Applied Mathematical Modelling 39 (2015) 72447265

Contents lists available at ScienceDirect

Applied Mathematical Modelling

journal homepage: www.elsevier.com/locate/apm

Seismic ground motion in micropolar elastic half-space

Maheshreddy Gade, S.T.G. Raghukanth
Dept. of Civil Engineering, IIT Madras, India

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

Article history: In this article, analytic expressions for the frequency wave number spectra are derived for
Received 6 May 2013 the displacement and rotational components at the ground surface of a micropolar elastic
Received in revised form 14 February 2015 half-space. The ground motion is caused by buried sources described in terms of unit
Accepted 5 March 2015
impulsive force and micro-moments. Closed-form analytic expressions for both plane
Available online 18 March 2015
strain and anti-plane strain conditions are established. In the sequel, Greens functions
for reduced micro polar half-space where the couple stresses are zero and the stress tensor
Keywords:
is nonsymmetric are also presented. Numerical results are presented for anti-plane strain
Strong ground motion
Micro-rotations
case. The derived expressions can be used to simulate rotational and displacement
Reduced micro polar continuum components of ground motion during earthquakes.
2015 Elsevier Inc. All rights reserved.

1. Introduction

The dynamic response of an elastic half-space due to buried seismic sources has been a subject of great interest to seis-
mologists and engineers for computing the surface ground motion. Beginning with the pioneering work of Lamb [1] on elas-
tic waves generated by line sources, several articles on ground motion simulation in layered elastic media have appeared in
the technical literature [26]. In these studies, analytical expressions for three translational components have been derived
based on classical continuum mechanics theory. These analytical models are widely used for investigating the interior struc-
ture of the earth as well as earthquake source process through modeling of the recorded strong motion data [7,8].
On the other hand, studies on rotational components of ground motion and its effects on engineering structures are still
not clear. The difculty in measuring rotational components is mainly due to lack of technology in the strong motion devices
to observe such small rotations in wide frequency band and distance range. However, there have been many reports about
rotation of obelisks, grave stones and buildings during past earthquakes [9,10]. Many structural failures and much of the
damage caused by earthquakes have been attributed to rotational ground motions [11]. There have been evidences during
the past earthquakes and theoretical studies also show that contribution from rotational excitations reaches almost 50% of
the total response. Recently, with modern acquisition technology such as ber-optical or ring laser gyros, it has become
possible to measure rotational motions during earthquakes. Takeo [12] recorded near-eld rotational ground motion for
several events during earthquake swarm in Izu peninsula, Japan. Igel et al. [13] and Suryanto et al. [14] obtained rotational
time histories at teleseismic distances for Mw 8.1 Tokachi-Oki (2003) and Mw 6.3 Al Hoceima, Morocco earthquake (2004).
There have been efforts in the past to develop mechanistic models for computing rotational time histories. Bouchon and
Aki [15] simulated rotational ground motions in the near-eld due to buried strike-slip and dip-slip faults in the layered
medium. The rotational components induced by seismic waves are obtained from the space derivatives of the displacements.

Corresponding author.
E-mail address: raghukanth@iitm.ac.in (S.T.G. Raghukanth).

http://dx.doi.org/10.1016/j.apm.2015.03.006
0307-904X/ 2015 Elsevier Inc. All rights reserved.
 M. Gade, S.T.G. Raghukanth / Applied Mathematical Modelling 39 (2015) 72447265 7245

The simulated rotations were small compared to the amplitude of the translational components. Takeo [12] simulated
rotational ground motions based on the classical continuum mechanics and compared them with the recorded rotational
velocities of Mw 5.7 magnitude earthquake in Izu peninsula, Japan. These results showed that the recorded rotations are
several times larger than the simulated rotations computed from linear displacements. These differences have been attrib-
uted to the internal structure and discontinuities in the earths crust. In order to get reliable estimates of rotations, one will
have to model the microstructure of the medium itself as correctly as possible. In this regard, the linear theory of micropolar
elasticity developed by Eringen [16] will be a suitable tool in simulating all the six components of ground motion. In the
micropolar continuum theory, deformation is described not only by the displacement vector u but also by the rotation vector
/, which characterizes the micro-rotation of the particle. Kulesh et al. [1719] and Grekova et al. [20] studied wave
propagation in isotropic full and reduced Cosserat half-space. In this article, closed-form analytical expressions for the
Greens functions in a two-dimensional micropolar medium are established. A spatial and temporal Fourier transform is used
for solving the governing equations of motion. Analytical expressions are derived for the displacement and rotational
components at the ground surface.

2. Basic equations of linear theory of micropolar elasticity

In continuum theories, a physical body B is considered to be a collection of a set of material particles. The motion of each
material particle is described by a position vector identifying the location of each particle as a function of time. The number
of degrees of freedom associated with each material point in classical continuum theory is three. In microcontinuum
theories, it is assumed that there is a micro-structure around each material point which can deform and rotate indepen-
dently from the surrounding medium [21]. The intrinsic deformation of a material particle is represented by a geometrical
point and some deformable vectors, known as directors attached to the point that denotes the orientations and intrinsic
deformations of the material particle. By increasing the number of directors attached to the material point, higher order
microcontinuum theories can be derived. If the number of deformable directors is three, the material body is known as a
micromorphic continuum. The number of degree of freedom associated with each material point in a micromorphic contin-
uum is twelve. Assuming rigid directors, the degree of freedom at each material point reduces to six and the body is known
as micropolar continuum. In the micropolar continuum theory, deformation is described by the displacement vector u and
the rotation vector /, which characterizes the micro-rotation of the particle [22]. The equations of motion for micropolar
elastic solid are
rji;j qX i qui ;
1
i;
mji;j eijk rjk JY i J/
where rij , mij are the stress and couple-stress tensors respectively. mij is the body force per unit mass and Y i is the body
couple per unit mass and q is the density of the medium. J is the parameter describing the rotational inertia of the medium.
eijk denotes the unit anti-symmetric tensor. The kinematic relations for innitesimal strain (ij ) and curvature twist (uij )
tensor are expressed as
ij uj;i ekji /k ;
2
uij /j;i ;
The linear constitute laws for an isotropic micropolar elastic solid are
rij k2kk dij l j2ij l2ji ;
3
mij aukk dij buij cuji ;
where dij , is the Kronecker delta, k and l are the Lames constants respectively. a, b, c and j are the four new constants
describing the microstructure medium. For internal energy to be a positive quantity, the six elastic constants has to satisfy
the condition
3k 2l j P 0; l P 0; j P 0;
4
3a b c P 0; c P b P 0; c P 0:
By eliminating rij , mij , uij and ij using Eqs. (2), (3) and (1), we obtain a system of six coupled equations of linear
micropolar elasticity expressed in vector form as

Jw20
c21 rr  u  c22 r  r  u r/u 0;
2 5
w2 0;
c23 rr  /  c24 r  r  / 0 r  u  w20 /  /
2
where c1, c2, c3, c4 are the wave velocities
k 2l j lj abc c 2j
c21 ; c22 ; c23 ; c24 ; w2o : 6
q q qJ qJ qJ
7246 M. Gade, S.T.G. Raghukanth / Applied Mathematical Modelling 39 (2015) 72447265

It can be observed that, when j = 0 the governing equations for displacement and microrotation become uncoupled. In
this case, Eq. (5a) reduces to the displacement equations of motion of the classical elasticity. Eq. (5b) represents a hypotheti-
cal body in which only rotations occur.

3. Plane waves in an innite micropolar elastic solid

In classical elasticity, two types of waves namely longitudinal and shear waves exist in an innite medium. In micropolar
theory, the microstructure of the medium reveals new types of waves not encountered in the classical continuum mechanics.
According to Helmholtz decomposition, the eld equations (Eq. (5)) in a micropolar elastic medium can be simplied by
separating the displacement and microrotation components into potential and solenoidal parts in the following form
u rq r  W; r  W 0; 7

/ rn r  U; r:U 0: 8
By substituting the above equation in Eq. (5) and in absence of body forces and couples we obtain the following system of
equations

c21 r2 q  q
0; 9

c23 r2 n  w20 n  n 0; 10

Jw20
c22 r2 W 0;
rUW 11
2

w20
c24 r2 U  w20 U 0:
rWU 12
2
It can be observed that the equations for vector potentials W and U are coupled. Following Partt and Eringen [23], the
plane waves advancing in the positive direction can be expressed as
fq; n; W; Ug fa; b;A; Bg expikn  r  v t; 13
where a, b are complex constants. A, B are complex constant vectors, k is the wave number, v is the phase velocity and n is
the unit normal vector. Substituting q from Eq. (13) in Eq. (9) and using Eq. (7), the displacement vector can be obtained as
u ik1 an expik1 n  r  v 1 t; 14
where
k 2l j
v 21 : 15
q
It can be noted from Eq. (14) that the displacement of the particles is in the same direction as that of waves traveling at
velocity v 1 . Such a motion represents a train of longitudinal waves. These waves are similar to the dilatational waves
encountered in the classical continuum mechanics. Similarly substituting the expression for n from Eq. (13) in Eqs. (10)
and (8) yields
/ rn ik2 bn expik2 n  r  v 2 t; 16
where the velocity of these waves can be expressed as
abc
v 22  : 17
qJ 1  q2Jxj2
It can be observed from Eq. (16) that the microrotation vector points in the direction of propagation of waves. These new
dispersive waves are known as longitudinal microrotation waves. The speed (v 2 ) of these waves depend on frequency and
are not encountered in the classical theory. Similarly Eqs. (13), (11) and (12) yield

2 Jw20
x2  c22 k A i n  B 0; 18
2

Jw20
n  A x2  w20  c24 B 0: 19
2
Solving Eq. (19) for B gives

Jw20
B n  A: 20
2x  w20  c24
2
 M. Gade, S.T.G. Raghukanth / Applied Mathematical Modelling 39 (2015) 72447265 7247

The above equation indicates that the three vectors n, A and B are mutually perpendicular to each other. The waves asso-
ciated with W are known as transverse displacement waves and those waves associated with U are known as transverse
microrotation waves. By eliminating A and B from the above equations and solving for velocities of the coupled waves,
we get
s
 2
w20
c2 c24 1 w20 Jw4
v 2
3;4 2 2  2
c24  c22 20 : 21
2k 2 4 k 4k

The upper sign (+) corresponds to the velocity v3 of the coupled transverse displacement wave and the lower sign ()
corresponds to the velocity v4 of the coupled transverse microrotational waves. The coupled transverse displacement waves
represent the counterpart of the classical shear waves and reduces to them when j and c are zero. The coupled transverse
microrotational waves are new and are not encountered in the classical theory. The above expressions can be nd in Eringen
[24].

4. Ground motion at free surface of half-space due to a buried sources

Consider a homogeneous micropolar elastic half-space with free surface y = 0 and assume the wave normal n to lie in the
xy-plane. The particle motion is invariant with respect to z if the wave normal is in xy-plane. Then, the displacements u and
microrotations / are functions of x, y and t. The derivative with respective to z is zero. Subsequently, the six coupled equa-
tions (Eq. (5)) can be written as.
" # " #
@ 2 ux @ 2 uy 2 @ uy
2
@ 2 ux Jw2 @/z @ 2 ux
C 21 2
 C2  2
0  2 0;
@x @x@y @x@y @y 2 @y @t
" # " #
2 2 2 2
@ ux @ uy 2 @ uy @ ux Jw20 @/z @ 2 uy
C 21 C 2    2 0;
@x@y @y2 @x2 @x@y 2 @x @t
" #  
2 2 2
@ /z @ /z w2 @uy @ux @ /
C 24 0   w20 /z  2z 0;
@x2 @y2 2 @x @y @t
" 2
# " 2 # 22
2 2
@ /x @ / y @ /y @ /x 2
w @uz @2/
C 23  C 24  0  w20 /x  2x 0;
@x2 @x@y @x@y @y2 2 @y @t
" # " #
2 2
2
@ /x @ /y @ /y @ /x 2 2
w @uz @ 2 /y
C 23 C 24   0  w20 /y  0;
@x@y @y 2 @x 2 @x@y 2 @x @t 2
" #  
@ 2 uz @ 2 uz Jw20 @/y @/x @ 2 uz
C 22 2
2
  2 0:
@x @y 2 @x @y @t

It can be observed that, two displacement components ux, uy and one microrotation component /z are coupled through
Eqs. (22)(ac). Eqs. 22(df) represents coupled equations for two microrotations /x, /y and one displacement component uz.
Further, from Eq. (3) it can be shown that the stresses rxx, ryy, rxy, ryx, mxz, myz, mzx and mzy depend only on ux, uy and /z and
the stresses rxz, ryz, rzy, rzx, mxx, myy, myx and mxy are functions of /x, /y and uz. This makes it possible to resolve the 3D prob-
lem into two parts, where one is plane strain case corresponding to ux, uy and /z and the other is anti-plane strain case
corresponding to /x, /y and uz.

4.1. Plane strain

The Cartesian coordinates x and z are directed along the surface of the half-space and the y-axis is perpendicular to it. The
solution to the buried source can be treated effectively as a half-space problem involving stress discontinuities across the
plane y = y0. For this purpose it is convenient to view the half-space as being composed of an upper and lower region of
the same material dened as.
I Region I; jxj < 1; 0 < y 6 y0 ;

II Region II; jxj < 1; y0 < y 6 1:

For a plane strain problem, the displacement and rotation vector is represented as u = (ux, uy, 0) and / = (0, 0, /z ). Two
potential functions, q and W, can be dened to describe the displacements in x- and y-directions.
@q @ W @q @ W
ux ; uy  ; /z /: 23
@x @y @y @x
7248 M. Gade, S.T.G. Raghukanth / Applied Mathematical Modelling 39 (2015) 72447265

The equations of motion for plane deformation parallel to xy plane in terms of the potentials are
!
@2q @2q @2q
c21 2
2
 2 0;
@x @y @t
!
2 2
@ W @ W Jw2 @2W
c22 2
2 0 /  2 0; 24
@x @y 2 @t
! !
2 2
@ / @ / w20 @ 2 W @ 2 W @2/
c24  w 2
0 /   0:
@x2 @y2 2 @x2 @y2 @t 2

The constitutive equations (Eq. (3)) in terms of potentials can be written as

!
2 @2q @2W
rxx kr q 2l j 2 ;
@x @x@y
!
@2q @2W
ryy kr2 q 2l j 2  ;
@y @x@y
@2q @2W @2W 25
rxy 2l j l 2  l j 2  j/;
@x@y @y @x
@2q @2W @2W
ryx 2l j  l 2 l j 2 j/;
@x@y @x @y
@/ @/ @/ @/
mxz b ; myz b ; mzx c ; mzy c :
@x @y @x @y
The buried sources inside the medium can be effectively treated as stress discontinuities across the plane y = y0. The three
loading cases considered in this study are shown in Fig. 1(a)(c). The magnitude of these three loads acting on the line y = y0
are exp[i(kx + xt)] per unit length. These can be expressed as
ryy x; y y0 ; t  ryy x; y y0 ; t expikx xt; 26

ryx x; y y0 ; t  ryx x; y y0 ; t expikx xt; 27

myz x; y y0 ; t  myz x; y y0 ; t expikx xt: 28

The displacements and micro-rotations are required to be continuous across the plane y = y0. These are given as
ux x; y y0 ; t ux x; y y0 ; t; 29

uy x; y y0 ; t uy x; y y0 ; t; 30

/x; y y0 ; t /x; y y0 ; t; 31

Fig. 1. (a)(c) Three loading conditions for plane strain problem.

 M. Gade, S.T.G. Raghukanth / Applied Mathematical Modelling 39 (2015) 72447265 7249

The boundary conditions in a half-space are given by the vanishing of stresses at the free surface which is
ryy x; 0; t ryx x; 0; t myz x; 0; t 0: 32
The solutions of Eq. (24) can be written as

~y expikx xt;
qx; y; t q
Wx; y; t W~ y expikx xt; 33
~
/x; y; t /y expikx xt:

Here k is the real wave number along the x direction and x stands for frequency. Substituting the above equation in Eq.
(24), the governing equations can be rewritten as
2  
d q~ x2 2
~ 0;
2
k q
dy c21
2
~  2 
d W x 2 ~ J w20 ~
2
k W / 0; 34
dy c22 2c22
2~   2~ 2
d / x2 w20 2 ~ w2 d W k w20 ~
2
 k /  02 W 0:
dy c24 2
c4 2c4 dy 2
2c24

In region I, the solution to the equation can be expressed as

~y AI eiv 0 yy0 BI eiv 0 yy0 ;
q
~ y DI eiv 1 yy0 DI eiv 2 yy0 DI eiv 1 yy0 DI eiv 2 yy0 ;
W 35
1 2 3 4
   
~
/y #1 DI1 eiv 1 yy0 DI3 eiv 1 yy0 #2 DI2 eiv 2 yy0 DI4 eiv 2 yy0 :

In region II and below the seismic source

~y AII eiv 0 yy0 ;

q
~ y DII eiv 1 yy0 DII eiv 2 yy0 ;
W 36
1
 2  
~
/y #1 DII1 eiv 1 yy0 #2 DII2 eiv 2 yy0 ;

where
 
2c22 2 x2
#1 k v 2
1  ;
Jw20 c22
 
2c22 2 x2
#2 k v 2
2  :
Jw20 c22
m0, m1 and m2 are the vertical wave numbers dened through,
s q q
x2 2 2 2
v0 k ; v1 a1  k ; v2 a2  k ;
c21
s 37
2 2
c2 c2 x2 w20 J  4c22 w20 c22  c24 4 w20 w20 Jc22 c24  4c22 c22  c24 2 w40 w20 J  4c22
a1;2 2 2 42  x x :
2c2 c4 8c22 c24 4 4
4c2 c4 4 4
8c2 c4 64c2 c4 4 4

The total numbers of unknown coefcients are 9. These constants have to be found from the above nine boundary con-
ditions. The stress free boundary condition on the surface of the half-space (Eq. (32)) results in
1 h


i
DI2 2kv 0 Jw20 2l jC3 eiy0 v 2 v 0 BI  2C1 C3 eiy0 v 1 v 2 DI3  C1 C4  C5 C2 C3 e2iv 2 y0 DI4 ;
C5 C2 C3  C1 C4
38

C6  iy0 v 1 v 2 I 
DI1 e2iv 1 y0 DI3 e D4  eiy0 v 2 v 1 DI2 ; 39
C7

C8  iy0 v 2 v 0 I 
AI e2iv 0 y0 BI e D2  eiy0 v 0 v 2 DI4 : 40
C9
7250 M. Gade, S.T.G. Raghukanth / Applied Mathematical Modelling 39 (2015) 72447265

From the continuity of displacements and rotations along the plane y = y0 (Eqs. (29)(31)), the constants AII ; DII1 and DII2 can
be expressed as
 
II I
C10 BI  2kc22 v 1 v 2 v 1 DI3 v 2 DI4
A A ; 41
C11
  
2 I
2v 0 x2  c22 k v 22 kB  v 2 DI4 v 1 v 2 C12 DI3
DII1 DI1 ; 42
v 1  v 2 C11
  
2 I
2v 0 x2  c22 k v 21 kB  v 1 DI3  v 1 v 2 C13 DI4
DII2 DI2  ; 43
v 1  v 2 C11
where
   
C1 Jw20 lk2  v 21 l j  2j x2  c22 k2 v 21 ;
   
C2 Jw20 lk2  v 22 l j  2j x2  c22 k2 v 22 ;
  
C3 v 1 c22 k2 v 21  x2 2
kk v 20 v 20 2l j ;

  
C4 v 2 c22 k2 v 22  x2 2
kk v 20 v 20 2l j ;

C5 k2 c22 Jw20 v 0 v 1 v 2 v 21  v 22 2l j2 ;
 
C6 v 2 c22 k2 v 22  x2 ;
 
C7 v 1 c22 k2 v 21  x2 ;

C8 kv 2 c22 v 21  v 22 2l j;
  
C9 c22 k2 v 21  x2 2
kk v 20 v 20 2l j ;

 
C10 j2 c22 v 1 v 2  v 0 x2  c22 k2  v 1 v 2 ;
 
C11 j2 c22 v 1 v 2 v 0 x2  c22 k2  v 1 v 2 ;
 
C12 j2 c22 v 1  v 2  v 0 x2  c22 k2 v 1 v 2 ;

C13 j2 c22 v 1  v 2 v 0 x2  c22 k2 v 1 v 2 :

   
Six coefcients AI ; DI1 ; DI2 ; AII ; DII1 and DII2 are expressed in terms of other three coefcients BI ; DI3 ; and DI4 in above
equations (Eqs. (38)(43)). These three coefcients have to be estimated by using stress discontinuity boundary condition
along the plane y = y0 (Eqs. (26)(28)).

4.1.1. Results for specic source distributions

Loading case I:
r~ yy x; y y0 ; t  r~ yy x; y y0 ; t 0;

r~ yx x; y y0 ; t  r~ yx x; y y0 ; t expikx xt;

~ yz x; y y0 ; t  m
m ~ yz x; y y0 ; t 0;

2
kc2
BI ; 44
2v 0 x2 l j
 M. Gade, S.T.G. Raghukanth / Applied Mathematical Modelling 39 (2015) 72447265 7251

2
c22 k v 22  x2
DI3  ; 45
2x2 l jv 21  v 22

2
c22 k v 21  x2
DI4 ; 46
2x2 l jv 21  v 22
Loading case II:
r~ yy x; y y0 ; t  r~ yy x; y y0 ; t expikx xt;

r~ yx x; y y0 ; t  r~ yx x; y y0 ; t 0;
~ yz x; y y0 ; t  m
m ~ yz x; y y0 ; t 0

1
BI 2
; 47
2k 2l jk v 20

2
kc22 k v 22  x2
DI3  ; 48
2c22 v 1 k 2l jv 21  v 22
2
kc22 k v 21  x2
DI4 : 49
2c22 v 2 k 2l jv 21  v 22
Loading case III:

r~ yy x; y y0 ; t  r~ yy x; y y0 ; t 0;

r~ yx x; y y0 ; t  r~ yx x; y y0 ; t 0;

~ yz x; y y0 ; t  m
m ~ yz x; y y0 ; t exp ikx xt;

BI 0; 50

iJw20
DI3
; 51
4c22 v 1 b v 21  v 22
iJw20
DI4 
: 52
4c22 v 2 b v 21  v 22
The obtained solutions are valid for the loading cases shown in Fig. 1 where the loads act periodically all over the plane
y = y0 and harmonically in time. The corresponding solutions for the three unidirectional unit impulses can be obtained by
using the Fourier integral representation of the Dirac delta function.

4.2. Anti-plane strain

For a anti-plane strain problem, the displacement and rotation vector is represented as u = (0, 0, uz) and / = (/x , /y , 0). Two
potential functions, n and U can be dened to describe the rotations about x- and y-directions.
@n @ U @n @ U
/x ; /y  ; uz uz : 53
@x @y @y @x
The equations of motion for plane deformation parallel to xy plane in terms of the potentials are
!
@2n @2n @2q
c23
2
2  w20 n  2 0;
@x @y @t
! !
2 2 2 2
@ uz @ uz 2
Jw0 @ U @ U @ 2 uz
c22   0; 54
@x2 @y2 2 @x2 @y2 @t2
!
@2U @2U w20 @2U
c24  w 2
0 U uz  0:
@x2 @y2 2 @t 2
7252 M. Gade, S.T.G. Raghukanth / Applied Mathematical Modelling 39 (2015) 72447265

The constitutive equations (Eq. (3)) in terms of potentials can be written as

 
@uz @n @ U
rxz l j j  ;
@x @y @x
 
@uz @n @ U
ryz l j j ;
@y @x @y
 
@uz @n @ U
rzx l j  ;
@x @y @x
 
@uz @n @ U
rzy l j ;
@y @x @y
!
2 @2n @2U 55
mxx ar n b c ;
@x2 @x@y
!
@2n @2U
myy ar2 n b c  ;
@y2 @x@y
@2n @2U @2U
myx b c b 2 c 2 ;
@x@y @y @x
@2n @2U @2U
mxy b c b 2 c 2 ;
@x@y @x @y
The buried sources inside the medium can be effectively treated as stress discontinuities across the plane y = y0. The three
loading cases considered in this study are shown in Fig. 2(a)(c). The magnitude of these three loads acting on the line y = y0
are exp[i(kx + xt)] per unit area. These can be expressed as
myy x; y y0 ; t  myy x; y y0 ; t expikx xt; 56

myx x; y y0 ; t  myx x; y y0 ; t expikx xt; 57

ryz x; y y0 ; t  ryz x; y y0 ; t expikx xt: 58

The displacements and micro-rotations are required to be continuous across the plane y = y0. These are given as
/x x; y y0 ; t /x x; y y0 ; t; 59

/y x; y y0 ; t /y x; y y0 ; t; 60

uz x; y y0 ; t uz x; y y0 ; t; 61
boundary conditions in a half-space are given by the vanishing of stresses at the free surface which is
myy x; 0; t myx x; 0; t ryz x; 0; t 0: 62

Fig. 2. (a)(c) Three loading conditions for anti-plane strain problem.

 M. Gade, S.T.G. Raghukanth / Applied Mathematical Modelling 39 (2015) 72447265 7253

The solutions of Eq. (54) can be written as

nx; y; t ~ny expikx xt;

Ux; y; t U ~ y expikx xt; 63
~z y expikx xt:
uz x; y; t u
Here k is the real wave number along the x direction and x stands for frequency. Substituting the above equation in
Eq. (54), the governing equations can be rewritten as
2  
d ~n x2
w20 2 ~
2
 k n 0;
dy c23 c23
2  2  2~ 2
d u~z x 2 w20 d U k w20 ~
2
 k u~ z  J J U 0; 64
dy c22 2c22 dy2 2c22
2~  2 
d U x w20 2 ~ w20 ~
2
2
 2  k U 2 uz 0:
dy c4 c4 2c4

In region I, the solution to the equation can be expressed as

~ny AI eiv 0 yy0 BI eiv 0 yy0 ;
~ y DI eiv 1 yy0 DI eiv 2 yy0 DI eiv 1 yy0 DI eiv 2 yy0 ;
U 65
1 2 3 4
   
~z y 11 DI1 eiv 1 yy0 DI3 eiv 1 yy0 12 DI2 eiv 2 yy0 DI4 eiv 2 yy0 :
u

In region II, the solution to the equation can be expressed as

~ny AII eiv 0 yy0 ;
~ y DII eiv 1 yy0 DII eiv 2 yy0 ;
U 66
1 2

~z y 11 DII1 eiv 1 yy0 12 DII2 eiv 2 yy0 ;

u
where
 
2c24 2 w2 x2
11 2
k v 21 20  2 ;
w0 c4 c4

 
2c24 2 w20 x2
12 k v 2
2  :
w20 c24 c24

m0, m1 and m2 are the vertical wave numbers dened through,

s q q
x2 w2
v 0 2  20  k2 ; v 1 a1  k2 ; v 2 a2  k2 ;
c3 c3
2
s 67
2 2
c22 c24 x2 w0 J  4c22 w20 c22  c24 4 w20 w20 Jc22 c24  4c22 c22  c24 2 w40 w20 J  4c22
a1;2  x x :
2c22 c24 8c22 c24 4c42 c44 8c42 c44 64c42 c44
The total numbers of unknown coefcients are 9. These constants have to be found from the above nine boundary
conditions. The stresses are zero along the free surface (y = 0). From the Eq. (62) the constants AI ; DI1 ; and DI2 can be
expressed as
1 h   i
DI2 K6 K1 2kv 0 b ceiy0 v 0 v 2 BI  2bv 21  ck2 eiy0 v 1 v 2 DI3  K1 K5 K2 K3 e2iv 2 y0 DI4 ; 68
K1 K4  K2 K3

K2 h iy0 v 1 v 2 I i
DI1 e2iv 1 y0 DI3 e D4  eiy0 v 2 v 1 DI2 ; 69
K1
h    i
K7 eiy0 v 1 v 0 DI1  eiy0 v 0 v 1 DI3 K8 eiy0 v 2 v 0 DI2  eiy0 v 0 v 2 DI4
AI e2iv 1 y0 BI : 70
K5
The continuity of displacements and rotations along the plane y = y0 (Eqs. (59)(61)) results in
1 h  i
AII AI K9 BI  2kc24 v 1 v 2 v 1 DI3 v 2 DI4 ; 71
K10
7254 M. Gade, S.T.G. Raghukanth / Applied Mathematical Modelling 39 (2015) 72447265

1 h   i
2 I
DII1 DI1 v 1 v 2 K11 DI3 2v 0 x2  w20  c24 k v 22 kB  v 2 DI4 ; 72
v 1  v 2 K10

1 h   i
2 I
DII2 DI2 v 1 v 2 K12 DI4  2v 0 x2  w20  c24 k v 21 kB  v 1 DI3 ; 73
v 1  v 2 K10
where
h    i
2
K1 v 1 2l j a b cv 20 ak x2  w20  c24 k2 v 21 ja b cw20 k2 v 20 ;
h    i
2
K2 v 2 2l j a b cv 20 ak x2  w20  c24 k2 v 22 ja b cw20 k2 v 20 ;
  
2 2 2
K3 a b cv 20 ak bv 21  ck v 0 v 1 k b c2 ;
  
2 2 2
K4 a b cv 20 ak bv 22  ck v 0 v 2 k b c2 ;

  
2 2 2
K5 a b cv 20 ak bv 22  ck  v 0 v 2 k b c2 ;

 
2
K6 a b cv 20 ak ;

K7 kv 1 b c;

K8 kv 2 b c;

2 2
K9 k c24 v 1 v 2  v 0 x2  w20  c24 k  v 1 v 2 ;
 
2 2
K10 k c24 v 1 v 2 v 0 x2  w20  c24 k  v 1 v 2 ;

2 2
K11 k c24 v 1  v 2  v 0 x2  w20  c24 k v 1 v 2 ;
 
K12 k2 c24 v 1  v 2 v 0 x2  w20  c24 k2 v 1 v 2 :
   
Six coefcients AI ; DI1 ; DI2 ; AII ; DII1 and DII2 are expressed in terms of other three coefcients BI ; DI3 ; and DI4 in above
equations (Eqs. (68)(73)). These three coefcients have to be estimated by using the stress discontinuity condition along
the plane y = y0 (Eqs. (56)(58)).

4.2.1. Results for specic source distributions

Loading case I:

myy x; y y0 ; t  myy x; y y0 ; t 0;

myx x; y y0 ; t  myx x; y y0 ; t 0;

ryz x; y y0 ; t  ryz x; y y0 ; t exp ikx xt;

BI 0; 74

2
iw0
DI3 ; 75
4c24 v 1 l jv 21  v 22
2
iw0
DI4  ; 76
4c24 v 2 l jv 21  v 22
Loading case II:
myy x; y y0 ; t  myy x; y y0 ; t 0;

myx x; y y0 ; t  myx x; y y0 ; t exp ikx xt;

 M. Gade, S.T.G. Raghukanth / Applied Mathematical Modelling 39 (2015) 72447265 7255

ryz x; y y0 ; t  ryz x; y y0 ; t 0;

2
kc4
BI ; 77
2bv 0 w20  x2

2
x2  w20  c24 k v 22
DI3  ; 78
2bv 21  v 22 w20  x2

2
x2  w20  c24 k v 21
DI4 ; 79
2bv 21  v 22 w20  x2
Loading case III:
myy x; y y0 ; t  myy x; y y0 ; t expikx xt;

myx x; y y0 ; t  myx x; y y0 ; t 0;

ryz x; y y0 ; t  ryz x; y y0 ; t 0;

1
BI 2
; 80
2a b ck v 20
 
2
k x2  w20  c24 k v 22
DI3 2
; 81
2v 1 c24 v 21  v 22 a b ck v 20

2
kx2  w20  c24 k v 21
DI4  2
: 82
2v 2 c24 v 21  v 22 a b ck v 20
The obtained solutions are valid for the loading cases shown in Fig. 2 where the loads act periodically all over the plane
y = y0 and harmonically in time. The corresponding solutions for the three unidirectional unit impulses can be obtained by
using the Fourier integral representation of the Dirac delta function.

5. Reduced micropolar elastic half-space

The above derived expressions can be used to simulate Greens functions in a micropolar medium. The main limitation
with the micropolar theory is that six material constants are required to simulate ground motion. These constants are not
directly available for rock and soil medium. To circumvent these difculties, Schwartz et al. [25] proposed reduced micropo-
lar theory to model granular materials. In this theory, the material point in a continuum is dened by three displacements u
and three rotations / which are kinematically independent. In reduced micropolar theory the particles are free to rotate
(micro rotation) but the continuum does not offer resistance to micro-rotation, unlike micropolar theory. The stress tensor
is asymmetric and couple stress tensor is zero. By this assumption three micropolar constants a; b; c are eliminated.
Previously Grekova et al. [20] studied surface wave propagation in reduced Cosserat medium. The equations of the motions
in terms of u and / can be expressed as
Jw20
c21 rr  u  c22 r  r  u 0;
r/u
2 83
w20 0;
r  u  w20 /  /
2
where
k 2l j lj 2j
c21 ; c22 ; w2o : 84
q q qJ

5.1. Ground motion at free surface of half-space due to a buried source

5.1.1. Plane strain

For a plane strain problem, the displacement and rotation vector is represented as u = (ux, uy, 0) and / = (0, 0, /z ). Two
potential functions, q and W, can be dened to describe the displacements in x- and y-directions.
@q @ W @q @ W
ux ; uy  ; /z /: 85
@x @y @y @x
7256 M. Gade, S.T.G. Raghukanth / Applied Mathematical Modelling 39 (2015) 72447265

The equations of motion for plane deformation parallel to xy plane in terms of the potentials are
!
@2q @2q @2q
c21 2
2
 2 0;
@x @y @t
! 86
2 2
@ W @ W Jw2 @2W
c22 2
2 0 /  2 0:
@x @y 2 @t
The constitutive equations (Eq. (3)) in terms of potentials can be written as
!
@2q @2W
rxx kr2 q 2l j ;
@x2 @x@y
!
2 @2q @2W
ryy kr q 2l j  ;
@y2 @x@y
87
@2q @2W @2W
rxy 2l j l 2  l j 2  j/;
@x@y @y @x
@2q @2W @2W
ryx 2l j  l 2 l j 2 j/:
@x@y @x @y
The buried sources inside the medium can be effectively treated as stress discontinuities across the plane y = y0. The mag-
nitude of these three loads acting on the line y = y0 are exp[i(kx + xt)] per unit area. These can be expressed as
ryy x; y y0 ; t  ryy x; y y0 ; t expikx xt; 88

ryx x; y y0 ; t  ryx x; y y0 ; t expikx xt: 89

The displacements and micro-rotations are required to be continuous across the plane y = y0. These are given as
ux x; y y0 ; t ux x; y y0 ; t; 90

uy x; y y0 ; t uy x; y y0 ; t: 91
The boundary conditions in a half-space are given by the vanishing of stresses at the free surface which is
ryy x; 0; t ryx x; 0; t 0 92
The solutions of Eq. (86) can be written as
~y expikx xt;
qx; y; t q
Wx; y; t W~ y expikx xt; 93
~
/x; y; t /y expikx xt:
Here k is the real wave number along the x direction and x stands for frequency. Substituting the above equation in Eq.
(86), the governing equations can be rewritten as
2  
d q~ x2 2
~ 0;
2
k q
dy c21
2  2  94
~
d W x 2
k
2 ~ J w0 /
W ~ 0:
2
dy c22 2c22
~ as
~ can be expressed in terms of potential W
From Eq. (83b) /
" 2~
#
~ w20 2~ d W
/ k W 2 : 95
2x2  w20 dy
~ and the solution to the Eq. (94) can be expressed as
Substitute equation (95) in Eq. (94b) for /
In region I
~y AI eiv 0 yy0 BI eiv 0 yy0 ;
q
96
~ y DI eiv 1 yy0 DI eiv 1 yy0 :
W 1 2

In region II
~y AII eiv 0 yy0 ;
q
97
~ y DII eiv 1 yy0 ;
W 1
 M. Gade, S.T.G. Raghukanth / Applied Mathematical Modelling 39 (2015) 72447265 7257

where m0, and m1 are the vertical wave numbers dened through,
v
 
s u
u4w2  x2 x2  k2 c2 Jk2 w4
x 2
2 t 0 2 0
v0 k ; v1 : 98
c21 4c22 w20  x2  Jw40

Grekova et al. [20] studied wave propagation in isotropic reduced Cosserat half-space. v 0 0 and v 1 0 correspond to
the dispersion relations for the compression and shear-rotational waves in the reduced Cosserat medium. The compression
wave is similar to classical continuum. The shear wave is coupled with rotation wave and the phase velocity of the shear
wave is function of frequency.
The total number of unknown coefcients is 6. These constants have to be found from the above six boundary conditions.
The displacements are continuous along the plane (y = y0). From the Eqs. (90) and Eq. (90).
DII1 C3 BI DI1 C1 DI2 ; 99

AII C1 BI AI  C2 DI2 : 100

The stresses are zero along the free surface (y = 0). From the Eq. (92).
AI PPBI SPDI2 ; 101

DI1 PSBI SSDI2 ; 102

where PP, PS, SP and SS are reection coefcients. It can be shown that as j ? 0, these reection coefcients are as same as
linear elastic reection coefcients reported in Aki and Richards [26].where
C5 C4  C6 e2iy0 v 0
PP ;
C5 C4 C6

2C4 C6 eiy0 v 1 v 0
SP ;
C5 C4 C6

2C5 eiy0 v 1 v 0
PS  ;
C5 C4 C6

C5 C4  C6 e2iy0 v 1
SS ;
C5 C4 C6
2
k  v 0v 1
C1 2
;
k v 0v 1

2kv 1
C2 2
;
k v 0v 1

2kv 0
C3 2
;
k v 0v 1

2l jkv 1
C4 2
;
kk v 20 v 20 2l j

C5 2kv 0 2l jx2  w20 ;

C6 2lk2  v 21 w20  x2 j2v 21 x2 w20 k2  v 21 :

Results for specic source distributions
Loading case I:
ryy x; y y0 ; t  ryy x; y y0 ; t 0;

ryx x; y y0 ; t  ryx x; y y0 ; t expikx xt;

x2  w20
DI2 2  ; 103
k v 2
1 2lx2  w20 j2x2  w20
7258 M. Gade, S.T.G. Raghukanth / Applied Mathematical Modelling 39 (2015) 72447265

kx2  w20
BI  2  : 104
v 0 k v 2
1 2lx2  w20 j2x2  w20
Loading case II:
ryy x; y y0 ; t  ryy x; y y0 ; t expikx xt;

ryx x; y y0 ; t  ryx x; y y0 ; t 0;

k
DI2 2
; 105
2v 1 k 2l jk v 20

1
BI 2
: 106
2k 2l jk v 20
The obtained solutions are valid for the loading cases shown in Fig. 1(a) and (b) where the loads act periodically all over
the plane y = y0 and harmonically in time. The corresponding solutions for the two unidirectional unit impulses can be
obtained by using the Fourier integral representation of the Dirac delta function.

5.1.2. Anti-plane strain

For a anti-plane strain problem, the displacement and rotation vector is represented as u = (0, 0, uz) and / = (/x , /y , 0). The
equations of motion for plane deformation parallel to xy plane
!  
@ 2 uz @ 2 uz Jw20 @/y @/x @ 2 uz
c22 2   2 0: 107
@x2 @y 2 @x @y @t

The constitutive equations (Eq. (3a)) in terms of displacements can be written as

@uz
ryz l j  j/x : 108
@y
The buried sources inside the medium can be effectively treated as stress discontinuities across the plane y = y0. The mag-
nitude of the load acting on the line y = y0 is exp[i(kx + xt)] per unit area. This can be expressed as
ryz x; y y0 ; t  ryz x; y y0 ; t expikx xt: 109
The displacement is required to be continuous across the plane y = y0. These are given as
uz x; y y0 ; t uz x; y y0 ; t: 110
The boundary conditions in a half-space are given by the vanishing of stresses at the free surface which is
ryz x; 0; t 0: 111
The solution of Eq. (107) can be written as
~ x y expikx xt;
/x x; y; t /
/ x; y; t /~ y y expikx xt; 112
y
~ z y expikx xt:
uz x; y; t u
Here k is the real wave number along the x direction and x stands for frequency. Substituting the above equation in Eq.
(107), the governing equation can be rewritten as
2
!
d u~z Jw2 ~
c22 ~ y  @ /x
0 ik/
2
x2  k c22 u
~ z 0: 113
2 2 @y
dy
~ x and /
From Eq. (83b) / ~ y can be expressed in terms of potential u
~ z as

~x w20 ~z
@u
/ ;
2w20  x @y
2

2
114
~y ikw0
/ ~z :
u
2w20  x2
Substitute equation (114) in Eq. (113), after rearranging (113)
2
d u~z
2
v 20 u
~ z 0; 115
dy
 M. Gade, S.T.G. Raghukanth / Applied Mathematical Modelling 39 (2015) 72447265 7259

where
s
2
4x2  k c22 w20  x2 Jw40
v0 :
4c22 w20  x2  Jw40
In region I, the solution to the eq. can be expressed as

~ z y AI eiv 0 yy0 BI eiv 0 yy0 :

u 116
In region II, the solution to the equation can be expressed as

~ z y AII eiv 0 yy0 :

u 117
The total number of unknown coefcients is 3. These constants have to be found from the above three boundary condi-
tions. Displacement is continuous along the plane (y = y0). From Eq. (110).

AII AI BI : 118
I
The stress is zero along the free surface (y = 0). From Eq. (111), the constant A can be expressed as

AI BI e2iv 0 y0 : 119
I
Using the stress discontinuity boundary condition along the plane y = y0 reported in Eq. (109), the constant B can be
obtained as
i
BI ; 120
2v 0 K1
where
jw20
K1 l j  ;
2w20  x2
As j ? 0, it can be shown that these constants (BI and AI) are equivalent to linear elastic constants [26]. The obtained
solutions are valid for the loading cases shown in Fig. 2c where the loads act periodically all over the plane y = y0 and
harmonically in time. The corresponding solutions for the unidirectional unit impulse can be obtained by using the
Fourier integral representation of the Dirac delta function.

6. Numerical results

Ground motions have been simulated numerically to illustrate the application of derived analytical expressions. Time
histories are estimated for anti-plane strain problem. Seven material constants (k, l, c, a, b, j and j) are required to dene
linear elastic micropolar material. Reduced micropolar theory requires two additional material constants j and j in addition
to Lames constants (k, l). In the present study, the material properties (k, l, c, and j) are taken from the work of Eringen
[24]. The following values of material parameters are used for numerical simulation, k = 7.59 GPa, l = 1.89 GPa, c = 2.63 kN,
j = 0.015 GPa, q = 2180 kg/mm3 and j = 0.196 mm2. The other two material constants a and b are approximately taken as c.
The response of micropolar elastic half space is estimated for three different loading cases shown in Fig. 2a. The response for
unit impulsive sources as shown in Fig. 3 are obtained by taking Fourier transform.

Fig. 3. (a)(c) Three unit impulsive force/micro-moments applied at (0, Y0) corresponds to anti-plane strain problem.
7260 M. Gade, S.T.G. Raghukanth / Applied Mathematical Modelling 39 (2015) 72447265

Fig. 4. Frequency-wavenumber spectras of uz ; /x and /y .

 M. Gade, S.T.G. Raghukanth / Applied Mathematical Modelling 39 (2015) 72447265 7261

Z 1 Z 1
1
Gpz x; y; t uz k; y; x expikx xtdkdx;
2p 1 1
Z 1 Z 1
1
Gp/x x; y; t / k; y; x expikx xtdkdx; 121
2p 1 1 x
Z 1 Z 1
1
Gp/y x; y; t / k; y; x expikx xtdkdx;
2p 1 1 y

where Gpz ; Gp/x and Gp/y are the Greens functions of displacement uz , rotation component /x and rotation component /y due to
unit impulsive source applied in the p-direction. uz ; /x and /y are the response of the medium for loadings as shown in the
Fig. 2. The innite integrals in Eq. (121) can be numerically evaluated by discrete wave number method developed by
Bouchon [27]. In this method, the innite integrals are replaced by summations by introducing spatial periodicity in the
loading. Selecting the appropriate cut-off frequency and wavenumber values is very important in discrete wave number
method. These cut-off values should be selected such that no important information on displacement/rotation is left outside
the integration limits. The cut-off values are selected from frequency (x)-wavenumber (k) spectras of uz ; /x and /y . The
corresponding x  k spectras are plotted in Fig. 4. The cut-off frequency x 2 rad=s and the cut-off wavenumber
k 0:004 rad=m. The nal solution is obtained by Fourier transform in the complex frequency domain. A review of the
discrete wave number method is available in Bouchon [27]. The ground motions are calculated at four locations on the free
surface for a unit impulsive force/micro moment applied at a depth of 10 km (Y0 = 10 km). The three constants BI ; DI3 and DI4
for unit impulsive force (Fyz) are estimated by using Eqs. (74)(76), for all frequency (x) and wavenumber (k) ranges. Other
three constants AI ; DI1 and DI2 are calculated from Eqs. (68)(70). The two potentials n; U and displacement uz are estimated
using Eqs. (65) and (63). The rotational components are obtained from potentials by using Eq. (53). These are substituted in
Eq. (121) to obtain the displacement and rotation time histories for a unit impulsive force. The simulated time histories are
shown in Fig. 5. In a similar fashion Greens functions are calculated for both unit impulsive micro-moment applied in
x-direction (Myx) and y-direction (Myy) from Eq. (121) using the constants reported in Eqs. (77)(79) and Eqs. (80)(82).
The simulated time histories at four locations on the surface of the half-space are reported in Figs. 6 and 7. From Figs. 5
7, it can be observed that the arrival time of the rst wave increases with increase in the distance from the source. The peak
values of the simulated displacements are higher than the amplitude of the micro-rotations for all the three loading cases.
The unit impulsive micro-moment applied about x-axis (Myx) has produced higher displacements and rotations compared to
other two impulsive loads.
Fig. 8 shows the simulated displacement and rotation time histories for reduced micropolar elastic half-space. The two
constants AI and BI are estimated by using Eqs. (119) and (120) for unit impulsive force for all frequency (x) and wavenum-
ber (k) ranges. The displacements and rotations are estimated from Eqs. (114) and (116). These are substituted in Eq. (112)
and the Greens functions are simulated using Eq. (121). The obtained peak displacements and micro-rotations in a reduced
micropolar half-space are lower than compared to micropolar half-space (Fig. 5). This can be attributed to the presence of
additional microrotational waves in micropolar elastic half-space.

Fig. 5. Response of micropolar elastic half space for unit impulse force (Fyz).
7262 M. Gade, S.T.G. Raghukanth / Applied Mathematical Modelling 39 (2015) 72447265

Fig. 6. Response of micropolar elastic half space for unit impulse micro-moment (Myx).

Fig. 7. Response of micropolar elastic half space for unit impulse micro-moment (Myy).

It will be interesting here to compare the simulated displacements with linear elastic half-space solution available in Aki
and Richards [26]. Accordingly the displacement (uz) time history is simulated for an unit impulse force (Fyz) applied at a
depth of 10 km (Y0 = 10 km) for Reduced micropolar half space with j = 0 (linear elastic half-space), Reduced micropolar half
space and Micropolar half space. The material constants are kept same for all the four cases. The ground displacements are
estimated at four locations on the surface of the half-space. Fig. 9 shows the comparison of displacement time histories for all
the three mediums. The simulated displacement time history of reduced micropolar for j = 0 is also valid for a linear elastic
medium. The peak amplitude values for linear elastic and reduced micropolar elastic half space are close to each other. The
micro effects on the simulated displacement time histories are more pronounced in the full micropolar medium compared to
reduced micropolar medium. The peak displacement in a full micropolar elastic medium is also higher than compared to the
reduced and linear elastic half-space.
 M. Gade, S.T.G. Raghukanth / Applied Mathematical Modelling 39 (2015) 72447265 7263

Fig. 8. Response of reduced micropolar elastic half space for unit impulse force (Fyz).

Fig. 9. Comparison of displacements for unit impulse force (Fyz).

In addition to Lames constants only two additional material constants rotational inertia (j) and kappa (j) are
required to dene reduced micropolar medium. In this present study j = 0.196e6 m2 is used to simulate response of
the reduced micropolar medium. For earth medium j value may be very high because of the presence of big stones, hard
rocks and heterogeneity. Ground motions are simulated for four different values of j whose order varies from 106 to
104 m2 to understand the effect of j on ground motions. The simulated ground motions are presented in Fig. 10. It
can be observed form Fig. 10, that the effect of j on translation motion (uz) is negligible. The rotational motions increase
with increase in the rotational inertia of the medium. The effect of rotational inertia is observed only in rotational
motions. The rotational inertia (j) and kappa (j) values for reduced micropolar and additional three material constants
(a, b and c) for full micropolar earth medium can be estimated once sufcient recorded rotational ground motions data
becomes available.
7264 M. Gade, S.T.G. Raghukanth / Applied Mathematical Modelling 39 (2015) 72447265

Fig. 10. Effect of rotational inertia (j) on ground motions.

7. Summary and conclusions

In this article fundamental solutions for both micropolar and reduced micropolar half space have been presented. The
problem is divided into plane strain and anti-plane strain by assuming that the waves are traveling in XY plane only.
Greens functions for buried concentrated force and micro-moment have been derived. Fourier integral representation is used
to derive the fundamental solutions. In addition to longitudinal and shear waves, micro-rotational waves are been observed
in micropolar theory. A total of six material constants are required to fully dene an elastic micropolar medium. For reduced
micropolar half-space the number of constants required to dene the medium is three. Given these material constants one
can simulate translation and micro-rotation time histories for a given unit impulse force and micro-moment. The derived
analytical expressions can be used to simulate displacement eld at the ground surface due to an earthquake with known
rupture characteristics. The earthquake sources are generally modeled as a moment tensors. The advantage with micropolar
theory is that one can apply directly the impulsive micro-moments in the half space. This theory can be used to simulate
both rotational and translation ground motions for earthquakes. Discrete wave number method can be used to numerically
evaluate the innite integrals (Eq. (121)) for estimating displacement and rotation time histories. The closed form analytical
expressions of frequency wave number spectra obtained in this study can be directly used to understand the temporal and
spatial variability of seismic ground motion. The corresponding correlation characteristics of ground motions can be
estimated from the frequency wave number spectrum. Numerical results have been presented for anti-plane strain case.
The displacement and rotational time histories due to a unit impulsive force and micro-moment are simulated for both
reduced and full micropolar medium (Figs. 58).
In the present study, expressions for surface displacement eld have been derived using wave propagation in a single
layered elastic half-space. These expressions can be further used to obtain Greens functions for layered half-space using
propagator based formalism [6].
The major limitation with the micropolar theory is that additional material constants and rotational inertia of the
medium are required to simulate ground motion. No studies have been done till date to estimate these constants for rock
and soil medium. However with recent developments in strong motion instrumentation, it has become possible to measure
rotational motions during the earthquake. With the help of strong motion data, one can estimate the micropolar material
properties of the rock medium.

References

[1] H. Lamb, On the propagation of tremors over the surface of an elastic solid, Philos. Trans. R. Soc. 203 (1904) 142.
[2] M. Bouchon, A simple method to calculate Greens functions for elastic layered media, Bull. Seismol. Soc. Am. 71 (1981) 959971.
[3] J.E. Luco, R.J. Apsel, On the Greens functions for a layered half-space. I, Bull. Seismol. Soc. Am. 73 (1983) 909929.
[4] B. Chouet, Representation of an extended seismic source in a propagator-based formalism, Bull. Seismol. Soc. Am. 77 (1987) 1427.
[5] R. Zhang, L. Zhang, M. Shinozuka, Seismic waves in a laterally inhomogeneous layered medium. I: theory, J. Appl. Mech. ASME 64 (1997) 5058.
[6] S.T.G. Raghukanth, Source mechanism model for ground motion simulation, Appl. Math. Model. 32 (2008) 14171435.
 M. Gade, S.T.G. Raghukanth / Applied Mathematical Modelling 39 (2015) 72447265 7265

[7] S.H. Hartzell, T.H. Heaton, Inversion of strong ground motion and teleseismic waveform data for the fault rupture history of the 1979 Imperial valley,
California, earthquake, Bull. Seismol. Soc. Am. 73 (6) (1983) 15531583.
[8] S.T.G. RaghuKanth, R.N. Iyengar, Strong motion compatible source geometry, J. Geophys. Res. 113 (2008) B04309.
[9] R. Yamaguchi, T. Odaka, Field study of the Izu-Hanto-oki earthquake of 1974, Spec. Bull. Earthquake Res. Inst. Univ. Tokyo 14 (1974) 241255.
[10] V.W. Lee, M.D. Trifunac, Rocking strong earthquake accelerations, Int. J. Soil Dyn. Earthq. Eng. 6 (2) (1987) 7589.
[11] J.E. Luco, Torsional response of structures to obliquely incident seismic SH waves, Earthq. Eng. Struct. Dyn. 4 (1976) 207219.
[12] M. Takeo, Ground rotational motions recorded in near-source region of earthquakes, Geophys. Res. Lett. 25 (1998) 789792.
[13] H. Igel, U. Schreiber, A. Flaws, B. Schuberth, A. Velikoseltsev, A. Cochard, Rotational motions induced by the M8.1 Tokachi-oki earthquake, September
25, 2003, Geophys. Res. Lett. 32 (2005) L08309.
[14] W. Suryanto, H. Igel, J. Wassermann, A. Cochard, B. Schuberth, D. Vollmer, F. Scherbaum, U. Schreiber, A. Velikoseltsev, First comparison of array-
derived rotational ground motions with direct ring laser measurements, Bull. Seismol. Soc. Am. 96 (6) (2006).
[15] M. Bouchon, K. Aki, Strain, tilt, and rotation associated with strong ground motion in the vicinity of earthquake faults, Bull. Seismol. Soc. Am. 72 (5)
(1982) 17171730.
[16] A.C. Eringen, Linear theory of micropolar elasticity, J. Math. Mech. 15 (1966) 909923.
[17] M.A. Kulesh, V.P. Matveenko, I.N. Shardakov, Construction and analysis of an analytical solution for the surface Rayleigh wave within the framework of
the Cosserat continuum, J. Appl. Mech. Tech. Phys. 46 (2005) 556563.
[18] M.A. Kulesh, V.P. Matveenko, I.N. Shardakov, Propagation of surface elastic waves in the Cosserat medium, Acoust. Phys. 52 (2006) 186193.
[19] M. Kulesh, Waves in linear elastic media with microrotations, part 1: isotropic full cosserat model, Bull. Seismol. Soc. Am. 99 (2009) 14161422.
[20] E.F. Grekova, M.A. Kulesh, G.C. Herman, Waves in linear elastic media with microrotations, part 2: isotropic reduced cosserat model, Bull. Seismol. Soc.
Am. 99 (2009) 14231428.
[21] A.C. Eringen, Linear theory of micropolar viscoelasticity, Int. J. Eng. Sci. 5 (1967) 191204.
[22] C.B. Kafadar, A.C. Eringen, Micropolar media-I the classical theory, Int. J. Eng. Sci. 9 (1971) 271305.
[23] V.R. Partt, A.C. Eringen, Reection of plane waves from at boundary of a micropolar elastic half-space, J. Acoust. Soc. Am. 45 (1969) 12581272.
[24] A.C. Eringen, Microcontinuum Field Theories: Foundations and Solids, vol. 487, Springer, New York, 1999.
[25] L.M. Schwartz, D.L. Johnson, S. Feng, Vibrational modes in granular materials, Phys. Rev. Lett. 52 (1984) 831834.
[26] K. Aki, P.G. Richards, Quantitative Seismology: Theory and Methods, vol. I, WH Freeman, New York, 1980.
[27] M. Bouchon, A review of the discrete wave number method, J. Geophys. Res. 160 (2003).