dynamic anisotropy and primitive waveforms in discrete

28
Dynamic anisotropy and primitive waveforms in discrete elastic systems D.J. Colquitt 1 I.S. Jones 2 A.B. Movchan 1 N.V. Movchan 1 R.C. McPhedran 3 1 Department of Mathematical Sciences University of Liverpool 2 School of Engineering John Moores University 3 CUDOS, School of Physics University of Sydney British Applied Mathematics Colloquium 2012 DJC gratefully acknowledges the support of an EPSRC research studentship (EP/H018514/1). ABM and NVM acknowledge the financial support of the European Community’s Seven Framework Programme under contract number PIAPP-GA-284544-PARM-2. RCM acknowledges the support of the ARC through its Discovery Grants Scheme. Colquitt et al. | British Applied Mathematics Colloquium 2012 1 / 24

Upload: others

Post on 17-May-2022

5 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Dynamic anisotropy and primitive waveforms in discrete

Dynamic anisotropy and primitive waveforms in discreteelastic systems

D.J. Colquitt1 I.S. Jones2 A.B. Movchan1 N.V. Movchan1

R.C. McPhedran3

1Department of Mathematical SciencesUniversity of Liverpool

2School of EngineeringJohn Moores University

3CUDOS, School of Physics

University of Sydney

British Applied Mathematics Colloquium 2012

DJC gratefully acknowledges the support of an EPSRC research studentship (EP/H018514/1). ABM and NVM acknowledgethe financial support of the European Community’s Seven Framework Programme under contract number

PIAPP-GA-284544-PARM-2. RCM acknowledges the support of the ARC through its Discovery Grants Scheme.

Colquitt et al. | British Applied Mathematics Colloquium 2012 1 / 24

Page 2: Dynamic anisotropy and primitive waveforms in discrete

Outline

1 Introduction

2 A uniform (scalar) square latticeGreen’s function & primitive wave forms

3 Triangular latticesScalar problem

Green’s function & primitive waveforms

Vector problem - planar elasticityGoverning EquationsDispersive propertiesPrimitive Waveforms

Colquitt et al. | British Applied Mathematics Colloquium 2012 2 / 24

Page 3: Dynamic anisotropy and primitive waveforms in discrete

| Introduction

Existing Literature

Dispersion & design of structures for controlling stop bands inconcentrated mass systems

P. Martinsson & A. Movchan, QJMAM 56 (2003), pp. 45–64.

Primitive waveforms in scalar lattices

Ayzenberg-Stepanenko & Slepyan, J Sound Vib 313 (2008), pp812–821.Osharovich et al., Continuum Mech Therm 22 (2010), pp. 599–616.Langley, J Sound Vib 197 (1996), pp. 447–469.Langley, J Sound Vib 201 (1997), pp. 235–253.

Dynamic homogenization of periodic media

Craster et al., QJMAM 63 (2010), pp.497–519.Craster et al., Proc R Soc A 466 (2010), pp. 2341–2362.Craster et al., JOSA A (2011), pp. 1032–1040.

Colquitt et al. | British Applied Mathematics Colloquium 2012 3 / 24

Page 4: Dynamic anisotropy and primitive waveforms in discrete

| A uniform (scalar) square lattice

A uniform square lattice

Label each node by n ∈ Z2

Introduce counting vectorsei = [δ1i , δ2i ]

T .

Lattice vectors: t1 = [`, 0]T ,t2 = [0, `]T .

Translation matrixT = [t1, t2].

Position of nth node:x = n⊗ T .

Colquitt et al. | British Applied Mathematics Colloquium 2012 4 / 24

Page 5: Dynamic anisotropy and primitive waveforms in discrete

| A uniform (scalar) square lattice

A uniform square lattice

(n)

t1

t2

Label each node by n ∈ Z2

Introduce counting vectorsei = [δ1i , δ2i ]

T .

Lattice vectors: t1 = [`, 0]T ,t2 = [0, `]T .

Translation matrixT = [t1, t2].

Position of nth node:x = n⊗ T .

Colquitt et al. | British Applied Mathematics Colloquium 2012 4 / 24

Page 6: Dynamic anisotropy and primitive waveforms in discrete

| A uniform (scalar) square lattice

A uniform square lattice

(n)

t1

t2

(n− e1) (n + e1)

(n + e1 + e2)

(n− e1 − e2)

(n− e2)

(n + e2)

(n− e1 − e2)

(n− e1 + e2)Label each node by n ∈ Z2

Introduce counting vectorsei = [δ1i , δ2i ]

T .

Lattice vectors: t1 = [`, 0]T ,t2 = [0, `]T .

Translation matrixT = [t1, t2].

Position of nth node:x = n⊗ T .

Colquitt et al. | British Applied Mathematics Colloquium 2012 4 / 24

Page 7: Dynamic anisotropy and primitive waveforms in discrete

| A uniform (scalar) square lattice | Green’s function & primitive wave forms

Uniform Square LatticeGreen’s function for time-haromic out-of-plane shear

Balance of linear momentum at node n connected to set of nodesP(n) = {n± e1,n± e2} (for time-harmonic deformations)∑

p∈P(n)

u(p)− |P(n)|u(n) + ω2u(n) = δm0δn0

Discrete Fourier transform:

F : u(m) 7→ U(ξ) =∑m∈Z2

u(m) exp {−iξ · (n⊗ T )}

Inverse transform:

Green’s Function

u(n) =1

π2

∫ π

0

∫ π

0

cos(n1ξ1) cos(n2ξ2)

ω2 − 4 + 2(cos ξ1 + cos ξ2)dξ1dξ2.

Colquitt et al. | British Applied Mathematics Colloquium 2012 5 / 24

Page 8: Dynamic anisotropy and primitive waveforms in discrete

| A uniform (scalar) square lattice | Green’s function & primitive wave forms

Primitive Wave Forms at a Saddle Point

Dispersion Diagram

−2

0

2

−2

0

2

0

0.5

1

1.5

2

2.5

3

ξ1

ξ2

ω

−A−B

BA

σ(ω, ξ) = 0

Slowness Contour

ξ1

ξ 2−3 −2 −1 0 1 2 3

−3

−2

−1

0

1

2

3

σ(2, ξ) = 0

Colquitt et al. | British Applied Mathematics Colloquium 2012 6 / 24

Page 9: Dynamic anisotropy and primitive waveforms in discrete

| A uniform (scalar) square lattice | Green’s function & primitive wave forms

Two stationary points of different kinds

−10 −5 0 5 10

−10

−8

−6

−4

−2

0

2

4

6

8

10

m

n

ω = 2

−10 −5 0 5 10

−10

−5

0

5

10

m

n

ω = 2√

2

Colquitt et al. | British Applied Mathematics Colloquium 2012 7 / 24

Page 10: Dynamic anisotropy and primitive waveforms in discrete

| Triangular lattices

Uniform Triangular Lattice

Colquitt et al. | British Applied Mathematics Colloquium 2012 8 / 24

Page 11: Dynamic anisotropy and primitive waveforms in discrete

| Triangular lattices

Uniform Triangular Lattice

(n)t2

t1

cell n = [n1, n2]T ∈ Z2

x(n) = n⊗ T , T = [t1, t2] , t1 = [`, 0]T , t2 = `[1,√

3]T/2

ei = [δ1i , δ2i ]T

Colquitt et al. | British Applied Mathematics Colloquium 2012 8 / 24

Page 12: Dynamic anisotropy and primitive waveforms in discrete

| Triangular lattices

Uniform Triangular Lattice

(n)t2

t1

(n + e1)(n− 2e1) (n− e1) (n + 2e1) (n + 3e1)

(n + e2) (n + e1 + e2)(n− 2e1 + e2)

(n− e1 + e2) (n + 2e1 + e2)

(n− e2)

(n + e1 − e2)

(n− 2e1 − e2)

(n− e1 − e2)

(n + 2e1 − e2, 1)

(n + 3e1 − e2, 2)

cell n = [n1, n2]T ∈ Z2

x(n) = n⊗ T , T = [t1, t2] , t1 = [`, 0]T , t2 = `[1,√

3]T/2

ei = [δ1i , δ2i ]T

Colquitt et al. | British Applied Mathematics Colloquium 2012 8 / 24

Page 13: Dynamic anisotropy and primitive waveforms in discrete

| Triangular lattices | Scalar problem

Uniform Triangular LatticeDispersive properties

Fourier Transform:[ω2 − 6 + 2 cos ξ1 + 4 cos(ξ1/2) cos(ξ2

√3/2)

]︸ ︷︷ ︸

σ(ω,ξ)

U(ξ) = 1,

−20

2

−5

0

50

1

2

3

ξ1

ξ2

ω

AB

C −A

−C

−B

σ(ω, ξ) = 0

−3 −2 −1 0 1 2 3

−3

−2

−1

0

1

2

3

ξ1ξ 2

σ(2√

2, ξ) = 0

Colquitt et al. | British Applied Mathematics Colloquium 2012 9 / 24

Page 14: Dynamic anisotropy and primitive waveforms in discrete

| Triangular lattices | Scalar problem | Green’s function & primitive waveforms

Uniform Triangular LatticeGreen’s function for time-harmonic out-of-plane shear

u(n) =

√3

4π2

2π∫0

dξ1

2π/√3∫

0

dξ2cos[(n1 + n2/2)ξ1] cos[n2ξ2

√3/2]

σ(ω, ξ)

−15 −10 −5 0 5 10 15−10

−5

0

5

10

x1

x 2

Colquitt et al. | British Applied Mathematics Colloquium 2012 10 / 24

Page 15: Dynamic anisotropy and primitive waveforms in discrete

| Triangular lattices | Vector problem - planar elasticity | Governing Equations

In-plane vector elasticity: Equilibrium equationsDisplacement amplitude field for time-harmonic waves

u(1) =[u(1)1 , u

(1)2 , θ(1)

]T

u(2) =[u(2)1 , u

(2)2 , θ(2)

]TF(i)1 = ES∂xu1(x),

F(i)2 = −EI∂3xxxu2(x),

F(i)3 = −EI∂2xxu2(x).

The axial deformation satisfies

(∂2xx − ω2ρ`/(ES)u1(x) = 0,

with u1(0) = u(1)1 , u1(`) = u

(2)1 .

The flexural deformation solves

[∂4xx − ω2ρS/(EI )]u2(x) = 0,

with u2(0) = u(1)2 , u2(`) = u

(2)2 ,

∂2xxu2|x=0 = θ(1), ∂2xxu2|x=` = θ(2).

Inclusion of flexural deformationallows for micropolar rotations.

Colquitt et al. | British Applied Mathematics Colloquium 2012 11 / 24

Page 16: Dynamic anisotropy and primitive waveforms in discrete

| Triangular lattices | Vector problem - planar elasticity | Governing Equations

F(1) =

A︷ ︸︸ ︷a11 0 00 a22 a230 a23 a33

u

(1)1

u(1)2

θ(1)

+

B︷ ︸︸ ︷b11 0 00 b22 b230 −b23 b33

u

(2)1

u(2)2

θ(2)

a11 = η cot η

a22 = βλ3 (coshλ sinλ− cosλ sinhλ) / [2 (1− cosλ coshλ)]

a23 = βλ2 sinλ sinhλ)/ [2 (1− cosλ coshλ)]

a33 = βλ (sinλ coshλ− cosλ sinhλ))/ [2 (1− cosλ coshλ)]

b11 = −η csc η

b23 = βλ2 (coshλ− cosλ) / [2 (1− cosλ coshλ)]

b33 = βλ (sinλ− sinhλ) / [2 (cosλ coshλ− 1)]

with β = 2EI/(S`2), η = ω√ρ, λ4 = 2ω2ρ/β.

Colquitt et al. | British Applied Mathematics Colloquium 2012 12 / 24

Page 17: Dynamic anisotropy and primitive waveforms in discrete

| Triangular lattices | Vector problem - planar elasticity | Governing Equations

Balance of MomentumUniform elastic triangular lattice

1, J1

Nodal displacement in the nth cell: u(n)

Introduce the matrices:

A(j) = R(j)TAR(j) and B(j) = R(j)TBR(j)

R(j) - matrix of rotation by angle jπ/3Equations of motion

ω2u(n) = B(0)u(n + e1) + B(1)u(n + e2) + B(2)u(n− e1 + e2)

+ B(3)u(n− e1) + B(4)u(n− e2)

+ B(5)u(n + e1 − e2) +∑

0≤j≤5A(j)u(n)

Colquitt et al. | British Applied Mathematics Colloquium 2012 13 / 24

Page 18: Dynamic anisotropy and primitive waveforms in discrete

| Triangular lattices | Vector problem - planar elasticity | Dispersive properties

Dispersion Equation

Fourier transformed equation

ω2U = σ(ω; ξ)U

The dispersion equation is then det[σ(ω, ξ)− ω2I

]= 0.

−2

0

2

−5

0

5

0

200

400

600

800

ξ1

ξ2

f [H

z]

Colquitt et al. | British Applied Mathematics Colloquium 2012 14 / 24

Page 19: Dynamic anisotropy and primitive waveforms in discrete

| Triangular lattices | Vector problem - planar elasticity | Dispersive properties

Slowness Contours

Isotropic for small ω, as expected.

Characteristic hexagonal contour at 323Hz.

Saddle point & intersection of slownesscontours.

Further saddle point at 615.8 Hz.

“Switching” of preferential directionsunique feature of elastic lattice.

Colquitt et al. | British Applied Mathematics Colloquium 2012 15 / 24

Page 20: Dynamic anisotropy and primitive waveforms in discrete

| Triangular lattices | Vector problem - planar elasticity | Primitive Waveforms

−2 0 2

−5

−4

−3

−2

−1

0

1

2

3

4

5

ξ1

ξ 2

f = 323 Hz

Saddle Point

−2 0 2

−5

−4

−3

−2

−1

0

1

2

3

4

5

ξ1

ξ 2

f = 428.6 Hz

Saddle Point

−2 0 2

−5

−4

−3

−2

−1

0

1

2

3

4

5

ξ1

ξ 2

f = 615.8 Hz

Preferential directionsColquitt et al. | British Applied Mathematics Colloquium 2012 16 / 24

Page 21: Dynamic anisotropy and primitive waveforms in discrete

| Triangular lattices | Vector problem - planar elasticity | Primitive Waveforms

Forcing Frequency of 323 Hz (Pass Band)Horizontal Forcing

F

−2 0 2

−5

−4

−3

−2

−1

0

1

2

3

4

5

ξ1

ξ 2

Colquitt et al. | British Applied Mathematics Colloquium 2012 17 / 24

Page 22: Dynamic anisotropy and primitive waveforms in discrete

| Triangular lattices | Vector problem - planar elasticity | Primitive Waveforms

Forcing Frequency of 323 Hz (Pass Band)Vertical Forcing

F

−2 0 2

−5

−4

−3

−2

−1

0

1

2

3

4

5

ξ1

ξ 2

Colquitt et al. | British Applied Mathematics Colloquium 2012 18 / 24

Page 23: Dynamic anisotropy and primitive waveforms in discrete

| Triangular lattices | Vector problem - planar elasticity | Primitive Waveforms

Forcing Frequency of 323 Hz (Pass Band)Concentrated Moment

F

−2 0 2

−5

−4

−3

−2

−1

0

1

2

3

4

5

ξ1

ξ 2

Colquitt et al. | British Applied Mathematics Colloquium 2012 19 / 24

Page 24: Dynamic anisotropy and primitive waveforms in discrete

| Triangular lattices | Vector problem - planar elasticity | Primitive Waveforms

Forcing Frequency of 615.8 Hz (Saddle Point)Horizontal Forcing

F

−2 0 2

−5

−4

−3

−2

−1

0

1

2

3

4

5

ξ1

ξ 2

Colquitt et al. | British Applied Mathematics Colloquium 2012 20 / 24

Page 25: Dynamic anisotropy and primitive waveforms in discrete

| Triangular lattices | Vector problem - planar elasticity | Primitive Waveforms

Forcing Frequency of 615.8 Hz (Saddle Point)Vertical Forcing

F

−2 0 2

−5

−4

−3

−2

−1

0

1

2

3

4

5

ξ1

ξ 2

Colquitt et al. | British Applied Mathematics Colloquium 2012 21 / 24

Page 26: Dynamic anisotropy and primitive waveforms in discrete

| Triangular lattices | Vector problem - planar elasticity | Primitive Waveforms

Forcing Frequency of 615.8 Hz (Saddle Point)Concentrated Moment

F

−2 0 2

−5

−4

−3

−2

−1

0

1

2

3

4

5

ξ1

ξ 2

Colquitt et al. | British Applied Mathematics Colloquium 2012 22 / 24

Page 27: Dynamic anisotropy and primitive waveforms in discrete

| Triangular lattices | Vector problem - planar elasticity | Primitive Waveforms

Forcing Frequency of 428.6 Hz (Saddle Point)

Horizontal Forcing Vertical Forcing Concentrated Moment

Shape of primitive waveform determined by slowness contour

Existence of primitive waveforms not associated with saddle points asin scalar case (Ayzenberg-Stepanenko et al.)

Frequency dependent “switching” of waveform orientations formonotonic lattice is a novel feature of elastic lattice

Elastic lattice “allows selection” of dominant orientation viatype/orientation of applied forcing

Colquitt et al. | British Applied Mathematics Colloquium 2012 23 / 24

Page 28: Dynamic anisotropy and primitive waveforms in discrete

| Summary

Summary

Statically isotropic lattices can exhibit very strong dynamic anisotropy

Qualitative behaviour of fundamental solution can be predicted fromthe Bloch-Floquet problem

Established the presence of primitive waveforms in elastic latticespreviously only demonstrated in scalar lattice

Offered an explanation for these effects via analysis of slownesscontours

Established the importance of the slowness contours in these primitivewaveforms

Inclusion of micropolar interactions allows for additional primitivewaveforms

Additional details can be found inColquitt et al. Dynamic anisotropy & localization in elastic latticesystems. Waves in Random & Complex Media. To appear (doi:10.1080/17455030.2011.633940).

Colquitt et al. | British Applied Mathematics Colloquium 2012 24 / 24