You are on page 1of 33

A wavelet based model of one-dimensional periodic structure for wave

propagation analysis

Parikshit A. Sonekara) and Mira Mitrab)

Department of Aerospace Engineering,

Indian Institute of Technology Bombay,

Powai Mumbai 400076,


India

(Dated: March 23, 2009)

Wave propagation in periodic structure 1


Abstract
In this paper, a wavelet based method is developed for wave prop-

agation analysis of a generic multi-coupled one-dimensional periodic


structure. The formulation is based on periodicity condition and use

dynamic stiffness matrix of the periodic cell obtained from finite el-

ement or other numerical methods. The proposed numerical scheme


uses Daubechies scaling function and enables both time and frequency

domain analysis under arbitrary loading conditions. This is unlike the


Fourier transform based analysis which are restricted only to frequency

domain analysis. Here, in frequency domain, the dispersion character-

istics of the periodic structures, especially the band gap features are
studied. In addition, the method is implemented to simulate time do-
main wave response under impulse loading condition. The two exam-

ples considered are periodically simply-supported beam and periodic


frame/truss structures. In all the cases, the response obtained using

the present formulation is compared with the exact solution.

PACS numbers: 43.20.Bi, 43.20.Mv, 43.20.Gp

2
I. INTRODUCTION

Research on periodic structures, particularly study of their dynamic properties dates


back to eighteen century1 and started mostly to understand the wave characteristics of
lattice structures. The interesting feature shown by periodic structure is the bandgap phe-

nomenon i.e they exhibit certain frequency bands within which no wave propagates resulting

in filtering capabilities.

In early 70s and before, research activities were directed towards engineering periodic
structures2,3 like periodically supported beams and plates. Since then a lot of work on

modeling and analysis of periodic structures has been reported in literature4–8 . This is

primarily because such periodic structures are often encountered in engineering like periodi-

cally stiffened fuselage and wing of an aircraft, periodic layered media, bridge with repeated
truss-like structures. In addition, their band gap characteristics are utilized in applications

like vibration attenuation9,10 , piezoelectric transducers11,12 . Recently with the advent of

nano-structured materials, for example, carbon nanotubes and their composites, study of
their periodic properties will help in exploring their functional properties.

This paper aims at developing a generalized numerical scheme based on Floquet theo-
rem to simulate the overall wave response of one-dimensional multi-coupled linear periodic

structure in both time and frequency domains. These simulations will help to understand
the dynamic characteristics for different applications as mentioned earlier. The proposed

scheme is based on the use of Daubechies scaling function13 as approximation bases. These
basis functions are bounded in both time and frequency domains. The localized nature of

the functions in time allows accurate simulation of time domain responses unlike the Fourier

transform based methods which are restricted only to frequency domain analysis. However,
the localization in time domain causes reduced accuracy in frequency domain. As a result,

in this wavelet based method, the frequency dependent wave characteristics can be obtained

a)
parikshit@aero.iitb.ac.in
b)
mira@aero.iitb.ac.in; www.aero.iitb.ac.in/~mira; Corresponding author

3
accurately only up to a certain fraction of the Nyquist frequency14 .

The formulation starts with obtaining the dynamic stiffness matrix KD of a periodic cell.

Here, this matrix is obtained in the wavelet domain in contrast to the conventional dynamic

stiffness matrix which is obtained in frequency domain through Fourier transform. In this

paper, two different methods of deriving KD are described. In the first technique, KD is

calculated from the mass and stiffness matrices obtained from finite element (FE) followed by
wavelet transform. The second method is the wavelet based spectral finite element (WSFE)15

technique which directly formulates KD in wavelet domain starting from the governing
partial differential equations. In brief, WSFE follows FE procedure in transformed wavelet

domain. While the second method is computationally efficient and can model waveguides

of rather high complexities, FE method can model waveguides of any arbitrary geometry.

The FE method for such problems however involves huge computational cost.

Next, the periodicity of the structure is considered using Floquet theorem to develop

the numerical scheme for wave propagation analysis. The method is described in detail
in the following sections. The method developed is then implemented first to obtain the

dispersion relation i.e frequency dependence of the propagation constants for different ex-
amples of periodic structures. These include periodically simply-supported Euler-Bernoulli
and Timoshenko beams, periodic frame structures. Next, the time domain wave responses

of these periodic structures due to broad-band impulse loading are simulated. These re-

sponses obtained through the periodicity assumption are compared with the exact solution.

The response obtained using the proposed method is referred as the periodic solution in the

remaining parts of the paper.

II. DYNAMIC STIFFNESS MATRIX

As mentioned earlier, the first step in formulation of the proposed wavelet based nu-

merical scheme for periodic structure is to obtain the dynamic stiffness matrix KD of the
periodic cell. This can either be derived using finite element (FE)16 to obtain the equation of

4
motion (EOM) as ordinary differential equation (ODE) in time. Next, Daubechies scaling
function13 approximation is used in time to derive KD in wavelet domain. The method

is described briefly in the Subsection II.B. FE modeling for wave propagation analysis is
however computationally costly, sometime prohibitive. This is because the mesh size should

be small enough to accurately capture the higher modes of vibration participating in wave

propagation. This drawback of FE has motivated researchers to develop alternative nu-


merical schemes for this purpose. The other technique that can be implemented to obtain

KD is one such alternate modeling scheme referred as wavelet based spectral finite element
(WSFE)14,15 . WSFE is tailored especially for wave propagation analysis in finite dimen-

sional structural waveguides of higher complexities and is computationally very efficient.

The method is explained shortly in the Subsection II.C. However, prior to these, a very
brief introduction to Daubechies wavelets is provided in the next Subsection II.A.

A. Daubechies Compactly Supported Wavelets

A concise review of orthogonal basis of Daubechies wavelets13 is provided. Wavelets,


ψj,k (t) forms compactly supported orthonormal basis for L2 (R). The wavelets and the
associated scaling functions ϕj,k (t) are obtained by translation and dilation of single functions

ψ(t) and ϕ(t) respectively.

ψj,k (t) = 2j/2 ψ(2j t − k), j, k ∈ Z (1)

ϕj,k (t) = 2j/2 ϕ(2j t − k), j, k ∈ Z (2)

The scaling functions ϕ(t) are derived from the dilation or scaling equation,
X
ϕ(t) = ak ϕ(2t − k) (3)
k

and the wavelet function ψ(t) is obtained as


X
ψ(t) = (−1)k a1−k ϕ(2t − k) (4)
k

ak are the filter coefficients and they are fixed for specific wavelet or scaling function ba-
sis. For compactly supported wavelets only a finite number of ak are nonzero. The filter

5
coefficients ak are derived by imposing certain constraints on the scaling functions which
are as follows. (1) The area under scaling function is normalized to one, (2) The scaling

function ϕ(t) and its translates are orthonormal and and (3) the wavelet function ψ(t) has
M vanishing moments. The number of vanishing moments M denotes the order N of the

Daubechies wavelet, where N = 2M .

Let Pj (f )(t) be the approximation of a function f (t) in L2 (R) using ϕj,k (t) as basis, at
a certain level (resolution) j, then
X
Pj (f )(t) = cj,k ϕj,k (t), k ∈ Z (5)
k

where, cj,k are the approximation coefficients.

B. Dynamic Stiffness Matrix From FE

FE modeling for structural dynamics problem generally provide a ODE involving the
mass M and stiffness matrices K relating the displacement vector u containing the nodal
degrees of freedom which are be arbitrarily referred as u. Similarly F is the force vector

containing nodal forces F . For an undamped structure, the ODE representing the EOM is

obtained as,

Mü + Ku = F (6)

Here, M and K represent the mass and stiffness matrices of the periodic cell. Let each

of the displacements u(t) of the displacement vector u (and similarly, force F (t) of the
force vector F) be considered to be discretized at n points in the time window [0 tf ]. Let

τ = 0, 1, 2 · · · , n − 1 be the sampling points, then

t = ∆tτ (7)

where, ∆t is the time interval between two sampling points. The function u(t) (similarly,
F (t)) can be approximated or transformed by scaling function ϕ(τ ) at an arbitrary scale as,
X
u(t) = u(τ ) = ûk ϕ(τ − k), k∈Z (8)
k

6
where, ûk are the approximation coefficients. These approximation coefficients can be de-
rived by multiplying both sides of Eqn. 8 by the translates of scaling function ϕ(τ − j).

Next, using their orthogonality properties, we get,


Z
ûj = u(τ )ϕ(τ − j) (9)

Similarly, considering acceleration ü(t) as an independent variable v(t), it can be written as,

X X
v(t) = ü(t) = vk ϕ(τ − k) = uk ϕ00 (τ − k) (10)
k k

Again, multiplying both sides by translates of ϕ(τ − k) and taking inner product we get,

Z j+N −2
X
v̂j = v(τ )ϕ(τ − j) = ûk Ω2j−k (11)
k=j−N +2

where, N is the order of Daubechies scaling function as discussed earlier. Ω2j−k is the second
order connection coefficients defined as,
Z
Ω2j−k = ϕ00 (τ − k)ϕ(τ − j)dτ (12)

For compactly supported wavelets, Ω2j−k (τ − k) and also the other order of derivatives17 are

non-zero only in the interval k = j − N + 2 to k = j + N − 2. While dealing with finite length

data sequence, problems arises at the boundaries. It can be observed from the Eqn. 10, that

certain coefficients ûj near the vicinity of the boundaries (j = 0 and j = n − 1) lies outside

the time window defined by j = 0, 1, 2, · · · , n − 1. Several approaches have been developed

to handle these boundary conditions. Here, a wavelet extrapolation technique proposed

by Amaratunga and Williams18 is implemented. The details of the implementation is also


presented in previous paper15 by the author. This boundary treatment technique is found

effective for time integration using wavelets.

After the treatment of the boundaries, the Eqn. 10, can be written as a matrix equation

of the form,

{v̂j } = Γ2 {ûj } (13)

7
where, Γ2 is the connection coefficient matrix obtained after imposing the boundary condi-
tions. Through similar derivation, it can be shown that,

2
Γ2 = Γ1 (14)

where, Γ1 is the connection coefficient matrix corresponding to first order derivative. Next,
the coupled matrix equation given by Eqn. 13 can be decoupled through eigenvalue decom-

position of the matrix Γ2 as,


2
Γ1 = Γ2 = ΦΠΦ−1 (15)

where, Φ is the eigenvector matrix and Π is the diagonal eigenvalue matrix containing the

eigenvalues which can be arbitrarily referred as −λ2j , −iλj being the eigenvalues of Γ1 .

Substituting Eqn. 15 into Eqn. 13, we get,

˜ j = −λ2j ũj
ṽj = ü (16)

where,

ũj = Φ−1 ûj and ṽj = Φ−1 v̂j (17)

Now, multiplying both sides of Eqn. 6 by translates of scaling functions and taking the
inner product, we get
˜ j + Kũj = F̃j
Mü (18)

Substituting Eqn. 17 in Eqn. 18, it can be written as,

−λ2j Mũj + Kũj = F̃j (19)

or,

(K − λ2j M)ũj = F̃j , j = 0, 1, 2, · · · , n − 1

KDj ũj = F̃j (20)

Here, λj represents the circular frequency ωj used in conventional dynamic stiffness matrix.
λj matches ωj exactly, but only up to a certain fraction pN of the Nyquist frequency ωnyq 14 .

8
C. Wavelet Spectral Finite Element

In the previous Subsection the details of formulating KDj from FE in wavelet domain has
been described. As mentioned earlier, KDj can also be obtained using WSFE method. In

this Subsection, the method14,15 is described briefly and in a generalized way considering the
example of an isotropic Timoshenko beam. Here, KDj is obtained directly from the governing

elasto-dynamic equations of the structure. The governing equations of a Timoshenko beam

is given as,
· 2 ¸
∂ w ∂φ ∂2w
GA − = ρA (21)
∂x2 ∂x ∂t2
· ¸
∂ 2φ ∂w ∂ 2φ
EI 2 + GA − φ = ρI 2 (22)
∂x ∂x ∂t
where, GA and EI are the shear and bending stiffness. ρA and ρI are the corresponding

inertias. w(x, t) and φ(x, t) are the transverse displacement and rotation respectively. The

force boundary conditions are given as,


· ¸
∂w
GA −φ =V (23)
∂x
∂φ
EI =M (24)
∂x
where, V and M are shear force and bending moment respectively.
Even here, first, the variables w(x, t) and φ(x, t) are approximated using ϕ(t) similar to
Eqn. 8 as,
X X
w(x, t) = w(x, τ ) = ŵk (x)ϕ(τ − k) and φ(x, t) = φ(x, τ ) = φ̂k (x)ϕ(τ − k) (25)
k k

Here, however, the approximation coefficients ŵk and φ̂k are function of spatial location
x. Next, Eqns. 25 are substituted in Eqns. 21 and 22. Following this, both sides of the

equations are multiplied by the translates of ϕ(t) and taking inner products on either sides

give,
" # j−N +2
d2 ŵj dφ̂j X
GA − = ρA ŵj Ω2j−k (26)
dx2 dx k=j+N −2
· ¸ j−N +2
X
d2 φ̂j dŵj
EI 2 + GA − φ̂j = ρI φ̂j Ω2j−k (27)
dx dx k=j+N −2

9
The above equations can be written in matrix form as,
"½ ¾ ( )#
d2 ŵ dφ̂ 2
GA 2
− = ρAΓ1 {ŵ} (28)
dx dx
( ) ·½ ¾ ¸
d2 φ̂ dŵ 2
EI 2
+ GA − {φ̂} = ρIΓ1 {φ̂} (29)
dx dx

These matrix equations can be decoupled through eigenvalue analysis as described earlier

and the decoupled form of Eqns. 28 and 29 are of the form,


" #
d2 w̃j dφ̃j
GA − = −λ2j ρAw̃j (30)
dx2 dx
· ¸
d2 φ̃j dw̃j
EI 2 + GA − φ̃j = −λ2j ρI φ̃j (31)
dx dx

Similarly, the transformed and decoupled form of the force boundary conditions given by
Eqns. 23 and 24 can be written as,
· ¸
dw̃j
GA − φ̃j = Ṽj (32)
dx
dφ̃j
EI = M̃j (33)
dx

The solutions of the decoupled ODEs given by Eqns. 30 and 31 can be assumed as,
4
X l
w̃j = Cjl Rj1l e−ikj x (34)
l=1
4
X l
φ̃j = Cjl Rj2l e−ikj x (35)
l=1

Substitution of the above assumed solution in Eqns. 30 and 31 results in a polynomial

eigenvalue problem (PEP), solution of which gives Rj1l and Rj2l as eigenvectors and kjl as the

eigenvalues. The constants Cjl are unknown and obtained from the displacement and force
(Eqns. 32 and 33) boundary conditions at two nodes of the WSFE defined by x = 0 and x = L

with L being the length of the element. Next, following the procedure of FE, the stiffness

matrix relating the transformed nodal displacements to transformed nodal forces given by
Eqn. 20 is obtained. This stiffness matrix is similar to the dynamic stiffness matrix KDj in

wavelet domain derived from FE model as described in the previous Subsection. However,
since WSFE is based on exact solution of the governing differential equation the transformed

10
wavelet domain, one element is sufficient to model waveguides of arbitrary length in absence
of discontinuity. This make the simulations computationally efficient unlike FE solution

where large number of elements are required to accurately predict the wave responses of
a structure. For a structural assembly, the elemental KDj obtained from WSFE can be

assembled to form the global dynamic stiffness matrices similar to FE.

III. MATHEMATICAL FORMULATION

In this section, the formulation to obtain the wave response of a periodic structure

utilizing its periodic property is explained in detail. The formulation proceeds with the
dynamic stiffness matrix KDj for the periodic cell. KDj relates the transformed nodal

displacements ũj and forces F̃j . The subscript j is dropped hereafter for simplified notations

and all the following equations should be evaluated for j = 0, 1, 2, · · · , n − 1.

The elements of vector ũ can be grouped into ũL , ũR and ũI . Here, ũL , ũR and

ũI respectively represent the displacements of LHS, RHS and intermediate nodes of the

periodic structure. Similarly, the force vector F̃ is also clustered into F̃L , F˜R and F̃I . For a
periodic structure with intermediate nodes, the displacements corresponding to them (ũI )
are condensed out such that ũ contains only ũL and ũR which are of length M each. Thus,

for any arbitrary periodic cell, the Eqn. 20 can be written as,
    
   
 KD
LL
KD   ũL   F̃L 
LR

  = (36)
RL RR 
 
 
 F̃R 
KD KD ũR

For a periodic structure ũL and ũR are related as,

ũR = e−ξd ũL = e−µ ũL (37)

where, d is the length of the periodic cell and µ is generally referred as propagation constant.

Similarly, F̃L and F̃R follows the relation,

F̃R = −e−µ F̃L (38)

11
Substituting Eqns. 37 and 38 into Eqn. 36, we can pose a PEP as,

[KD RL + (KD LL + KD RR )e−µ + KD LR e−2µ ]ũL = 0 (39)

The solution of the eigenvalue problem provides an eigenvector matrix Λ and a diagonal

matrix with the eigenvalues, e−µ as the diagonal terms. Since Eqn. 39 is a quadratic equation,

the eigenvalues will be a set of complex pairs corresponding to ±µR ± iµI .

In addition, due to the periodic nature, Eqn. 37 written for a single periodic cell can be
extended for a periodic structure with Nc number of cells as,

ũR = e−ξNc d ũL = e−µNc ũL (40)

For similar reasons, the generalized solution of the displacement ũNx at LHS boundary of

the Nx th cell of a periodic structure can be written as,


2M−1
X 2M
X
ml −µl Nx l
ũm
Nx = l
CΛ e + C l Λml e−µ (Nc −Nx ) (41)
l=1,3,5... l=2,4,6...

where, ũm
Nx is the m
th
element of the vector ũNx representing the mth degree of freedom.

C l are the unknown constants to be derived from the boundary conditions. The particular
form of the solution shown in Eqn. 41 is considered to avoid ill-conditioning of the matrices.

For a single cell, Nc = 1 in Eqn. 41, with Nx = 0 giving ũL and Nx = 1 representing ũR .
Similarly, for a periodic structure with number of cell equal to Nc , substituting Nx = 0 in

Eqn. 41 gives the displacements ũL at LHS boundary and Nx = Nc gives the displacements

ũR at the RHS boundary. The Eqn. 41 can be written in matrix form as,

ũNx = ΛΘ{C} (42)

where, Θ is a diagonal matrix with the diagonal terms


h i
1 1 2 2 l l 2M 2M
e−µ Nx e−µ (Nc −Nx ) e−µ Nx e−µ (Nc −Nx ) . . . e−µ Nx e−µ (Nc −Nx ) . . . e−µ Nx e−µ (Nc −Nx ) .

Substituting the boundary conditions at Nx = 0 and Nx = Nc in the matrix equation given

by Eqn. 42, we get,

ũL = ΛΘNx =0 {C} = T11 {C} (43)

12
ũR = ΛΘNx =Nc {C} = T12 {C} (44)
   
 ũL 
 
 T11 
ũ = =  {C} = [T1 ] {C} (45)
 ũR 
  T12

Next, substituting Eqn. 37 in the first row of Eqn. 36, we can relate ũL to F̃L as,
£ ¤
KD LL + e−µ KD LR ũL = F̃L (46)

Substituting the generalized solution of the displacements given by Eqn. 41 in the above

equation 46, the force F̃Nx at LHS boundary of the Nx th cell of a periodic structure can be

written as,
2M−1
X X 2M
l l
F̃Nmx = (KD LLmp + e−µ KD LRmp )Λpl C l e−µ Nx
l=1,3,5... p
2M
X 2M
X l l
+ (KD LLmp + e−µ KD LRmp )Λpl C l eµ (Nc −Nx ) (47)
l=2,4,6... p

The above equation can be written in a matrix form as,

F̃Nx = ΥΘ{C} (48)

P2M l
where, Υml = p (KD LLmp + e−µ KD LRmp )Λpl . Thus, similar to the displacements the

force boundary conditions at Nx = 0 and Nx = Nc can be written as,

F̃L = ΥΘNx =0 {C} = T21 {C} (49)

F̃R = ΥΘNx =Nc {C} = T22 {C} (50)


   
 
 F̃L 
 T21 
F̃ = =  {C} = [T2 ] {C} (51)
 F̃R 
  T22

From Eqns. 45 and 51, the transformed force F̃ and displacement ũ can be related as,

F̃ = T2 T1 −1 ũ = Tũ (52)

Knowing the force applied on the LHS and RHS boundaries of the periodic structure, the

displacements at these boundaries can be derived using the above equation. In addition, after
the boundary displacements ũ are known, the constants {C} can be obtained from Eqn. 45.

13
A Ncell = 7 B

FIG. 1. Schematic of periodic simply-supported beam

Y transverse

X axial
Ncell = 6

A B

FIG. 2. Schematic of periodic truss/frame

Substituting these {C} in Eqn. 42 one can obtain the displacement at any intermediate point.
The displacement uNx in time domain can be obtained through the inverse transforms of

ũNx

IV. FREQUENCY DOMAIN ANALYSIS

In this Section, the frequency domain wave characteristics of periodic structures are stud-

ied. The parameter studied is the propagation constant, µ for different periodic structural

configurations. Two examples of periodic structures are considered. First is a periodically

simply-supported beam shown in Fig. 1. Second is a periodic frame structure shown in

Fig. 2. The members of the frame are Euler-Bernoulli beam with three degrees of freedom,
namely axial and transverse displacements, and rotation. In all the cases, the structural

members are considered to be aluminum with E = 70 GPa, ν = 0.3 and ρ = 2700 kg/m3 .

The cross-sectional area is A = 0.05 × 0.01 m2 .

In Figs. 3(a) and (b), the real (µr ) and imaginary (µi ) parts of µ are plotted with
respect to the frequency up to 2.5 kHz for the simply-supported beam. For such mono-

14
1.5 4
Timoshenko Beam Euler−Bernoulli Beam
Euler−Bernoulli Beam Timoshenko Beam

Propagation Constant µ (Imaginary)


1
Propagation Constant µ (Real)

2
0.5

0 0

−0.5

−2
−1

−1.5 −4
0 0.5 1.0 1.5 2.0 2.5 0 0.5 1.0 1.5 2.0 2.5
Frequency (kHz) Frequency (kHz)

(a) (b)

FIG. 3. (a) Real and (b) imaginary parts of the propagation constant (µ) for periodic

simply-supported beam with d = 0.5 m.

coupled periodic structure only one wave mode is present19 as seen in the figure. The length
of each periodic cell i.e, the length of the beam between two supports is 0.5 m. The plots

are obtained and compared for both Euler-Bernoulli and Timoshenko beam theories. The

pattern of µ is similar to that reported in the reference20 for articulated Euler-Bernoulli


beam. As expected, µr shows band gaps within which propagation will not occur. In the

later part of the paper, the overall wave response of the structure due to broadband impulse

load also exhibit this band gap feature. µi also shows similar trend, however, the non-zero
band has wider width as compared to the real part. In addition, since µ has both real

and imaginary parts, the waves are dispersive in nature, i.e they will be attenuated as they
propagate. The comparison between the Euler-Bernoulli and Timoshenko beam shows slight

difference between the two with the Timoshenko beam having a higher propagation speed.

Figs. 4(a) and (b) plots µr and µi for the frame shown in Fig. 2 with d = 0.25 m and the

members as Euler-Bernoulli beam with three degrees of freedom, namely, axial, transverse
and rotational at each nodes. This also represent a multi-coupled system with six degrees

of freedom coupling. The symmetry, however results only in three unique wave modes as
observed from the figure. As mentioned earlier, µ is obtained numerically through solution

15
4

10

Propagation Constant µ, (Imaginary)


2
Propagation Constant (µ)

0 0

−5
−2

−10

−4
0 1.0 2.0 3.0 0 1.0 2.0 3.0
Frequency (kHz) Frequency (kHz)

(a) (b)

FIG. 4. (a) Real and (b) imaginary parts of the propagation constant (µ) for periodic frame

structure of Euler-Bernoulli beam with d = 0.25 m.

of a PEP and hence it is difficult to separate out µ corresponding to each mode. The waves

are however dispersive in nature due to non-negligible imaginary component of µ and exhibit

banded nature.

V. TIME DOMAIN ANALYSIS

Here, the time domain wave response of periodic structures due to broadband impulse

load with different frequency content are simulated using the proposed method. The re-

sponses obtained are also compared with the exact solution. Similar to the frequency domain

analysis performed in the previous section, the examples of periodic simply-supported beam

and truss with Euler-Bernoulli beam members are considered here. The loads are applied

at one end point (A) and the responses are measured at the other end (B) shown in Figs. 1

and 2. Frequency content of the impulse load used is shown in Fig. 5. Similar impulse load

with lower frequency content of 2.2 kHz has also been used for some of the examples.

In Fig. 6(a), the time domain responses of a periodic simply-supported beam with d =

0.25 m due to impulse load of frequency content 4.4 kHz are plotted. The periodic simply-
supported beam is shown in Fig. 1 and has Ncell = 20. The load is applied at A and the

16
0 2 4 6 8 10 12
Frequency (kHz)

FIG. 5. Frequency content (4.4 kHz) of the applied impulse load.

2
0.4 10
Periodic Periodic
Exact Exact

1
0.2 10
Rotational Velocity (rad/s)

0
0 10

−1
−0.2 10

−2
−0.4 10
0.004 0.008 0.012 0.016 0.02 0 0.5 1.0 1.5 2.0 2.5 3.0
Time (sec) Frequency (kHz)

FIG. 6. (a) Time history and (b) FRFs of the response of periodic simply-supported Euler-

Bernoulli beam with d = 0.25 m & Ncell = 20 for broadband impulse load with frequency

content 4.4 kHz

response is measured at point B. The response obtained considering periodicity is compared

with the exact solution obtained using WSFE or FE methods. The periodic and exact

solutions matches exactly and the corresponding frequency response functions (FRFs) are
presented in Fig. 6(b) for frequency range 0 to 3.0 kHz. The FRFs show three band gaps or

stop bands between 0 to 3.0 kHz. This is in congruence with that predicted by the dispersion
relation obtained from the frequency domain analysis.

17
0
0.8 10

−1
0.4 10
Transverse Velocity (mm/s)

−2
0 10

−3
−0.4 10

−4
−0.8 Periodic 10
Exact

−5
−1.2 10
0 0.004 0.008 0.012 0.016 0.02 0 0.5 1.0 1.5 2.0 2.5
Time (sec) Frequency (kHz)

(a) (b)

FIG. 7. (a) Time history and (b) FRF of the transverse wave response of periodic frame

structure for broadband load of 2.2 kHz

In conclusion, it can be said that for periodically simply-supported beam, the time
domain impulse response obtained through consideration of periodicity matches the exact

solution. The above mentioned trends are observed for the periodic beam structure irre-

spective of the number of cells, Ncell , the length of each cell d.

Next, the time domain wave responses of periodic truss structure with Ncell = 10 and
d = 0.125 m shown in Fig. 2 are studied. Each of the members of the truss is an Euler-
Bernoulli beam with three degrees of freedom as mentioned earlier. The impulse with

frequency content of 2.2 kHz is applied at A and the wave responses are measured at B.
The periodic responses are compared with the exact response.

In Fig. 7(a), the transverse velocity i.e the velocity in Y −direction is plotted. Even

here, it can be seen that the periodic solution matches exactly with the exact solution. This

is similar to the example of simply-supported periodic beam. Other examples of periodic

structures with or without kinematic constraints have been studied and it has been observed
that the periodic formulation gives the exact solution for structures. The corresponding

FRFs are presented in Fig. 7(b) for frequency between 0 to 3.0 kHz. Here, two stop bands
are observed between approximately 0.9 to 1.4 kHz and 2.15 to 2.4 kHz.

18
VI. CONCLUSIONS

Here, a generalized Daubechies wavelet based numerical tool is developed for wave prop-

agation analysis in one-dimensional multi-coupled periodic structure. The method described

starts from the dynamic stiffness matrix of the periodic cell and can simulate time and fre-
quency domain responses of the structure using the periodicity condition. In this paper,

the dynamic stiffness matric is obtained using wavelet spectral finite element formulation.

Similar matrices can be obtained using finite element method and wavelet transform. The

necessity of using wavelet transform in obtaining the dynamic stiffness matrix of the cell and

also modeling of the periodic structure is that it allows simultaneous frequency and time
domain analysis unlike most of the other techniques available.

The proposed method is implemented to study wave behavior in periodically simply-


supported beam and periodic frame structures. In all the cases, the dispersion relation are

obtained and these plots show the band gap nature of periodic structure. The dispersion plot
of periodically simply-supported beam is similar to that reported in the literature. Next,

the time domain responses due to broadband impulse loads are simulated for the beam and

the frame structures. The response of the beam and the frame structure obtained from the
periodic and exact solutions are same. The technique will have important application in

homogenization of periodic structures which will be studied in future work.

References

1
L. Brillouin, “Wave propagation in periodic structures”, Dover Publications, New York

(1953).
2
D. J. Mead, “Free wave propagation in periodically supported, infinite beams”, Jl. Sound
Vib. 11, 181–197 (1970).
3
G. Sengupta, “Natural flexural waves and the normal modes of periodically supported
beams and plates”, Jl. Sound Vib. 13, 89–101 (1970).

19
4
D. Li and H. Benaroya, “Dynamics of periodic and near periodic structures”, Appl. Mech.
Rev. 45, 447–459 (1992).
5
R. S. Langley, “The response of two dimensional periodic structures to point harmonic
forcing”, Jl. Sound Vib. 197, 325–348 (1996).
6
D. J. Mead, “The forced vibration of one-dimensional multi-coupled periodic structures:

An application to finite element analysis”, Jl. Sound Vib. 319, 282–304 (2009).
7
S. Gonella and M. Ruzzene, “Homogenization of vibrating periodic lattice structures”,

App. Math. Modeling 32, 459–482 (2008).


8
S. Gonella and M. Ruzzene, “Homogenization and equivalent in-plane properties of two-

dimensional periodic lattices”, Int. Jl. Solids Struct. 45, 2897–2915 (2008).
9
M. Ruzzene and L. Airoldi, “Broad-band vibration attenuation in plates with periodic
arrays of shunted piezoelectric patches”, Jl. Acoust. Soc. Am. 123, 3573 (2008).
10
S. Asiri, A. Baz, and D. Pines, “Active periodic struts for a gearbox support system”,

Smart Mat. Struct. 15, 1707–1714 (2006).


11
M. Hofer, N. Finger, G. Kovacs, J. Schoberl, S. Zaglmayr, U. Langer, and R. Lerch, “Fi-

nite element simulation of wave propagation in periodic piezoelectric SAW structures”,


IEEE Trans. Ultrasonics Ferroelectrics and Frequency Control 53, 1192–1201 (2006).
12
S. Ballandras, M. Wilm, P. F. Edoa, A. Soufyane, V. Laude, W. Steichen, and R. Lardat,

“Finite element analysis of periodic piezoelectric transducers”, Jl. Appl. Phys. 93, 702

(2003).
13
I. Daubechies, “Ten lectures on wavelets”, CBMS-NSF Series on in Applied Mathematics,

SIAM, Philadelphia (1992).


14
M. Mitra and S. Gopalakrishnan, “Extraction of wave characteristics from wavelet based

spectral finite element formulation”, Mechanical Sys. Sig. Proc. 40, 2046–2079 (2006).
15
M. Mitra and S. Gopalakrishnan, “Spectrally formulated wavelet finite element for wave
propagation and impact force identification in 1-D waveguides”, Int. Jl. Solid Struct. 42,

4695–4721 (2005).
16
R. D. Cook, D. S. Malkus, M. E. Plesha, and R. J. Witl, “Concepts and applications of

20
finite element analysis”, John Wiley & Sons (ASIA), Singapore (2007).
17
G. Beylkin, “On the representation of operators in bases of compactly supported

wavelets”, SIAM Jl. Numer. Anal. 6, 1716–1740 (1992).


18
K. Amaratunga and J. R. Williams, “Wavelet-Galerkin solution of boundary value prob-

lems”, Arch. Comp. Meth. Eng. 4, 243–285 (1997).


19
D. J. Mead, “Wave propagation and natural modes in periodic system. I Mono-coupled
systems”, Jl. Sound Vib. 40, 1–18 (1975).
20
Y. Watanabe and N. Sugimoto, “Flexural wave propagation in a spatially periodic struc-
ture of articulated beams”, Wave Motion 42, 155–167 (2005).

21
List of Figures

FIG. 1 Schematic of periodic simply-supported beam . . . . . . . . . . . . . . . . . 14


FIG. 2 Schematic of periodic truss/frame . . . . . . . . . . . . . . . . . . . . . . . . 14

FIG. 3 (a) Real and (b) imaginary parts of the propagation constant (µ) for periodic
simply-supported beam with d = 0.5 m. . . . . . . . . . . . . . . . . . . . . 15

FIG. 4 (a) Real and (b) imaginary parts of the propagation constant (µ) for periodic

frame structure of Euler-Bernoulli beam with d = 0.25 m. . . . . . . . . . . 16

FIG. 5 Frequency content (4.4 kHz) of the applied impulse load. . . . . . . . . . . . 17

FIG. 6 (a) Time history and (b) FRFs of the response of periodic simply-supported

Euler-Bernoulli beam with d = 0.25 m & Ncell = 20 for broadband impulse


load with frequency content 4.4 kHz . . . . . . . . . . . . . . . . . . . . . . 17

FIG. 7 (a) Time history and (b) FRF of the transverse wave response of periodic

frame structure for broadband load of 2.2 kHz . . . . . . . . . . . . . . . . . 18

22
A Ncell = 7 B

d
Y transverse

X axial
Ncell = 6

A B

d
1.5
Timoshenko Beam
Euler−Bernoulli Beam
1
Propagation Constant µ (Real)

0.5

−0.5

−1

−1.5
0 0.5 1.0 1.5 2.0 2.5
Frequency (kHz)
4
Euler−Bernoulli Beam
Timoshenko Beam
Propagation Constant µ (Imaginary)

−2

−4
0 0.5 1.0 1.5 2.0 2.5
Frequency (kHz)
10
Propagation Constant (µ)

−5

−10

0 1.0 2.0 3.0


Frequency (kHz)
Propagation Constant µ, (Imaginary)

0
2
4

−4
−2

0
1.0
2.0
Frequency (kHz)
3.0
0 2 4 6 8 10 12
Frequency (kHz)
0.4
Periodic
Exact

0.2
Rotational Velocity (rad/s)

−0.2

−0.4
0.004 0.008 0.012 0.016 0.02
Time (sec)
2
10
Periodic
Exact

1
10

0
10

−1
10

−2
10
0 0.5 1.0 1.5 2.0 2.5 3.0
Frequency (kHz)
0.8

0.4
Transverse Velocity (mm/s)

−0.4

−0.8 Periodic
Exact

−1.2
0 0.004 0.008 0.012 0.016 0.02
Time (sec)
0
10

−1
10

−2
10

−3
10

−4
10

−5
10
0 0.5 1.0 1.5 2.0 2.5
Frequency (kHz)