1 s2.0 S0267726111001278 Main

Soil Dynamics and Earthquake Engineering 31 (2011) 11591170
Contents lists available at ScienceDirect
Soil Dynamics and Earthquake Engineering

journal homepage: www.elsevier.com/locate/soildyn
Pilesoilpile interaction in pile groups with batter piles under

dynamic loads
Hasan Ghasemzadeh, Mehrnaz Alibeikloo n
K.N. Toosi University of Technology, Civil Faculty, Valie Asr, Tehran, Iran
a r t i c l e i n f o
abstract
Article history:
Received 28 July 2010
Received in revised form
16 April 2011
Accepted 18 April 2011
Available online 10 May 2011
A simple analytical solution is developed for computing the dynamic interaction tensor for oating pile
groups with batter piles. For this purpose, the governing differential equations are derived for an
unloaded batter oating pile closely spaced to another loaded pile with the same properties. The
reaction of soil against pile deformation is modeled by the springs and dashpots along the length of the
pile. The soil is assumed linear viscous elastic and the pile behavior is linear elastic. The closed form
solutions of governing equations are obtained using appropriate boundary conditions. The results are
veried and compared with rigorous and approximate analytical solutions for vertical piles. The
proposed method can be readily applied by engineers in the design of pile groups with batter piles.
& 2011 Elsevier Ltd. All rights reserved.
1. Introduction
In structures to support lateral dynamic loads like platforms,
bridges and machinery foundations batter piles are usually used.
Since the piles are used in groups, it is necessary to consider the
effect of each pile on another pile in the group. For this purpose
the pilesoilpile interaction factor can be used in the form of the
ratio of the displacement of an unloaded pile to the displacement
of a loaded pile because of soil deformation and oscillation.
Dynamic behavior of piles is far from completely understood as
the soilpile interaction is very complex. Due to complexity of
pilesoilpile interaction, there are no readily applicable methods
available that would include values of dynamic interaction factors
especially in the presence of batter piles.
Poulos [1,2] introduced the concept of interaction factors.
Interaction factors for each degree of freedom of the pile head
have been obtained by resource to integral equation-based
methods [35] and nite element formulations [6,7], as well as
by using simple but physically sound approximations [8,9].
Static interaction factors are not applicable to the dynamic
analysis of pile groups, except perhaps at very low frequencies of
oscillation. Indeed, dynamic studies of pile groups have demonstrated
that the dynamic response of pile groups may differ substantially
from their static response. Nevertheless, Kaynia and Kausel [10,11]
have shown that even for dynamic loads, Poulos superposition
procedure remains an excellent engineering approximation, provided
of course that dynamic interaction factors are used for each frequency
of interest. Dobry and Gazetas [12] have developed a simple
Corresponding author. Tel.: 989121698286.

E-mail addresses: ghasemzadeh@kntu.ac.ir (H. Ghasemzadeh),
Mehrnazalibeikloo@yahoo.com (M. Alibeikloo).
0267-7261/$ - see front matter & 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.soildyn.2011.04.005
analytical solution for computing dynamic impedances of oating

rigidly capped pile groups with consideration to pilesoilpile interaction. Despite its simplicity, the results of that method were in
reasonable agreement with more rigorous solutions. Cairo et al. [13]
method makes use of the closed-form stiffness matrices derived by
Kausel and Roesset [14] to analyze pile groups under vertical
harmonic vibration in layered soil.
Today several numerical methods have been done for dynamic
analysis of piles and pile groups [1518].
Mylonakis and Gazetas [19,20] have developed a new analytical
method in order to determine pilesoilpile interaction factor for
axially and laterally loaded vertical piles in layered soil by considering the presence of receiver pile, which is also done by Ghadimi [21]
for axially loaded piles and by Ravanshenas [22] for piles under
lateral harmonic vibrations. It is obvious that these methods present
smaller interaction factors and economical design than other
approaches. Despite the signicant progress in understanding
dynamic pile group behavior, in all methods mentioned above for
calculating dynamic interaction factors, piles are vertical whereas
using batter piles in some pile groups is inevitable. In case of batter
piles, there is an interaction tensor of second order and four
interaction factors should be determined. Therefore, in this research
a practical and simple method has been developed for calculating
dynamic interaction tensor in pile groups with batter piles. The
presence of receiver pile and interaction of receiver pile with the soil
around it is also considered.
2. Pile model
To calculate dynamic pilesoilpile interaction factor of pile
groups with batter piles an one dimensional continuous model is
developed in which the reaction of soil is modeled by springs and
1160
H. Ghasemzadeh, M. Alibeikloo / Soil Dynamics and Earthquake Engineering 31 (2011) 11591170

(i t)
on the other pile (source pile)
V .e
P .e (i t)
Wj Wjj Ad Wii
Q . e (i t )
(i t)
H.e
x
K
Kx
z
C
Fig. 1. Springs and dashpots for pilesoil interaction.
dashpots along the length of the pile (Fig. 1). The soil is assumed
linear viscous elastic, homogeneous and isotropic and the pile
behavior is linear elastic. Piles are cylindrical, compressible and
inclined in a same plane. Dynamic loads are assumed harmonic in
axial and lateral. In batter piles axial and lateral loads on a source
pile (loaded pile) causes a displacement on a receiver pile, which
can be resolved in axial and lateral direction of the receiver pile.
The stiffness of these springs is complex functions of soil and
piles parameters. The real parts of these functions are soil
stiffness and the imaginary parts present soil damping. Load
distribution along batter pile length under axial load is assumed
similar with those for vertical pile. This assumption is acceptable
for practical inclination angles (less than 301) [5]. Stiffness
functions dened as [23]
Kz0 Gs Sw1 iSw2
Sw1 2pa0
J1 a0 J0 a0 Y1 a0 Y0 a0
J0 2 a0 Y0 2 a0
1
2a
4
2b
J02 a0 Y02 a0
p
shear wave velocity; Gs is the
In which Vs Gs =rs is the
p
shear modules of the soil; i 1; r0 is the pile radius; rs is the
soil mass density; a0 is the r0o/Vs is the dimensionless frequency;
o is the loading frequency; J0 and J1 are the rst kind of Bessel
function of the orders zero and one, respectively; Y0 and Y1 are the
second kind of Bessel function of the orders zero and one,
respectively
Sw2
Kx0 pGs a20 T

T
a0
4K1 b0 K1 a0 a0 K1 b0 K0 a0 b0 K0 b0 K1 a0

b0 K0 b0 K1 a0 a0 K1 b0 K0 a0 b0 a0 K0 b0 K0 a0
a0 i
p
1 i2bs
a0 i
b0 p
Z 1 i2bs
s
21ns
Z
12ns
where Wj is vector of total displacement of receiver pile, Wjj is

vector of displacement of receiver pile due to its loading, Wii is
vector of displacement of source pile due to its loading and Ad is
dynamic interaction tensor of second order. The relation above
can be written in the form below
( ) (
) "
#(
)
wj
wjj
wii
aaa ana
6
uj
ujj
uii
aan ann
where wj and uj are total axial and lateral (normal) displacement
of batter pile. wjj is axial displacement of this pile (receiver pile j)
due to its axial load and ujj is lateral displacement of this pile due
to its lateral load. wii is axial displacement of source pile (pile i)
and uii is lateral displacement of source pile. aaa is the dynamic
axialaxial interaction factor of pilesoilpile; aan is the dynamic
normalaxial interaction factor of pilesoilpile. ann is the
dynamic normalnormal interaction factor of pilesoilpile; ana
is the dynamic axialnormal interaction factor of pilesoilpile.
These interactions coefcients will be determined here.
3. Axial pilesoilpile interaction factor

When pile groups with batter piles are under axial harmonic
loads, axial displacement of source pile (loaded pile) causes both
axial and lateral displacement of receiver pile. Therefore, it is
necessary to dene two kinds of interaction factors for batter pile
groups under axial harmonic loads. These dynamic interaction
factors are: axialaxial interaction factor (aaa) and normalaxial
interaction factor (aan). Calculating dynamic interaction factors
for axially loaded batter pile groups has four consecutive steps. In
all steps, the formulation can be developed for vertical piles when
c 0.
3.1. Axial displacement of source pile
The source pile is subjected to an axial harmonic load at its
head. Axial displacement is obtained by considering one element
of a pile as illustrated in Fig. 2. Equilibrium equation is written
along the pile length
@
@2 w
@w
Kz0 w
AP s rp AP 2 c
@z0
@t
@t
7a
@2 w11 z0 ,t
@w11 z0 ,t
@2 w11 z0 ,t
Ep Ap
c
Kz0 w11 z0 ,t
2
@t
@t
@z02
7b
4a
Ap
4b
K
z
4c
p .
Ap .
z
4d
p
In which bs is the soil damping; i 1; ns is Poissons ratio of
soil; K0 and K1 are the modied second kind of Bessel function of
the orders zero and one, respectively.
In the system of two batter piles, we suppose that the
displacement of each pile (receiver pile) consists of two parts.
displacement due to its loading and displacement due to loading
dz
Ap
-
(A
p )d
z
z
Fig. 2. An element of batter pile with Winklers spring and damper for modeling
batter pile subjected to harmonic axial load.
1161
In above equations w11(z0 ,t) is axial displacement of the batter

pile, L is pile length, AP is cross sectional area of the pile, EP is
elasticity modulus of pile, m is pile mass per length and c is
coefcient of pile internal damping
inclination angles have positive signs. In the case that source pile
and receiver pile have the same inclination angle in value but
with different sings, piles are parallel.
w11 z0 ,t w11 z0 expiot
3.3. Modication of soil displacement at receiver pile position
By replacing Eq. (8) in (7a)

@2 w11 z0
Kz0 w11 z0 0
o mw11 z iocw11 z Ep Ap
@z0 2
2
The solution of this differential equation that is axial displacement of frictional pile can be written as
0
w11 z0 A11 eLz B11 eLz
10a
s
Kz0 ioco2 m
L
Ep Ap
10b
In Eq. (10b), A11 and B11 are calculated by using boundary

conditions in which axial force is P at the pile head and is zero at
the pile tip.
@2 w21 z0
@w21 z0
@2 w21 z0 ,t
EP AP
c
Kz0 w21 z0 ,t
@t
@t 2
@z02
w11 z0 ,s,tcosc1 c2 0
11a
P
1
eLL
B11
EP AP L eLL eLL
11b
3.2. Soil displacement at receiver pile position

At the location of the receiver pile without considering pile
effects, the soil displacement can be obtained by Eq. (12a) in
which the attenuation function j(s) is simply equal to dynamic
interaction factor proposed by Dobry and Gazetas [12]
w11 z0 ,s,t w11 z0 ,tjs
12a
r

r
bs os
ios
js 0 exp
exp
Vs
Vs
s
12b
2 or0
a0
Vs
12c
P.e (it
s
z
2
Fig. 3. Soil settlement along axial direction of source pile.
L
2
Gz0 A11 eLz B11 eLz A21 eLz B21 eLz
Kz0
o m ioc Kz0
2
r

r0
os
exp
i bs cosc1 c2
Vs
s
14a
14b
In Eq. (14a), A21 and B21 are calculated by using boundary

conditions in which axial force at the receiver pile head and pile
tip is zero.

G LLA11 eLL B11 eLL A11 B11 eLL
A21
15a
2
eLL eLL
B21
G LLA11 eLL B11 eLL A11 B11 eLL

2
eLL eLL
15b
Dynamic axialaxial interaction factor is dened as
aaa
It has been assumed for simplicity that the soil wedge between
ground surface and radial line of source pile head has a rigid body
motion equals to the soil under this wedge (Fig. 3). Fig. 3 also
shows that the sign of c is positive when piles are battered
outside the vertical direction. So in this paper we assume that
13
Solution of the above equation can be obtained by

w21 z0
P
1
eLL
A11
EP AP L eLL eLL
1
The presence of receiver pile modies (usually reduces) the

above displacement. Soil settlement in the axial direction of
source pile induces axial and lateral displacement in receiver
pile. In this step, axial displacement of receiver pile is calculated.
With the spring reaction being proportional to the relative
displacement of receiver pile respect to soil, w21 z0 ,t
w11 z0 ,t,scosc1 c2 , axial equilibrium of an element of receiver
pile is written as
w21 z0 0
w11 z0 0
16
The results are compared with approximate solution obtained

by Dobry and Gazetas [12], rigorous solution of Kaynia and Kausel
[11] and analytical method of Mylonakis and Gazetas [19] for
different piles spacing of vertical piles. As can be seen in Figs. 46,
the results obtained by present study are less than Dobry and
Gazetas [12] interaction factor because in Dobry and Gazetas
method the presence of receiver pile is not considered and so
their method is just suitable for short piles. The interaction factors
introduced by Dobry and Gazetas are used as soil displacement in
the location of receiver pile at the third step of this research. So
the method presented in this study leads to more economical
design in pile groups and this method is in acceptable agreement
with the rigorous results of Kaynia and Kausel [11]. The difference
between present method and rigorous solution of Kaynia and
Kausel can be due to the differences of substantial assumptions in
the two theoretical solutions. Moreover, in Kaynia and Kausel
method the presence of receiver pile is not considered. Since piles
are long and exible pile length interaction with soil is considerable and it reduces soil displacement around receiver pile even for
axial displacement so the value of interaction factors obtained in
this research are a little different with those obtained by rigorous
solution of Kaynia and Kausel [11]. As it has shown in Figs. 46,
the value of interaction factors calculated in this research are very
close to those obtained by Mylonakis and Gazetas [19] because in
both methods the presence of receiver pile is considered and both
methods have introduced dynamic interaction factors of Dobry
and Gazetas [12] as attenuation function. A little difference
between present study and analytical method of Mylonakis and
Gazetas is due to different complex-valued impedance
1162
Present Study
0.8
Real Part of aa
Real Part of aa
0.6
0.2
Present Study
Dobry & Gazetas (1988)
Kaynia & Kausel (1982)
Mylonakis & Gazetas (1998)
-0.2
-0.6
0.1
0.3
0.5
0.7
Dimensionless Frequency (a0)
0.9
0.1
0.3
0.9
Dobry & Gaztetas (1988)

0.4
-0.4
0.1
0.3
0.5
0.7
0.9
0.5
0.7
Imaginary Part of aa
0.1
-0.3
-0.7
Present Study
-1.1

Present Study
0.8

0.4
-1.5
-0.4
0.1
Fig. 4. Comparison of axialaxial interaction factor between present analysis and
other methods (Ep/Es 1000, L/d 20, rs/rp 0.8, ns 0.4, bs 0.05, S/D 2).
0.3
0.5
0.7
0.9
Fig. 6. Comparison of axialaxial interaction factor between present analysis and

0.8
Present Study
Real Part of aa

Mylonakis and Gazetas (1998)
0.4
-0.4
0.1
0.3
0.5
0.7
0.9
0.8
3.4. Lateral displacement at receiver pile position
Present Study
In this step, lateral (normal) displacement of receiver pile due

to axial displacement of source pile is calculated. It is obvious that
for vertical piles and parallel batter piles there is not such
interaction. With the spring reaction being proportional to the
relative displacement, u21 z0 ,tw11 z0 ,t,ssinc1 c2 , equilibrium of an element of receiver pile is written in x0 direction as
(Fig. 9)
0.4
representing soil stiffness. This paper utilizes frequency-dependent spring stiffness which is introduced by Novak et al. [23]
whereas in Mylonakis and Gazetas method the nite-elementbased spring stiffness of Makris and Gazetas [24] is used.
In Figs. 7 and 8, dynamic axialaxial interaction factors
obtained by Eq. (16) have been shown for different batter angles
and piles spacing for batter piles. It has been assumed that batter
angle of source pile and receiver pile is the same (c1 c2). As it is
obvious in Fig. 7 by increasing inclination angle, axialaxial
interaction factor is decreased. The other effective parameter in
interaction factor is piles spacing, as it has been shown in Fig. 8 by
increasing piles spacing, pilesoilpile interaction is being
deceased.
SFx0 0-V dV V m
-0.4
0.1
0.3
0.5
0.7
0.9
Fig. 5. Comparison of axialaxial interaction factor between present analysis with


d2 u21
du21
dz0
dz0 c
2
dt
dt
Kx0 u21 w11 z0 ,s,tsinc1 c2 dz0 0
SMy0 0-dM Vdz0
17a
17b
1163
0.2
0.4
S/D = 3
=0
S/D = 5
= 10
= 20
0.2
Real Part of aa
Real Part of aa
= 30
0
-0.1
-0.2
-0.2
-0.4
0.1
0.3
0.5
0.7
0.9
0.1
0.3
0.9
0.1
0.7
-0.2
=0
= 10
-0.3
0.5
-0.1
= 20
= 30
-0.1
S/D = 3
-0.2
S/D = 5
S/D = 10
-0.3
0.1

Fig. 7. Dynamic axialaxial interaction factor for different values of batter angles
(Ep/Es 1000, L/d 25, rs/rp 0.7, ns 0.4, bs 0.05, S/D 3).
8

dV
0
_
< dz
0 mu 21 c u 21 Kx0 u21 w11 z ,s,tsinc1 c2
d2 M
dz0 2
EP IP
d4 u
17c
21
dz0 4
2

d4 u21
d u21
du21
-EP IP
m
c
Kx0 u21 w11 z0 ,s,tsinc1 c2 0
04
2
dt
dz
dt
18
u21 z0 ,t U21 z0 expiot
19
By replacing Eq. (19) in (18)

-EP IP
d4 U21
Kx0 mo2 icoU21 Kx0 w11 z0 ,ssinc1 c2
dz04
20
The solution can be given by

0
U21 z0 A21 elz coslz0 B21 elz sinlz0 C21 elz coslz0 D21 elz sinlz0
0
GA11 eLz B11 eLz
21
In which A11 and B11 are obtained based on Eqs. ((15a) and
(15b)) and other parameters dened as
Kx0 jssinc1 c2
EP IP L4 Kx0 mo2 ico
js
0.9
-0.4
dV
: dz
0
0.3
0.5
0.7
0.1
S/D = 10
0.1
r

r0
bs os
ios
exp
exp
Vs
Vs
s
22a
22b
0.3
0.5
0.7
0.9
Fig. 8. Dynamic axialaxial interaction factor for different values of piles spacing
Ep =Es 1000, L=d 25, rs =rp 0:7, ns 0:4, bs 0:05, c1 c2 303 .

1=4
Kx0 mo2 ico
4EP IP
22c
By applying boundary conditions for receiver pile, other

coefcients of Eq. (21) are obtained. In this research it has been
assumed that the pile cap has lateral deection without any
rotation (the pile is xed to the cap) and piles are long and
exible.
(
U21 0
z0 C 123a
-A21 B21 0
21
y21 dU
0
dz0
8
8
>
>
< y21 0
< C21
z0 0d3 U21
V
E
I
0
>
>
P P dz0 3
: 21
: D21
GLB11 A11 L2 2l2

3
4l
GLB11 A11 L2 2l2
3
4l
23b
Dynamic normalaxial interaction factor is dened as:
aan
u21 z0 0
w11 z0 0
24
In Figs. 10 and 11, dynamic normalaxial interaction factors

obtained by Eq. (24) have been shown for different batter angles
and piles spacing for batter piles. In Fig. 10, as it was being
expected by increasing inclination angles, lateral displacement of
receiver pile due to axial displacement of source pile is increased.
So when using batter piles in piles groups, this kind of interaction
factor should be considered. Fig. 11 shows that by increasing piles
spacing, dynamic normalaxial interaction factor is deceased.
1164
0.4
S/D = 3
S/D = 5
Fig. 9. An element of receiver pile for considering normal deection.
Real Part of an
0.2
S/D = 10
-0.2
0.4
= 10
= 20
= 30
0.3
0.5
0.7
0.3
-0.2
-0.4
0.1
0.5
0.7
0.3
0.9
0.1
-0.1
S/D = 3
-0.3
S/D = 5
0
0.1
Imaginary Part of an
0.9
Imaginary Part of an
Real Part of an
0.2
-0.4
0.1
0.3
0.5
0.7
S/D = 10
0.9
-0.5
0.1
-0.1
0.3
0.5
0.7
0.9
Fig. 11. Dynamic normalaxial interaction factor for different values of piles
spacing Ep =Es 1000, L=d 25, rs =rp 0:7, ns 0:4, bs 0:05, c1 c2 303 .
-0.2
= 10
-0.3
= 20
= 30
-0.4
Fig. 10. Dynamic normalaxial interaction factor for different values of batter
angles (Ep/Es 1000, L/d 25, rs/rp 0.7, ns 0.4, bs 0.05, S/D 3).
Fig. 12. An element of batter pile and Winklers spring and dashpot for modeling
batter pile subjected to lateral load.
4. Lateral pilesoilpile interaction factor

In this research lateral loads refer to loads which are normal to
piles axes. So when source pile is subjected to lateral load, there
are two kinds of interaction factors: normalnormal interaction
factor (ann) and axialnormal interaction factor (ana). The procedure for calculating these interaction factors has also four consecutive steps and the results can be used for vertical piles by
replacing c 0.
of a pile as illustrated in Fig. 12. Equilibrium equation is written

along the pile

2

d u11
du11
0
dz0 Kx0 u11 dz0 0
dz
SFx0 0-V dV V m
c
dt
dt 2
25a
SMy0 0-dM Vdz0

4.1. Lateral displacement of source pile
The source pile is subjected to harmonic lateral load at its head
and lateral displacement is obtained by considering one element
8
dV
_
< dz
0 mu 11 c u 11 Kx0 u11
dV
: dz
0
d2 M
dz0 2
EP IP ddz0u11
4
25b
25c
-EP IP
2

d4 u11
d u11
du11
Kx0 u11 0
c
dt
dz04
dt 2
26a
u11 z0 ,t U11 z0 expiot
26b
Boundary conditions for receiver pile under lateral load are in

the form below
(
U11 0
z0 C 129a
-A11 B11 0
11
y11 dU
0
dz0
By replacing Eq. (26b) in (26a)
d4 U11 Kx0 mo2 ico
U11 0
EP IP
dz04
27
The solution for Eq. (27) is
1165
z0 0-
y11 0
3
V11 EP IP ddzU0 311
(
C11 D11 U0
- U
Ql
Q
0
K 0 mo2 ico
29b
where Q is lateral load acts on the pile head.
U11 z0 A11 elz0 coslz0 B11 elz0 sinlz0 C11 elz0 coslz0 D11 elz0 sinlz0

1=4
Kx0 mo2 ico
4EP IP
28a
4.2. Soil displacement at receiver pile position
28b
At the location of the receiver pile, if this pile were not present,
the soil displacement would be [12]
uS z0 ,s,t, y u11 z0 ,tfs, y
30a

s
0:5
bs a0 s
i:a0 s
exp
exp
D
r D
r D
30b

p
s
0:5
s
exp ia0
2
exp bs a0
D
D
D
2
30c
fs,0 2
f s,
sin2 y
2
fs, y js,0cos2 y j s,
Fig. 13. The location of piles in pile group under lateral load.
30d
In which r VLa/VS 3.4/p(1 ns) and VLa is Lysmer velocity. As

it has been shown in Fig. 13, y is the angle between the line of two
piles and the direction of the horizontal applied force.
1.2
0.4
Present Study
0.8
Real Part of nn
Real Part of nn
0
Present Study
-0.4
0.4

-0.8
Imaginary Part of nn
0
0.1
0.3
0.5
0.7
Dimensonless Frequency (a0)
0.9
0.3
0.9
0.5
0.7
-0.4
Present Study
-0.8

0.1
0.3
0.5
0.7
0.9
0.4
0.1
Present Study
0.2

0
-0.2
-0.4
-1.2
Dimensonless Frequency (a0)
Fig. 14. Comparison of normalnormal interaction factor between present analysis and other methods for vertical piles (Ep/Es 1000, L/d 20, rs/rp 0.7, ns 0.4,
bs 0.05, S/D 2, y 0).
0.1
0.3
0.5
0.7
0.9
Fig. 15. Comparison of normalnormal interaction factor between present analysis and other methods (Ep/Es 1000, L/d 20, rs/rp 0.7, ns 0.4, bs 0.05, S/D 5,
y 0).
1166
4.3. Modication of soil displacement at receiver pile position
receiver pile is zero.
The presence of receiver pile modies (usually reduces) the

lateral displacement of soil caused by normal displacement of
source pile. Since soil displacement is in the normal direction of
source pile, it induces axial and lateral (normal) displacement in
receiver pile. In this step, lateral displacement of receiver
pile is calculated. With the spring reaction being proportional
to the relative displacement of receiver pile and soil, u21(z0 ,t) u11
(z0 ,t,s,y)cos(c1 c2), lateral equilibrium of an element of receiver
pile is written as
C 021 D021
EP IP
d4 U21 z0
Kx0 mo2 icoU21 z0 Kx0 js, ycosc1 c2 U11 z0
dz0
31
The solution for above equation is

U21 z0 js, ycosc1 c2 UP UG
UP
32
h
i
QKx0
0
lz0 elz A021 coslz0 B021 sinlz0
Kx0 mo2 ico
33a
A021 0
B021
1
l
2 Kx0 mo2 ico
33b
UG elz C 021 coslz0 D021 sinlz0
33c
In Eq. (33c), C 021 and D021 are calculated by using boundary

conditions in which lateral force and rotation angle at the head of
3
QKx0
B0
2 Kx0 mo2 ico 21
34
Normalnormal interaction factor is dened as (Fig. 13)
ann
u21 z0 0
u11 z0 0
35
The results are compared with other methods for vertical piles
in Figs. 1416. The present method is in acceptable agreement
with Mylanokis and Gazetas [20] method because in both methods the presence of receiver pile and interaction of it with the soil
around is considered. As it has stated in Section 3 of this research,
a little difference between two results is due to difference
between the functions which are used for spring stiffnesses of
soil. The difference between present method and approximate
solution of Dobry and Gazetas [12] and rigorous solution of
Kaynia and Kausel [11] as mentioned before is the result of
considering presence of receiver pile in this research. Moreover,
there are substantial theoretical differences between present
method and rigorous solution of Kaynia and Kausel.
In Figs. 17 and 18, dynamic normalnormal interaction factors
obtained by Eq. (35) are shown for different inclination angles and
piles spacing for batter piles. It is assumed that batter angles of
source pile and receiver pile are the same (c1 c2). As it is
obvious in Fig. 17 by increasing batter angle, normalnormal
interaction factor is decreased. The other effective parameter in
0.4
Present Study
0.4
=0
= 10
= 20
= 30
0.3
Real Part of nn
Real Part of nn
0.8
0.2
0.1
0
-0.4
0.1
0.3
0.5
0.7
-0.1
0.9
0.1
0.5
0.7
0.9
0.8
Present Study
0.1
0.3
0.5
0.7
0.9

Kaynia& Kausel (1982)
0.4
0.3
-0.1
-0.2
-0.4
0.1
0.3
0.5
0.7
0.9
Fig. 16. Comparison of normalnormal interaction factor between present analysis and other methods (Ep/Es 1000, L/d 20, rs/rp 0.7, ns 0.4, bs 0.05, S/D 10,
y 0).
=0
= 10
= 20
= 30
-0.3
Fig. 17. Dynamic axialaxial interaction factor for different values of batter angles
(Ep/Es 1000, L/d 25, rs/rp 0.7, ns 0.4, bs 0.05, S/D 3, y 0).
39b
G Kx0 mo2 ico
39c
S/D = 3
S/D = 5
Real Part of nn
0.1
G=Kz0 Z
G=Kz0 2 Z2
D21
0.2
1167
S/D = 10
2
2EP AP l
Kz0
39d
0
0
w21G A21 eLz B21 eLz
40a
s
Kz0 iocmo2
L
Ep Ap
-0.1
-0.2
0.1
0.3
0.5
0.7
40b
0.9
0.1
S/D = 3
S/D = 5
S/D = 10
0
-0.1
Fig. 19. An element of receiver pile for considering axial deection.
0.02
-0.2
0.1
0.3
0.5
0.7
= 10
= 20
= 30
0.9
interaction factor is pile spacing, Fig. 18 shows that by increasing

piles spacing, pilesoilpile interaction is decreased.
Real Part of na
0.015
Fig. 18. Dynamic normalaxial interaction factor for different values of piles spacing
Ep =Es 1000, L=d 25, rs =rp 0:7, ns 0:4, bs 0:05, c1 c2 303 , y 0.
0.01
0.005
4.4. Axial displacement at receiver pile position

0
In the last step, axial displacement of receiver pile is calculated. There is no such interaction for vertical piles and parallel
batter piles. With the spring reaction being proportional to the
relative displacement of receiver pile and soil, w21 z0 ,t
u11 z0 ,t,s, ysinc1 c2 , equilibrium of an element of receiver pile
is written in z0 direction as (Fig. 19)
@2 w21 z0
Kz0 mo2 icow21 z0 Kz0 U11 z0 js, ysinc1 c2
@z02
36
The solution for Eq. (36) is

w21 z0 js, ysinc1 c2 w21p w21G
w21p
l
Kx0 mo2 ico
elz0 C21 cos lz0 D21 sin lz0

1=4
Kx0 mo2 ico
4Ep Ip

C21
38a
38b
0.9
0.1
0.3
0.9
0.5
0.7
-0.005
-0.01
-0.015
-0.02
= 10
= 20
= 30
-0.025
G=Kz0 D21 1
37
0.3
0.5
0.7
Imaginary Part of na
Ep Ap
0.1
39a
Fig. 20. Dynamic axialnormal interaction factor for different values of batter
angles (Ep/Es 1000, L/d 25, rs/rp 0.7, ns 0.4, bs 0.05, S/D 3).
1168
S/D=3
0.025
0.01
S/D=10
0.015
0.005
-0.005
Imaginary Part of na
Real Part of na
S/D=5
0
-0.01
-0.02
S/D = 3
S/D = 5
S/D = 10
-0.015
0.1
0.3
0.5
0.7
-0.03
0.1
0.9
0.3
0.5
0.7
0.9
Fig. 21. Dynamic axialnormal interaction factor for different values of piles spacing Ep =Es 1000, L=d 25, rs =rp 0:7, ns 0:4, bs 0:05, c1 c2 303 .
0
0.1
Present Study
0.6
0.4
0.2
0
0.1
0.3
0.5
0.7
Dimensionless Frequecy (a0)
0.9
Imaginary Part of uM
Real Part of uM
0.8
0.3
0.5
0.7
0.9
-0.2
Present Study
-0.4

-0.6
Fig. 22. Comparison of auM between present analysis and other methods for vertical piles (Ep/Es 1000, L/d 20, rs/rp 0.7, ns 0.4, bs 0.05, S/D 2, y 0).
0.2
Real Part of uM
0
0
Present Study
-0.2

-0.1
Present Study
-0.2

-0.3
-0.4
0.1
0.3
0.5
0.7
0.1
0.9
0.3
0.5
0.7
0.9
Fig. 23. Comparison of auM between present analysis and other methods for vertical piles (Ep/Es 1000, L/d 20, rs/rp 0.7, ns 0.4, bs 0.05, S/D 5, y 0).
By applying boundary conditions for receiver pile, A21 and B21

are obtained
z0 1-w21 0-A21 0
41a
2
z0 0-P21 z0 Ep Ap
@w21
l
0-B21
C D
@z0
LG 21 21
41b
Dynamic axialnormal interaction factor is dened as
ana
w21 z0 0
U11 z0 0
42
Dynamic axialnormal interaction factors obtained by Eq. (42)

are shown in Figs. 20 and 21 for different inclination angles and piles
spacing for batter piles. Fig. 20 shows that by increasing batter
Present Study
0.4
0.2
5. The effect of moment on interaction tensor of batter piles

0
-0.2
0.1
0.3
0.5
0.7
In order to study the inuence of moment on free-headed

batter piles, the above steps for lateral interaction factors should
be followed with boundary conditions that rotation angle at the
head of the pile is not zero. For calculating lateral displacement of
receiver pile due to moment on free-headed batter pile, Gazetas
et al. [25] suggested reduction factor as bellow
0.9
auM fs, y2
0.2
angles, axial displacement of receiver pile due to normal displacement of source pile is increased. And it is clear in Fig. 21 that by
increasing piles spacing, dynamic axialnormal interaction factor
decreased.
-0.2
Present Study
-0.4

-0.6
0.1
0.3
0.5
0.7
0.9
Fig. 24. Comparison of auM between present analysis and other methods for
vertical piles (Ep/Es 1000, L/d 20, rs/rp 0.7, ns 0.4, bs 0.05, S/D 10, y 0).
0.6
0.6
S/D = 3
Interaction Factor
43
In which auM is interaction factor for lateral displacement of

receiver pile due to moment on free-headed source pile. In this
research the presence of receiver pile and batter pile is considered
and the results are compared with those obtained by approximate
solution of Dobry and Gazetas [12], rigorous solution of Kaynia and
Kausel [11] and Mylonakis and Gazetas [20] method for vertical
piles (Figs. 2224). Since in this research the presence of receiver
pile is considered, the interaction values of present study are less
than those obtained by Dobry and Gazetas and are in reasonable
agreement with rigorous solution of Kaynia and Kausel. In addition,
in present study and Mylonakis and Gazetas [20] method the
presence of receiver pile is considered, so the difference of two
methods is due to different attenuation function and spring stiffnesses of soil. Generally the values of interaction factors for moments
are very small and they can be ignored in calculating interaction
factors specially, for batter piles.
aa
na
nn
an
0.4
0.2
S/D = 5
Interaction Factor
Real Part of uM
0.6
1169
aa
na
nn
an
0.4
0.2
0
0
20
30
10
40
Batter Angle (degree)
Interaction Factor
0.6
S/D = 10
20
30
10
40
aa
na
nn
an
0.4
0.2
10
20
30
40
Fig. 25. Comparison of dynamic interaction factors (L/d 25, Ep/Es 1000, rs/rp 0.7,ns 0.4, bs 0.05, a0 0.5).
1170
6. Comparison of dynamic interaction factors

As it is shown in Fig. 25, interaction factors are compared for
different values of batter angles and pile spaces. It can be
concluded that except axialnormal interaction factor, for common batter angles (151o c o 301) the value of interaction factors
are considerable and should be calculated in interaction tensor for
analyzing pile groups with batter piles. Fig. 25 also shows that
even in S/D 10 pilesoilpile interactions are important and
considering interaction factors for this pile spaces causes more
accurate design of pile groups.
7. Conclusions
A very simple method is developed for estimating dynamic
pilesoilpile interaction factors of piles in piles group with batter
piles. The approach leads to simple analytical expressions, and it
can be easily understood and applied by engineers familiar with
using interaction factors in design of pile groups. The basis of
the model is a generalized Winkler-type model for pilesoil and
pilesoilpile interaction analysis. Extensive comparison with the
analytical results obtained by other results conrmed the validity
of the diagonal factors of interaction tensor. The method
presented herein is considered the presence of receiver pile and
interaction of it with soil. So interaction factors obtained by this
method leads to more economical design.
Present method permits key parameters to be evaluated
through closed-form expressions. As it can be seen in parametric
study part, batter angle is the most effective parameter in
interaction factors of piles group with batter piles. So choosing
an optimum batter angle for pile groups needs more considerations and it should be to consider the displacement of receiver
pile in both direction (axial and normal).
Dynamic interaction tensor for batter piles can also be evaluated easily through applied graphs presented herein as well as
computing via closed-form solutions. By comparing dynamic
interaction factors, importance degree of them in designing and
analyzing batter piles group is obtained.
References
[1] Poulos HG. Analysis of the settlement of pile groups. J Geotechnique
EngASCE 1968;18:44971.
[2] Poulos HG. The displacement of laterally loaded piles. J Soil Mech Found
1971;97(5):71131.
[3] Poulos HG, Mattes NS. Settlement and load distribution analysis of pile
groups. Aust Geomech J 1971;1:1828.
[4] Buttereld R, Banerjee PK. The elastic analysis of compressible piles and pile
groups. Geotechnique 1971;21(1):4360.
[5] Poulos HG, Davis EH. Pile foundation analysis and design. New York: Wiley;
1980.
[6] Naylor DJ, Hooper JA. An effective stress nite element analysis to predict
the short and long term behavior of a pile-raft foundation on London
clay. In: Proceedings of the settlement of structures symposium, Cambridge,
1975.
[7] Qttaviani M. Three-dimensional nite element analysis of vertically loaded
pile groups. Geotechnique 1975;25(2):15974.
[8] Randolph MF, Worth CP. An analysis of vertical deformation of pile groups.
Geotechnique 1979;29(4):42339.
[9] Scott RF. Foundation analysis. Englewood Cliffs: Prentice-Hall; 1981.
[10] Kaynia AM, Kausel E. Dynamic behavior of pile groups. In: Proceedings of the
second international conference on numerical methods in offshore piling,
Austin, Texas, 1982. p. 50932.
[11] Kaynia AM, Kausel E. Dynamic stiffness and seismic response of pile groups.
Research report no. R82-03. Cambridge, MA: Massachusetts Institute of
Technology; 1982.
[12] Dobry R, Gazetas G. Simple method for dynamic stiffness and damping of
oating pile groups. Geotechnique 1988;38(4):55774.
[13] Cairo R, Conte E, Dente G. Analysis of pile groups under vertical harmonic
vibration. J Comput Geotechnique 2005;32:54554.
[14] Kausel E, Roesset JM. Stiffness matrices for layered soils. Bull Seismol Soc Am
1981;71(6):174361.
[15] Kucukarslan S, Banerjee PK. Behavior of axially loaded pile group under
lateral cyclic loading. J Eng Struct 2003;25:30311.
[16] Xiao X, Chi S, Lin G, Alfano J. Simplied method and parametric sensitivity
analysis for soilpile dynamic interaction under lateral seismic loading. J Adv
Build Technol 2002;1:7718.
[17] Padron LA, Aznarez JJ, Maeso O. BEMFEM coupling model for the dynamic
analysis of piles and pile groups. J Eng Anal Boundary Elem 2007;31:
47384.
[18] Tahghighi H, Konagai K. Numerical analysis of nonlinear soilpile group
interaction under lateral loads. J Soil Dyn Earthquake Eng 2007;27:
46374.
[19] Mylonakis G, Gazetas G. Dynamic Vertical vibration and additional distress of
grouped piles in layered soil. Soil Found 1998;38(1):114.
[20] Mylonakis G, Gazetas G. Lateral vibration and internal forces of grouped piles
in layered soil. J Geotechnique EngASCE 1999;125(1):1625.
[21] Ghadimi M. Pilesoilpile interaction factor for piles under statically and
dynamic axial loads. MSc thesis, K.N. Toosi University of Technology, Tehran,
Iran, 2007, in Persian under supervision of Dr. Ghazavi.
[22] Ravanshenas P. Pilesoilpile interaction factor for laterally loaded piles. MSc
thesis, K.N. Toosi University of Technology, Tehran, Iran, 2007, in Persian
under supervision of Dr. Ghazavi.
[23] Novak M, Nogami T, Aboul-Ella F. Dynamic soil reactions for plain strain case.
J Eng Mech DivASCE 1978;104(4):9539.
[24] Makris N, Gazetas G. Dynamic pilesoilpile interaction. Part II: lateral and
seismic response. Earthquake Eng Struct Dyn 1992;21(2):14562.
[25] Gazetas G, Fan K, Kaynia AM, Kausel E. Dynamic interaction factor for oating
pile groups. J Geotechnique EngASCE 1991;117(10):153148.

1 s2.0 S0267726111001278 Main

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

1 s2.0 S0267726111001278 Main

Загружено:

Авторское право:

Доступные форматы

Soil Dynamics and Earthquake Engineering 31 (2011) 11591170

Contents lists available at ScienceDirect

Soil Dynamics and Earthquake Engineering

Pilesoilpile interaction in pile groups with batter piles under

Corresponding author. Tel.: 989121698286.

analytical solution for computing dynamic impedances of oating

H. Ghasemzadeh, M. Alibeikloo / Soil Dynamics and Earthquake Engineering 31 (2011) 11591170

on the other pile (source pile)

Fig. 1. Springs and dashpots for pilesoil interaction.

Kx0 pGs a20 T

4K1 b0 K1 a0 a0 K1 b0 K0 a0 b0 K0 b0 K1 a0

where Wj is vector of total displacement of receiver pile, Wjj is

3. Axial pilesoilpile interaction factor

H. Ghasemzadeh, M. Alibeikloo / Soil Dynamics and Earthquake Engineering 31 (2011) 11591170

In above equations w11(z0 ,t) is axial displacement of the batter

w11 z0 ,t w11 z0 expiot

3.3. Modication of soil displacement at receiver pile position

By replacing Eq. (8) in (7a)

w11 z0 A11 eLz B11 eLz

In Eq. (10b), A11 and B11 are calculated by using boundary

3.2. Soil displacement at receiver pile position

Fig. 3. Soil settlement along axial direction of source pile.

Gz0 A11 eLz B11 eLz A21 eLz B21 eLz

In Eq. (14a), A21 and B21 are calculated by using boundary

G LLA11 eLL B11 eLL A11 B11 eLL

Dynamic axialaxial interaction factor is dened as

Solution of the above equation can be obtained by

The presence of receiver pile modies (usually reduces) the

The results are compared with approximate solution obtained

H. Ghasemzadeh, M. Alibeikloo / Soil Dynamics and Earthquake Engineering 31 (2011) 11591170

Dobry & Gaztetas (1988)

Kaynia & Kausel (1982)

Kaynia & Kausel (1982)

Dimensionless Frequency (a0)

Fig. 6. Comparison of axialaxial interaction factor between present analysis and

Dobry & Gazetas (1988)

3.4. Lateral displacement at receiver pile position

In this step, lateral (normal) displacement of receiver pile due

Kaynia & Kausel (1982)

Mylonakis & Gazetas (1998)

Fig. 5. Comparison of axialaxial interaction factor between present analysis with

Kx0 u21 w11 z0 ,s,tsinc1 c2 dz0 0

SMy0 0-dM Vdz0

H. Ghasemzadeh, M. Alibeikloo / Soil Dynamics and Earthquake Engineering 31 (2011) 11591170

Dimensionless Frequency (a0)

By replacing Eq. (19) in (18)

The solution can be given by

GA11 eLz B11 eLz

By applying boundary conditions for receiver pile, other

GLB11 A11 L2 2l2

Dynamic normalaxial interaction factor is dened as:

In Figs. 10 and 11, dynamic normalaxial interaction factors

H. Ghasemzadeh, M. Alibeikloo / Soil Dynamics and Earthquake Engineering 31 (2011) 11591170

Fig. 9. An element of receiver pile for considering normal deection.

4. Lateral pilesoilpile interaction factor

of a pile as illustrated in Fig. 12. Equilibrium equation is written

SMy0 0-dM Vdz0

H. Ghasemzadeh, M. Alibeikloo / Soil Dynamics and Earthquake Engineering 31 (2011) 11591170

u11 z0 ,t U11 z0 expiot

Boundary conditions for receiver pile under lateral load are in

By replacing Eq. (26b) in (26a)

d4 U11 Kx0 mo2 ico

The solution for Eq. (27) is