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

net/publication/222191485

Analysis of capped pile groups subjected to horizontal and vertical load

Article  in  Computers and Geotechnics · January 2000

DOI: 10.1016/S0266-352X(99)00029-4

CITATIONS READS

102 961

2 authors, including:

John C. Small
The University of Sydney
182 PUBLICATIONS   3,243 CITATIONS   

SEE PROFILE

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

Slope stability View project

Analysis of piled raft load sharing under vertical and horizontal loading View project

All content following this page was uploaded by John C. Small on 14 October 2017.

The user has requested enhancement of the downloaded file.

 / ..

COMPUTERS
AND
GEOTE(HNICs
ET-SEVIER Computers and Geotechnics 26 (2000) I-21
www. elsevier.com/locate/compgeo

Analysis of capped pile groups subjected to

honzontal and vertical loads
H.H. Zhang*, J.C. Small
Department of civil Engineering, university of sydney, NSW 2006, Australia

Received 16 April 1999; received in revised form 17 September 1999; accepted 20 September 1999

Abstract
This paper presents a method of analysis for an off-ground cap supported by piles embed-
ded in a tayereO soil and subjected to horizontal and vertical loads. The cap is modelled as a
thin plate and the piles as elastic beams and the soil is treated as consisting of horizontal
layeri of different materials. Finite element theory is used to analyse the cap and piles while
finite layer theory is employed to analyse the layered soil. Using program APPRAF (Analysis
of piles and Piled RAft Foundations) to carry out the analysis described above, comparisons
of the behaviour of capped pile groups are made and factors affecting the displacements of
capped pile group foundations are examined. Finally, an example related to three types of
,oits wh.re the moduli increase with depth is illustrated. The results show that the present
method is a powerful and useful way to evaluate the behaviour of capped pile foundations
embedded in different types of soils and subjected to both vertical and horizontal loadings.
@) 2000 Published by Elsevier Science Ltd.
All rights reserved'

1. Introduction

Off-ground pile groups with stiff caps are widely used as foundations for struc-
tures and offshore platforms (wharves or drilling platforms). Generally, capped pile
groups will be subjected to both vertical loads transferred from the structure above
and horizontal loads caused by winds, waves, earth pressures or earthquakes. It is
therefore necessary to develop a method which can not only analyse capped pile
group foundations under vertical loading but can also analyse their behaviour when

{< Corresponding author.

0266-352X100/$ - see front matter @ 2000 Published by Elsevier Science Ltd. All rights reserved.
PII: S0 266-3 5 2x(99)00 029 -4
 H.H. Zhang, J.C. Small I Computers and Geotechnics 26 (2000 ) I-2 I

subjected to horizontal loading. Early investigations were generally only related to

gtorrpr of piles with a rigid cap (assuming the displacement of each pile is identical)
or related to a completely flexible cap (assuming the loading acting on each pile is
identical). Gradually, cap behaviour was included. Hongladaromp et al. [1] have
explored a capped pile group with flexibility, but pile-pile and pile-soil surface
interactions were ignored. Kuwabara [2] used a boundary element method to ana-
lyse the behaviour of a capped pile group. Clancy and Randolph [3] and Poulos [4]
presented, respectively, dfl approximate numerical solution mainly based on finite
ilement theory and Mindlin's solution. By introducing finite layer theory, Ta and
Small [5,6] have analysed piled raft systems subjected to vertical loads in layered
soil. H-owever all the previous work on the analysis of piled raft systems has only
been related to vertical loads. In this paper, off-ground capped pile group systems
subjected to both vertical and horizontal loads and embedded in homogenous elastic
soiis or elastic soils where the modulus increases with depth, have been analysed by
combining finite elements to model the cap and the piles with finite layer theory to
model the soil. The method can also be used for on-ground caps, but this is left for
further work.

2. Method of analYsis

For a capped pile group foundation, external loads including vertical and hor-
izontal concentrated or uniform loads and moments in each of the three (orthogo-
nal) axis directions may be applied to the cap and are transferred from the cap to the
piles and then through the piles to the soil. Based on this consideration, analysis of a
capped pile group system is carried out by separating a capped pile group into three
puiL' the cap, the group of piles and the layered soil. The cap is assumed to be a
ihin elasti" plut. utta ttt. finite element method developed by Bogner et al. [7] is
employed toanalyse the cap. Element division of the cap should be such that the pile
head fits within one element of the cap. Each cap element has four nodes and 24
degrees of freedom. The contact forces applied to the cap elements that are con-
n.Jt.d to the pile heads are assumed to be uniform loads on the cap and con-
centrated loads on the piles as shown in Fig. 1.
A pile is modelled as a beam and a simple finite element method is used to analyse
the pile. The part of a pile which is embedded in the layered soil is divided into the
sarne number of elements as the number of layers in the soil while the upper exposed
part of the pile may just be taken as one element or more if it is of some length. A
series of vertical or horizontal ring loads are assumed to act on the soil interfaces
(corresponding to each node along the pile shaft) but a circular load is assumed to
act on the soil at the pile base. Point vertical and horizontal forces are assumed to
act on the pile at each node. Interface forces transferred from the cap to the pile
heads may be simplified into equivalent concentrated loads, and these loads are
considered as external forces for the group of piles. Torsional loadings are not con-
sidered along the pile shafts, and so the analysis is limited to where torsion is not of
major concern.
 H.H. Zhang, J.C. Small lComputers and Geotechnics 26 (2000) I-21

First row of Piles

Pile 1

Rigid base

Fig. l. Schematic diagram of separated cap and piles'

A modified finite layer method based on that developed by Small and Booker [8] is
used to analyse the layered soil. Generally, the layered soil above
the base of a pile is
numer-
divided into 12-15 layers or more based on the soil properties and required
or
ical accuracy, while tle soil beneath the base of the pile may be divided into one
more layers according to the soil properties and strata'

2.1. Analysis of caP

The analysis of the cap and the piles is carried out separately and, in the
structural
free body
analysis, some nodes on the cap must be restrained from undergoing
rotations and translations. For convenience, in the present paper two corner
nodes
of the cap were chosen as points of restraint as shown in Fig' 1' At pin 1, the node is

completeiy fixed in all dirictions (i.e. six freedoms) and pin 2 is fixed only
in the y-
direction to resist the cap from rotating about the z-axis. Based on the above
contain-
assumption, the actual displacement {6r} at the centre of each cap element
ing a pile may be exPressed as

{6,}: t/.ltP,} * lalD,+ {blDy* (c}D,+ {dlT*+ le}0v + {fle' * {d'o} (l)

where

t/,1 : influence matrix of the Pinned caP

{P,} e#i; ii,*1., Mly,'. . . , P'l*, F1r,ry,,tu[i*'fir)', is the v99tor of inter-
iuJ. toiar unO -o*.its on the cap elements (coirtaining a pile), where P'r"
is the uniform interface load in the x-direction on element i, etc'
{6,} : 6f,;i, ;l ,,0i,,*,01r, . . . , 6l*, 6'113']",gi*,.0i)' is the vector of displacements at
ift...itt. of iire Zap elemenis (Containing a pile), where 6i' is the displace-
ment in the x-direciion for element i and 0i, represents the rotation
in the
direction of M* (see Fig' 1) for element i, etc'
 H.H. Zhang, J.C. Small lComputers and Geotechnics 26 (2000) I-21

}: displacements in the centre of the cap elements (containing a pile) under

the
{d,o
uppt.a loads on the pinned cap. The order of unknowns is the same as {6r}'
D* htiizontal translation of the cap at the first pinned point in the x-direction
Dy horizontal translation of the cap at the first pinned point in the y-direction
D, vertical translation of the cap at the first pinned point in the z-direction
0* : the rotation of the cap about the first pinned point in the direction of Mx
0y : the rotation of the cap about the first pinned point in the direction of Mt
0, : the rotation of the cap about the z-axis
{u} : (1,0,0,0,0, 1,0,0,0,b, 1,0,0,0,0,' ):
{b} : (0, l, o, o, o, o, 1,0,0,0,0, 1,0,0,0, )l
{'} : (0,0, 1,0,0,0,0, 1,0,0,0,0, 1,0,0, .... . j)'^.-
{d} : (0,0, xl, 1, 0, 0, 0, x2,1,0, .. . . . .0, 0, xn,]'9);
{.} : (0,0,./1,0, 1,0,0, Y2,0,1,......0,0,/r,0, 1)'
{f} : (r, ,1*r,0, 0, 0, yr,--*,r,0, 0, 0, . . . . . , Yn, -xr,0, 0, 0)T

where

xi, yi : co-ordinates of the central point of cap element i (containing a

pile) relative
to the first Pinned Point,
n : number of cap elements (containing a pile)'

The influence matrix of the pinned cap can be generated as follows' A

unit uni-
that
form horizontal load in the x-direction may be applied to the first cap element
contains a pile and the central displacements (in the x-, y- and z-directions)
plus
and
rotations of all elements containing a pile are computed' These displacements
pinned cap.
rotations will form the first column of the influence matrix [f'] of the
Then a unit uniform horizontal load in the y direction may be applied to
the same
(containing a pile) in
cap element and the central displacements of all cap elements
thi x-, y- andz-directions plus iotations can be calculated. These displacements will
and a unit
form the second column of t1r] and so on for a unit load in the z-direction
moment in the x- and y-directions, respectively'

2.2. Analysis of a group of piles embedded in a layered soil

Behaviour of the soil is here assumed to be elastic (although the

soil may consist
of horizontal layers having different elastic properties). This is a reasonable
assumption since working loads should be kept well below the
failure load and in
is linear'
most Jases, this is within the region where the load-deflection behaviour
Model test results
Thus, elastic theory can be used directly for settlement prediction.
curves of pile groups are roughly linear up to
[9] also show that ihe load-deflection
i;;J1.";iof I to I of the failure load and so in this paper the elastic theorv is used
to analyse the soil behaviour.
For the group of piles, the displacement of the head of pile i due to the
loads and
moments acting there as well as loads applied to the tops of
other piles may be
obtained by superposition using the following equation.
 H.H. Zhang, J.C. Sntall I Contputers and Geotechnics 26 ( 2000 ) I-2 I

6i: D'u't
j: I

where Ii1 is the flexibility coefficient denoting the displacement

at the top of pile i due
to a unit load (or moment) at the top of pile j; P1 is the pile-cap interaction
force
displacement at the
acting on the top of pile j; rz is the total number of piles. So the
as
top o1each pile under the loads transferred from the cap can be expressed
(2)
{6,0 }

where

[r'ol
{rro}
{ Aro } : displacements at the pile heads

The primary problem here is to generate the pile-soil influence matrix.

In thrs
paper , &n interaction method based on the finite layer theory
developed by Small
the layered soil.
and Booker t8l is introdLlced to generate the influence matrix of

2.2.1. Finite laYer theorY

applied
In order to simplify the equations governing the problem of circular loads
field quan-
to a layered soil, double Fourier transformations may be applied to all
this results
tities (displacements and loads). For general vertical and shear loading,
in two seis of finite layer equations that give the relationship between the trans-
formed stresses and transformed displacements at the interface of
each horizontal
soil layer.
The first set of equations may be written in flexibility form
(see [8])

fr'l{r'} : {a'} (3)

or in the altern ate stiffness form

fr'l{a'} : {r'} (4a)

where

at} : (U ro, (l qp, fJ Ug-)t

{
^,
,
i",'): iruo, rsr,-tr; 4*)t
U, (vet-
The transforms of the displacements at the layer interface are designated
p, m desig-
tical), ug (displacement in tle f-direction see Fig. 2) and the subscripts
of the stresses at the
nate'itre upper and lower faces of the soil layer. The transforms
shear stress in the
layer inteifaces are designated N (for vertical stress) and Tq for
upper and lower faces of the soil layer)'
5z-direction (p and tn ugain denote
 H.H. Zhang, J.C. Srnall lComputers and Geotechnics 26 (2000) I-2 I

+ -+ -+
-+ -+ I
-+ h" ---->

Fig. 2. Uniform shear loading over a circular region showing axis systems.

The second set of equations also results for the soil layer (called the uncoupled
terms)
(4b)
tK,,l'{61,} - lPiI
where

{a;} : {IJro,un".,)].
{P'"I : (Tro, -Ta^)'
where U, is the transform of displacement in the 4 direction and T, is the
transform
of the shear in the qz-dftection (see Fig. 2)'
These stiffness matrices may be assembled for each soil layer into
a global matrix,
(Jr,UE,U, at the layer interfaces'
and solved to give the transformed displacements
Upon addition of layer matrices all interface stresses cancel, and so the right-hand
side only consists of ihe transform of the load at the interface
where it is applied.

load-
2.2.1.1. Transform of horizontal loadings. The transform of a horizontal shear
may be shown to
ing in the x-axis diiection and applied uniformly over a circle
,.dr'tt. to the Hankel transform of the load
rr ih"al (pa) (6)
r-rc
p
 H.H. Zhang, J.C. Small I Computers and Geotechnics 26 (2000) I-2 1

where ft. is the shear load. The components of the shear load for use in the finite
layer Eqs. (4a) and (4b) are therefore

,,'p_ih"alt(pa)
"o*
ih"alt(Pa)
,^ _ ,rn,

and the angle e is as shown in Fig. 2.

For a hoiizontal ring load applied in the x-direction, the transform 4,' is given by

iH"_,
: fiJo@a) (1)
T,r

and again TE and T, can be found in terms of the angle e, where

T*" : transform of applied horizontal circular load

T*, : transform of applied horizontal ring load
h": applied horizontal circular load
H,: applied horizontal ring load
a: radius of loaded circle
/o(pa) : Bessel function of the first kind (order zero)
Jt(pa) Bessel function of the first kind (order one)
p: Hankel transform Parameter
i: tr:1 * make the solution for the horizontal load the complex part of the
solution.

2.2.L2. Transforms of vertical loadings. The transforms of the vertical ring loads
pile
that are appfiLd along the pile shaft or the uniform load that is applied at the
base can atso Ue found by using a double Fourier transform. The transform
reduces
to the Hankel transform of the load. For a uniform vertical load q the transform is
qalr@a) (B)
e, - p

and for a ring load

P* (e)
Q' -;ro@a)

where

Q, applierd uniform circular vertical load

Q, total applied ring vertical load
 H.H.Zhang,J-C.SmaltlComputersandGeotechnics26(2000)l-21

q : applied uniform circular vertical load

Pr : total applied ring vertical load.

The solution to the finite layer equations results in the transformed displacements
that may be inverted to yield the actual displacements. The real part is the dis-
placement due to the vertical loading and the complex part is the displacement due
io the horizontal loading. In the case of horizontal loading, the components of dis-
placement from both sets of Eqs (4a) and (4b) need to be combined to compute
the
displacement in say the x- or y-direction'
Th. flnit. layer iechnique can therefore be used to compute the response of the
loads at the
layered soil to ring loads at the soil layer boundaries or uniform shear
pile toe.

2.2.2. Generation of influence matrix of pile-soil system

Similar to the formation of the influence matrix of the cap, the pile-soil influence
matrix, f7rol, *uy be generated by using the interaction method for pile
groups
developed Ay Znung aid Smatl t101. A horizontal load which is equal
to the total
pile I in
horizontal load applied on the corresponding cap element may be applied to
the x-direction ani the displacements used to form the elements of the
first column
tiri. tn u similar *uy, ti. remaining columns of_l/rpl can be generated for loads
"i
in ifr. y- and z-directions and moments in the x- and y-directions. The advantage
of
this mlthod is that it can be used to solve capped pile group problems involving a

large group of Piles.

2.3. Analysis of capped pile groups

If the compatibilitY of displacements and the equilibrium of interaction forces

get
between the Pile heads and the cap are taken into accoutrt, we

{Er} : {a'o}
( 10)

(11)
tP,) - -{P,p}
Combination of Eqs. (l) , (z), (10) and (11) leads to

(l+t+ [r,ol11r,o1 - {alD*- {blDy- {c}D,- {dll*- {elly- lfle': t6.o} 02)
To guarantee force and moment equilibrium of the cap, the following equations
must be satisfied:

{o'l{r'o } : P*
(13)

{u'l{r,o} : Pv
(14)

{t'l{P'o} : P, (15)
 9
H.H.Zhang,J.C.SmalllComputersandGeotechnics26(2000)I-2]

(16)
{r/'}{P,ol - M*
( 17)
{r'}{P,o} - Mr
( 1 B)
{/'}{r,o} - M,

where

{o') (Ar,0,0,0,0, A2,0,0,0,0,,Ar,0,0,0,0)

{b'} (0, At,0,0, 0, 0, A2,0,0, 0, ,0, An,0,0,0)
{t'} (0,0,,4,,0,0, o,o, A2,o, o, .r.... ' o, o,ln, o, o)
{tl'} (0, O,,lit, , At,0,0,0, Azx2, A2,0, ,0,0, Anxn, An,0)
{u') io, o, zili ,0, At,0,0, A2y2,0, Az, , 0,0, Anxn,g, An)

{f'} i.qrrr, JAr*r,0,0,0, Azyz, -A2x2,0, 0,0, , Anl,' -Anx"0' 0' 0)

and where

Ai : thearea of element i in the cap, and P,s, Py, P, ate the total loads applied
to the
cap in the x-, and, z-directions and M', M, are the total moments applied to
!-
thi A, ii the total moment about the z-axis (at pin l) due to P' and Pr'
"up;
Solving Eqs. (12)-(18), we may obtain interaction forces on the pile heads
at the
pile-caplnt"ifacrs, and'the rigid body rotations and translations of the cap relative
io the pins. Substiiuting the risults into Eq. (1) it is possible to work out the actual
external for-
displacements of the ca-p elements that contain piles under the applied
,... By substituting the interaction forces into Eq. (2) it is also possible to calculate
the displacements of the pile heads'

3. Results

For convenience in the analysis of capped pile group foundations subjected to

vertical and horizontal loads, some dimensionless parameters can be defined as
follows
Definition Practical range
SID 2.5^-r8
Pile spacing ratio
Pile slenderness ratio Lr^lD 10-100
Pile-soil stiffness ratio KO, : EplE, 100-10000
Cap-soil stiffness ratio Ktt : E, lE, 100-10000

where Eo is the pile Young's modulus; E, is the soil Young's modulus;

s is spacing
between pile centres; is pile diameter and Z.* embedded pile length' In the fol-
I
lowing analyses, thePoisson's ratio of the cap and soil will be taken as 0'15
and
0.499, respectively, if not stated otherwise'
l0 H.H. Zhctng, J.C. Small I Contputers cmd Geotechnics 26 ( 2000 ) I-21

Uniform horrzontal load Qx or vertical load, is generally taken into account

Qr,
separately in the examples below. The normalised displacement for a capped Pile
group corresponding to horrzontal loads or vertical loads is defined as:
for capped pile groups under horizontal loads

Iu**: u**ErD lq*BrL, (1e)

for capped pile groups under vertical loads

f ur, : (20)
LtzzEtD lQrBrL,

where u** is the displacement of one of the pile heads in the x-direction induced by
horizontal loads; u' is the displacement of one of the pile heads in the z-direction
induced by vertical loads; B, and L, are the breadth and length of the cap, respec-
tively; D is pile diameter.

3.1 . Verification by comparisons with results for pile Sroups

3.1.L Comparisons pile group with two piles

for capped
An obvious characteristic of a capped pile group with only two piles is that when a
symmetric horizontal loading is applied to the cap in the direction perpendicular to
tire line connecting the two piles, the cap has no enhancing effect on the two piles
and the behavioui of the capped pile group is similar to that of a free head pile
group. If the load is applied parallel to the line joining the piles, the cap stiffness will
f,uu.-u large effect on the displacement. For verification of this, the pile group as
shown in the inset to Fig. 3 was analysed, where the ratio of embedded pile length,
L" , to diameter, D, was chosen to be 25. The ratio of the exposed pile length Ilo to
pile diameter was 2.5 and the pile-soil and cap-soil stiffness ratios are taken as 4000
ind 4285, respectively. The results of the analysis of the pile group and capped pile
group with only two piles subjected to horizontal loading in the two directions are
shown in Figs. 3a and b, respectively.
As can be seen in Fig. 3a under horizontal load in the x-direction, as expected, the
normalised displacement of the capped pile group is very close to that of the pile
group with a rigid cap but much lower than that of the pile group with a completely
[.*iUt. cap. Fig. 3a also shows that when pile spacing is less than 6 times the pile
diameter, the disptacement factor is sensitive to change in spacing. However, with
increase of pile spacing, the horizontal displacement is less and is not greatly affected
by the spacing.
Fig. 3t demonstrates that under a horizontal load in the y-direction, there will be
no difference among a capped pile group and two-pile group with a completely
flexible or rigid *ittr respect to normalised displacements. The results of Fig. 3b
indicate that the progtu* is probably working correctly. However, for problems
"up
with more piles, further comparisons are needed.
 H.H. Zhang, J.C. Sntall I Contputers and Geotechnics 26 (2000) I-21 1l

(at 0.6
-e- Capped Pile grouP
''tl" Piles with flexible caP
' ' A ' Piles with rigid caP

.^ xx
0.4

-J HJD= 2.5
0) Lon'/D = 25
E
q,
(J
lrr, = ur*ErD/P"
Js
o 0.3 H/Lom = 100
.9
E
ct
c)
.9
fr
zE
0.2

t 'L_
:ffiT:r x

uxx means disPlacement at the Pile

head in the x direction due to
horizontal load P, in the x direction ml
i
Ho

L"-

69 12

Pile spacing (S/D)

tbt 0.0

-+- Capped Pile grouP

''tl' ' Piles with flexible caP
.. A' ' Piles with rigid caP

\
Y
G
j
0.4

o
E
q)
()
(u

a
a
03
H"/D= 2.5
E
0)
L"JD = 25
.9 luyy = urrEuD/P,
o
E H/Lu,n = 100
2 o.z

uyy means displacement at the pile head

in the y direction due to horizontal load
P, in the y direction

0.0
12

Pile spacing (SlD)

Fig. 3. Comparison of 2-pile group with capped pile group: (a) P" only; (b) Py only.
|2 H.H, Zhatlg, J.C. Snlall lComputers and Geotechnics 26 (2000) I_21

3.1.2. Comparison with pile grottps

In an example given by Poulos [11], there is a fixed-head pile group with 16 piles as
shown in the inset to Fig. 4. The ratio of pile length to diameter is 25 and the pile
flexibility factor Kp is equal to 10-5 (defined as Kp : EpIp/ErL\^, I, is the second
moment of area of the pile). Here the Poisson's ratio for the soil is taken as 0.499
while 0.5 was used by Poulos Ull. EplE, was chosen to be 79.6 while Erf E"was
made large (85 250) to represent a rigid cap so that the behaviour of the capped pile
group would be like that of a fixed-head pile group although for the rigid cap some
head rotations may still occur. The cap is just clear of the ground (i.e. fl6 = 0.0).

(I,
x
o-
x
0-

46
Pite spacing, S/D

pile group'
Fig. 4. Typical horizontal load distributions in fixed head pile group and capped
 H.H. Zhang, J.C. Small I Contputers and Geotechnics 26 (2000) l-21 13

The load distributions against pile spacing are plotted in Fig. 4. We can observe that
for a pile spacing ratio of less than 4 there is poor agreement of the typical hor-
izontal load distributions between the fixed-head pile group and the capped pile
group. Load on piles 3 and 4 computed using the present method is higher than that
given by Poulos I l], while the load on piles I and 2 is lower, especially for a capped
pile group with close pile spacing. With increase of the pile spacing, the load dis-
tributions tend to become more uniform.
Because of the lack of agreement, a further comparison was made with the results
of El Sharnouby and Novak [12] for the same problem. It may be seen in Fig. 4 that
these results are in fairly good agreement with the present results.

0.30

0.25

xx

-a
+t
C
0.20
o
E
o
c)
(U
E.
.9,
E
6 0.1 5
o
N
o
.(f
o
.9,
E
E
l- 0.10
o
z

0.05

0.00
1 E+01 1 E+02 1 E+03 1 E+04 1 E+05 1 g+06

Pile-soil stitfness ratio, Ep/Es

Fig. 5. The effect of pile-soil stiffness ratio on deflection of capped pile groups under horizontal loads.
14 H.H. Zhang, J.C. Sntall I Computers qnd Geotechnics 26 (2000) I-21

3.2. Parametric analysis

In order to demonstrate the use of the present method for the analysis of laterally
or axially loaded capped pile groups, the most critical parameters were examined to
show their effects on the maximum displacement of a capped pile group. The
ratios of the embedded pile length to pile diameter (L" /D) were chosen to be 10,
20, 40,80. The Poisson's ratios of the soil and the cap were assumed to be 0.35
and 0.15, respectively. The ratio of the exposed pile length Ho to pile diameter
was 2.5 and the pile spacing ratio was taken as 5. The thickness of the cap was
0.5 m, while the breadth and length of the cap were chosen to be 18 and 13 m,
respectively.

3.2.1 .
The effect of soil and pile moduhts
Varying the pile-soil stiffness ratio Epf E,from 10r (flexible pile) to 106 (rigid pile)
and keeping the cap-soil stiffness ratio constant (E, / E':
2000), the present program
APPRAF was used to calculate displacements at the central point of each pile head.
The normalised displacements for pile 5 (maximum displacements generally occur
for this pile) are plotted in Figs. 5 and 6 against the pile-soil stiffness ratio.

0.08
urz= displacement at the top of pile No. 5 |

Il-. I Ho/D=z.S
lu.o= u=ErD/QrB,L,
I s/D=s.o
I
I
E/E' = 2ooo
Vs = 0.35
Vr = 0'15
0.06 H/L€m = 100
N
a
:c
o
E
o
(J
(tt
E.
.a
-tf
(o
0.04
6
C)
E Lrr/D = 20
o
Eo
.9
6
E
l-

zo
0.02

30
2c
1O 40 00
1E+02 1 E+03 1 E+04 1 E+05 1 E+06

Pile-soil stiffness ratio, EplEs

Fig. 6. The effect of pile-soil stiffness ratio on deflection of capped pile group under vertical loads.
 H.H. Zhang, J.C. Small lComputers and Geotechnics 26 (2000) I-21 l5

It may be seen from Fig. 5 that the horizontal displacements of the capped pile
group decrease rapidly with the pile-soil stiffness ratio. The results corresponding to
four pile lengths shows that there is almost no influence of the pile lengths on the
displacement of the capped pile group for a pile-soil stiffness ratio of less 103 and for
pile slenderness ratios in excess of 10. However, when the piles become substantially
rigid (Er/E, up to 106), the shorter piles will obviously produce higher displacement
than the longer piles.
Fig. 6 shows that for vertical loading and a pile-soil stiffness ratio of less than 100,
the displacement of the capped pile group increases sharply as the pile-soil stiffness
ratio reduces. For pile-soil stiffness ratios greater than 100, the displacement of
the capped pile group will decrease very slowly, especially when the piles are short.
Fig. 6 also shows that there are no obvious differences in the deflections of the group
with small pile-soil stiffness ratios (<100) for different pile lengths. However,
for pile-soil stiffness ratios of over 100, pile lengths have more effect on the
deflections of the group and the group with longer piles will produce lower vertical
displacements.

0.15
fl
u*x = displacement at the top of pile No. 5

lr
lrr* = ur*ErD/QrB,L,
ai2
T
J
c
o
E
o
(J
o 0.09
o. HJD = 2,5
,a
p S/D =5.0
o E/E' = 2000
c
o v.= 0.35
N
t-
o V,'=0'15
E 0.06 H/L.' = 100 L"JD = B0
o i

,g
o
E
zo
30
0.03
20
1O +OOO
0.00
1 E+01 1E+02 1 E+03 1 g+04 1 E+05 1 E+06

Cap-soil stiffness ratio, E,/Es

Fig. 7. The effect of cap-soil stiffness ratio on deflection of capped pile groups under horizontal loads'
16 H.H. Zhang, J.C' Small lComputers and Geotechnics 26 (2000) 1-21

3.2.2. The ffict of caP modulus

The pile-ioil stiffness ratio was chosen to be 2000 and the cap-soil stiffness ratio
(Er/Eri was varied from 10r to 106 in this example. Program APPRAF was used to
unofyi. the capped pile group under horizontal loading and vertical loading,
respectively. The displacements of the capped pile group under horizontal loading
and under vertical loading are shown in Figs. 7 and 8, respectively, against the cap-
soil stiffness ratio.
Fig.7 shows that if the capped pile group is subjected to horizontal loading only,
the increase of the cap-soil stiffness ratio will obviously reduce the horizontal dis-
placement of the group. However, once the cap becomes stiffer (a cap-soil stiffness
iatio > 1000), further increase of the capsoil stiffness will have a negligible influence
on the horizontal displacement of the group. Furthermore, the lengths of the piles
have almost no influence on the horizontal displacements of the group. On the
contrary, when the capped pile group is subjected to vertical loads only as shown in
Fig. 8, tiie cap-soil stiffness ratio has only a minor effect on the vertical deflection of the
grJrrp. Howwer, increase in pile length will greatly reduce the vertical displacements of

0.05

0.04

N
N

o
E
(u
o 0.03
o
o-
,9,
E
tr$
()
.F
o
E 0.02
0)
.9,
6
E
zo
0.01

0.00
1 E+01 1E+A2 1E+03 1E+04 1E+05 1 E+06

CaP-soil stiffness ratio, E/E'

Fig. g. The effect of capsoil stiffness ratio on deflection of capped pile groups under vertical
loads.
 H.H. Zhang, J.C. Small lComputers and Geotechnics 26 (2000) I-21 17

the group. For example, as shown in Fig. 8, the vertical displacemsnt of the group
wittr- a pile slenderness ratio of 80 may be reduced to approximately 4348% of that
of a capped pile group with LID : 10.

+ Pile

:T
''L
Case 1: Gase 2: Banerjee'ssoil Gase 3: parabolic soil
Gibson's soil Assuming E'u/E'o = 3 Assuming modulus
at any point given bY

Er(z) = Ero(/Lr')o't

Fig. 9. Models for soils where the modulus increases with depth.

--ts Case 1
E.b = 7 MPa
-*- Case 2
HJD = 2.5
-o- Case 3
L"r/D = 20 +F Homogeneous
lr*" = urrErpD/qrB,Lt

i 0.2
DDEsu

u"" = disp t at the toP of Pil

0.0
0 5 10 15 20

Pile spacing, S/D

Fig. 10. The effect of different soils on the displacement of capped pile groups under horizontal loading.
18 H.H. Zhang, J.C. Small lComputers and Geotechnics 26 (2000) I-21

3.3. Cappect pile groups embedded in soils where the modulus increases with depth

Three types of elastic soils where the modulus increases with depth are chosen in
this paper to demonstrate this method. The soil models are shown in Fig. 9. It is
assumed that the capped pile group with 12 piles as shown in Fig. I is embedded in
the different soil types. The stiffnesses of the different soils can be expressed as

Es(z)
- Eo * m(z/L"^)n
If m:0, the above equation represents a homogeneous soil with a constant elastic
modulus, i.e. Eo : Esbi if Eo:0 and n:I, it represents Gibson's soil; if Eso > 0
and n: 1, it is Banerjee's soil; and the general case is a parabolic variation of soil
modulus.
The soil modulus E,6 was taken as 7 MPa at the toe of the pile and the ratio
L" /D was taken as 20. The pile-soil stiffness ratio Ep/Erv and the cap-soil ratio

0.07
U-= maximum splacement of the +Case 1

--)+ Case 2
+Case 3
0.06 +F Homogeneous
Esb =7 MPa
Ho/D =
2.5
0.05 Ler/D = 20
lr.- = u22E56D/qrBrL,
H/L". = 100

0.03

0.02

0.01

0.00
10
Pile spacing, S/D

Fig. I l. The effect of different soils on the displacement of capped pile groups under vertical loading.
 H .H. Zhang , J.C. Sntall I Computers and Geotechnics 26 ( 2000 ) 1-2 I I 9

Er/Era were 4000 and 4285, respectively. The cap thickness was chosen to be 1.0 m.
The breadth and length of the cap were 20 and 15 m, respectively. Other parameters
were the same as those used in Section 3.2.
Figs. 10 and l1 show that whether the capped pile group is subjected to horizontal
loading or it is subjected to vertical loading, the pile spacing has a significant effect
on displacements of the group when the pile spacing is less than about 6 times the
pile diameter. Furthermore, they show that the homogeneous soil is one of the best
soils with respect to resistance of deflection of a capped pile group. On the contrary,
the Gibson's soil is the poorest one as may be expected. Moreover, beyond a value
of pile spacing of S/D:6, the effect of a change in spacing on the displacement of
the group examined will become much less.
For the capped pile group subjected to horizontal loading, moments in pile 1
corresponding to two pile spacings are plotted in Figs. 12 and 13. With a pile spa-
cing ratio of 2 as shown in Fig. 12, the largest value of the positive moment in pile 1
is much higher than that in pile I with a pile spacing ratio of 6 as plotted in Fig. 13.
This demonstrates that a relatively small pile spacing ratio (i.e. less than 6) may lead
to a large moment in the pile and is therefore a less economical use of the piles'

lM

-0.6 -0.2

Capped pile groups

subjected to horizontal
load

Pile 1

S/D=2
I

Eso = 7 MPa
Ho/D = 2.5
L"r/D = 20
lu = M/q*DBrL,
H/L"m = 100

-+F Homogeneous
+Case 1

-F Case 2
-x- Case 3

Fig. 12. Moment distribution in single pile embedded in different soils with close pile spacing.
20 H.H. Zhang, J.C. Small f Compttters and Geotechnics 26 (2000) I-2 1

lM

-0.6 -4.4

Capped pile group

subjected to horizontal
load

Pile 1

I
I

Esu = 7 MPa
Ho/D = 2.5
Lem/D = 20
lu = M/qrDBrL,
H/L"' = 100

+F Homogeneous
+Case 1

+Case 2
+(- Case 3

Fig. 13. Moment distribution in single pile embedded in different soils with large pile spacing.

4. Conclusions

In this paper, a powerful method for analysing the behaviour of capped pile
groups subjected to both lateral and vertical loading has been developed. The ana-
lysis of a capped pile group with only two piles, comparisons with solutions for
fixed-head pile groups, and investigations of the effect of soil, pile and cap para-
meters demonstrate that:

1. The method can be used to analyse the behaviour of a capped pile group
foundation subjected to vertical loading or horizontal loading as well as
moments in all axis directions.
2. Through comparison of the load distribution on the pile heads between a cap-
ped pile group (where the cap is very stiff) and a pile group with a rigid cap, it
is shown that the present method for analysis of a capped pile group can
reproduce previous results.
3. Pile-soil stiffness ratio plays an important role in the resistance of a capped pile
group to lateral deflection for pile-soil stiffness ratios of less than 1000. In this
 H.H. Zhang, J.C. Small I Computers and Geotechnics 26 (2000) I-21 2l

range, the displacements of the capped pile group will increase sharply with the
reduction of the pile-soil stiffness ratio.
4. For the example chosen, change in the pile-soil stiffness ratio had a large
influence on the displacement of the capped pile group under vertical loading
when this ratio was less than 100. When the pile-soil stiffness ratio is greater
than 100, change in the pile-soil stiffness ratio had only insignificant effect on
the displacement. Pile lengths have a relatively large effect on the deflections of
the capped pile group under vertical loading, but much less effect for hor-
izontal loading.
5. For lateral loading, small pile spacing (i.e. less than 6 times the pile diameter)
will not only result in a large deflection of a capped pile group under applied
load, but it can also cause larger moment in the piles.
6. The method can conveniently solve problems involving capped pile groups
embedded in different soil types. The results of analysis confirm, as expected,
that a Gibson's soil provides the poorest lateral restraint for capped pile
groups subjected to lateral loading.

References

tll Hongladaromp T, Chen N, Lee S. Load distribution in rectangular footings on piles. Geotechnical
Engineering 197 3 ;4(2):7 7 -90.
tzl Kuwabara F. An elastic analysis for piled raft foundations in a homogeneous soil. Soils and Foun-
datio ns 1989 ;29 (l) :82-92.
[3] Clancy P, Randolph MF. An approximate analysis procedure for piled raft foundations. Interna-
tional Journal for Numerical and Analytical Methods in Geomechanics 1993;17:849-69.
[4] Poulos HG. An approximate numerical analysis of pile-raft interaction. International Journal for
Numerical and Analytical Methods in Geomechanics 1994;18:73-92.
[5] Ta LD, Small JC. An approximation for analysis of raft and piled raft foundations. Computers
and
Geotechnics 1997 ;20(2):1 0 513.
Methods in
[6] Ta LD, Small JC. Analysis of piled raft systems in layered soils. Int J for Num and Anal
Geomechs 199 6:20: 57 -7 2.

[7] Bogner FK, Fox RL, Schmit LA. The generation of inter-element-compatible stiffness and mass
matrices by the use of interpolation formulas. In: Proc. Conf. on Matrix Methods in Structural
Mechanics, AFFDL-TR-66-80, 1965' p. 397443'
[8] Small JC, Booker JR. Finite layer analysis of layered elastic materials using a flexibility approach.
Part 2-circular and rectangular loadings. International Journal for Numerical Methods in Engineer-
ing 1986;23:959-78.
[9] Rao SN, Ramakrishna VGST, Rao MB. Influence of rigidity on laterally loaded pile
groups in
marine clay. Journal of Geotechnical and Geoenvirnmental Engineering, ASCE 1998;124(6):542-9.
[10] Zhang HH, Small JC. Analysis of axially and laterally loaded pile groups embedded
in layered soils.
In: Proceedings of 8th ANZ Conference on Geomechanics, Hobart (Tasmania), February, 1999.

[l l] Poulos HG. Behaviour of Laterally-Loaded Piles: II -

Pile Groups. J Soil Mech Found Div
(JSMFD), ASCE l97l ;97(SM5):733-5 l.
[12] El Sharnouby B, Novak M. Static and low-frequency response of pile groups. Can
Geotech J
1985:22:79-94.

View publication stats