption not fulfilled:

Intro No impervious wall
between the 2 patches
Method
parallel foam
Results
Conclusion

series
29
 Results on sample 2a
DRIVE

Intro
Method
Results
Conclusion
air

30
 Results on sample 2b
DRIVE

Intro
Method
Results
Conclusion
Solid

31
 Conclusion
DRIVE • Development of parallel assembling of Transfer
Matrix
Intro • Parallel assembling accounts for parallel
patchworks and surface non-homogeneity under
Method locally reacting assumption
Results • Can be assembled in series with classical transfer
matrix to account for multilayer configurations
Conclusion

32
 Conclusion
DRIVE • Validated for locally reacting acoustical mosaic
(honeycomb with impervious walls filled with materials)

Intro • Validated for non-locally reacting patchworks

when pressure diffusion between materials is
Method weak (patches with no impervious walls)
Results
Conclusion

33
 Future works
DRIVE • Test other and more complex configurations
• Remove assumptions to enable:
Intro • Excitations different from normal incidence
Method plane wave

Results • Interactions between cells (multipole matrix)

to account for pressure diffusion
Conclusion
• Poroelastic modelling of porous materials

See recent publications on this topic at:

http://ema.recherche.usherbrooke.ca/?page_id=264

34
 DRIVE

Intro
Method Thank you!
Results
Conclusion Questions?

35

View publication stats