You are on page 1of 14

Study on Effect of Ground Improvement

on Lateral Load Carrying Capacity of


Pile using Developed Simplified Soil–Pile
Stiffness Matrix

Tanumaya Mitra, Kalyan Kumar


Chattopadhyay & Ambarish Ghosh

Indian Geotechnical Journal

ISSN 0971-9555

Indian Geotech J
DOI 10.1007/s40098-020-00422-6

1 23
Your article is protected by copyright and
all rights are held exclusively by Indian
Geotechnical Society. This e-offprint is for
personal use only and shall not be self-
archived in electronic repositories. If you wish
to self-archive your article, please use the
accepted manuscript version for posting on
your own website. You may further deposit
the accepted manuscript version in any
repository, provided it is only made publicly
available 12 months after official publication
or later and provided acknowledgement is
given to the original source of publication
and a link is inserted to the published article
on Springer's website. The link must be
accompanied by the following text: "The final
publication is available at link.springer.com”.

1 23
Author's personal copy
Indian Geotech J
https://doi.org/10.1007/s40098-020-00422-6

TECHNICAL NOTE

Study on Effect of Ground Improvement on Lateral Load


Carrying Capacity of Pile using Developed Simplified Soil–Pile
Stiffness Matrix
Tanumaya Mitra1 • Kalyan Kumar Chattopadhyay1 • Ambarish Ghosh1

Received: 4 September 2019 / Accepted: 12 February 2020


 Indian Geotechnical Society 2020

Abstract In this present study, a Beam on Nonlinear Introduction


Winkler Foundation approach has been utilized to study the
effect of ground improvement on deformation character- Subsurface deposits having soft-to-medium-stiff clayey
istics of laterally loaded pile. Fourth-order governing dif- soils are often encountered on project sites. The given site
ferential equation of pile deformation has been solved, and location cannot be avoided due to several commercial
a simplified pile–soil stiffness matrix has been developed. reasons. Normally, the choice of foundation depends on the
The optimum depth (from ground surface) up to which type and intensity of load required to be transmitted and
engineering properties of soil require to be improved has characteristics of the subsurface soil deposits. When sub-
been ascertained using a computer code developed in surface has sufficient bearing capacity, shallow foundation
MATLAB utilizing the developed soil–pile stiffness is the most economical choice for structures subjected to
matrix. This optimum depth can be defined as the depth light-to-moderate loading. However, existence of soft-to-
beyond which the improvement in engineering properties medium-stiff deposits requires deep foundations for trans-
of soil (e.g., deformation modulus) has negligible effect on mitting load at deeper bearing strata. In addition to gravity
deformation characteristics of laterally loaded pile. Opti- load, foundations need to transmit lateral loads arising from
mum depth has been suggested for different slenderness wind, wave or seismic, traffic for bridges, lateral pressure
ratios of piles installed in soft-to-medium-stiff clayey soil on soil retaining structures, etc. Existence of such lateral
deposit, and statistical equations have been developed to load in extreme site conditions even makes the design of
find the optimum depth of ground improvement for dif- pile foundation more complicated (critical) owing to less
ferent pile slenderness ratios (L/D) and relative stiffness lateral resistance offered by the soft soil deposits at shallow
(Ep/Es) of pile. The study reveals that the optimum layer depth. The lateral resistance of pile can be increased by
improvement ratio (L1/L) reduces with the increased pile increasing the length up to a certain extent and diameter of
slenderness ratio (L/D) and increases with the increased pile or by increasing the stiffness of the surrounding soil.
relative stiffness (Ep/Es) of pile. Design charts have been Lateral resistance is greatly influenced by characteristics of
proposed to get a preliminary estimation of lateral pile pile and its surrounding soil up to a certain depth from the
capacity in virgin clay as well as improved clay for dif- surface. Increased stiffness of the surrounding soil may be
ferent layer improvement ratios (L1/L). achieved by using suitable ground improvement methods,
such as jet grouting, deep mixing or replacing the poor soil
Keywords Ground improvement  Pile–soil stiffness  with compacted fill, etc. If the depth from the ground
Pile head stiffness  Finite element method  Lateral load surface up to which soil strength may be improved to
reduce the pile head deformation and increase the lateral
& Tanumaya Mitra capacity of pile is known, then by improving soil up to that
tanumayamitra@gmail.com optimum depth may be used in practice for piles with
1 different slenderness ratios. Many researchers have carried
Department of Civil Engineering, Indian Institute of
Engineering Science and Technology, Shibpur, Howrah, out studies to investigate the behaviour of laterally loaded
India pile. The studies can be categorized in the following areas,

123
Author's personal copy
Indian Geotech J

viz. load transfer curve (p-y curve) methods [1–10], finite layer improvement ratio (L1/L) on the lateral-load-
element methods [11–17], subgrade reaction method carrying capacity of pile embedded in homogeneous
[18, 19], etc. In general, some methods [20, 21] are soil as well as in soil improved from the ground surface
available for piles in homogeneous soil. Davisson and Gill up to a certain depth.
[22] and Basu et al. [23] studied the behaviour of laterally 4. Optimum layer improvement ratios (L1/L) have been
loaded piles in layered soil. The concept of strain wedge estimated for different relative stiffnesses (Ep/Es) and
theory was utilized by Ashour et al. [24] and Ashour and pile slenderness ratios (L/D).
Norris [25], to study the characteristics of piles in layered 5. Design charts have been proposed in this study which
soils under lateral load. A very few theoretical studies have simulate the conditions actually existing in the field
been made in the area of increase in lateral capacity of pile and hence are more relevant for design of piles under
using ground improvement. lateral loads.
The derivation of the differential equation for the
deflection curve of a beam supported on an elastic foun-
dation has been given by Hetenyi [21]. This governing
Theoretical Formulation
differential equation (Eq. 1) has been used by previous
researchers (Poulos and Davis [20]) as the differential
The governing differential equation for deformation of pile
equation of the deflection of a laterally loaded pile, and
[20, 21] at any depth z (Fig. 1) under lateral load is
various solutions have been proposed based on FDM,
expressed as
FEM, etc. In the present study, this standard fourth-order
differential equation of pile deflection [20, 21] has been d4 u
EI þ kuðzÞ ¼ 0 ð1Þ
solved and a finite element-based approach has been uti- dz4
lized to investigate the behaviour of laterally loaded pile where EI is the flexural rigidity of the pile, u is the
embedded in soft-to-medium-stiff homogeneous clay bed deflection of pile at depth z, and k is the lateral stiffness of
as well as in an improved soil deposit. The element stiff- soil.
ness matrix has been arranged and presented in a compact Solution of Eq. (1) for pile deformation expressed as
and concise form, which can be easily used in any pro- z
 z z
gramming language like MATLAB or MATHCAD, and uðzÞ ¼ ekL B1 cos k þ B2 sin k
 L L
the use of expensive software package is not required. z z z
þ ekL B3 cos k þ B4 sin k ð2Þ
This study focuses on the determination of a simplified L L
soil–pile stiffness matrix which can be used to find the
optimum depth of ground improvement required to
increase the lateral-load-carrying capacity of pile. This
optimum depth signifies a depth from ground surface up to
which the stiffness of soil needs to be improved to reduce
the pile head deflection, and an improvement of soil
beyond this optimum depth will yield negligible benefit in
terms of pile head stiffness.
The scope of this paper is outlined as follows:
1. A simplified analysis accounting soil nonlinearity has
been developed to study the soil–pile interaction under
pure lateral loading. In numerical solution of the
problem, standard element stiffness matrix and a
computer code in MATLAB have been developed to
determine the pile deformation.
2. In order to check the validity and accuracy of the
proposed method of analysis, the results of the
experimental investigations and other theoretical solu-
tions have been compared with the results obtained (b)
from the proposed method.
3. An extensive study has been made to investigate the (a)
effect of various important parameters, i.e., slenderness
ratio (L/D) of pile, relative stiffness (Ep/Es) of pile, Fig. 1 a Discretized pile and b nodal force–displacement in typical
element

123
Author's personal copy
Indian Geotech J

Numerical Solution 2 3
1 0 1 0
6 ki ki ki ki 7
6  7
Matrix stiffness approach has been adopted for the 6 hi hi hi hi 7
6 7
numerical solution of the above differential equation 6 eki cos ki eki sin ki eki cos ki e sin ki 7
ki
6 7
(Eq. 1). The embedded pile is assumed to be made of n 4k ki ki ki ki ki 5
i ki
number of elements as shown in Fig. 1. For the ith element, e C1 e C2  eki C2 e C1
hi hi hi hi
Eq. (2) can be rewritten as 2 3
B11 B12 B13 B14
 
z z 6B 7 ð7Þ
uð z Þ ¼ e
ki hz
i B1 cos ki þ B2 sin ki 6 21 B22 B23 B24 7
hi hi 6 7
  4 B31 B32 B33 B34 5
ki hz z z
þ e i B3 cos ki þ B4 sin ki ð3Þ B41 B42 B43 B44
hi hi 2 3
1 0 0 0
where B1, B2, B3 and B4 are the integration constants, hi is 60 1 0 07
6 7
the height or length of the ith element, and pile–soil ¼6 7
40 0 1 05
parameter
sffiffiffiffiffiffiffiffiffiffiffiffi 0 0 0 1
rffiffiffiffiffiffiffiffi
k Let H and M be the end forces acting at the ends i and
ki ¼ hi  ð4Þ
4EI (i ? 1) of the element, then
8 9 8 9
End deformations of ith element (Fig. 1b) at node i (i.e., > Hi > > EIu000 ðzÞ at z ¼ 0 >
>
> >
> >
> >
>
< M = < EIu00 ðzÞ at z ¼ 0 =
at z = 0) and at node i ? 1 (i.e., at z = hi) can be expressed i
¼
> > > 000
as follows: > Hiþ1
> >
> > EIu ðzÞ at z ¼ hi > >
: ; > : >
;
8 9 Miþ1 EIu00 ðzÞ at z ¼ hi
>
> ui ¼ uð 0Þ > > 2
ki ki ki ki
3
< h ¼ u0 ð 0Þ >
> = 
i 6 hi h i hi hi 7
6 7
> k2i 6 0 1 0 1 7
> uiþ1 ¼ uðzÞ >
> >
>
6
¼ EI 2 6 k ki ki ki ki
7
ki ki 7
: ; hi 6 i eki C e C1  e C1  e C2 7
hiþ1 ¼ u0 ðzÞ 4 hi 2
hi hi hi 5
2 3 e sin ki e cos ki eki sin ki
ki ki
eki cos ki
1 0 1 0 8 9 2 3
6 ki ki ki ki 7> B1 > B11 B12 B13 B14
6  7>>
<
>
> 6B 7
6 h i h i h i h i 7 B2 = 6
6 21 B22 B23 B24 7
7
¼6 7
6 eki cos ki eki sin ki eki cos ki eki sin ki 7> B > 4 B31 B32 B33 B34 5 ð8Þ
6 7>> 3> >
4k ki ki 5: B4 ;
B41 B42 B43 B44
i ki ki k i ki ki 2
ki ki ki ki
3
e C1 e C2  e C2 e C1 
hi hi hi hi 6 hi hi hi hi 7
6 7
ð5Þ k2i 6
6 0 1 0 1 7
7
¼ 2EI 2 6 k ki ki ki ki ki ki 7
hi 6 i eki C e C  e C  e C 7
where C1 ¼ ðcos ki  sin ki Þ; C2 ¼ ðcos ki þ sin ki Þ; and u 4 hi 2
hi
1
hi
1
hi
2 5
and h are the horizontal translation and rotation at the ends eki sin ki eki cos ki eki sin ki eki cos ki
2 31
of the element, respectively. 1 0 1 0
6 ki ki ki ki 7
Integration constants (B1, B2, B3 and B4) can be eval- 6
6
 7
7
6 hi hi hi hi 7
uated by applying the following boundary conditions:  6 ki ki ki ki 7
6 e cos k e sin k e cos k e sin ki7
9 4k
i i i

BC  I : ui ¼ 1; hi ¼ 0; uiþ1 ¼ 0; hiþ1 ¼ 0 > i ki ki ki ki ki ki ki 5


>
= hi
e C1
hi
e C2  e C2
hi hi
e C1
BC  II : ui ¼ 0; hi ¼ 1; uiþ1 ¼ 0; hiþ1 ¼ 0
BC  III : ui ¼ 0; hi ¼ 0; uiþ1 ¼ 1; hiþ1 ¼ 0 > > Substituting the integration constants obtained from
;
BC  IV : ui ¼ 0; hi ¼ 0; uiþ1 ¼ 0; hiþ1 ¼ 1 Eq. (7) into Eq. (8), the following nodal force–
ð6Þ displacement relationship of the element has been
obtained.
For all the four boundary conditions, the integration 8 9 8 9
>
> Hi >> > u
constants can be evaluated from the following expression: < = < i > =
Mi hi
¼ ½Ku i ð9Þ
>
> H > : uiþ1 >
>
: iþ1 > ; hiþ1
;
Miþ1

where, ½Ku i is the stiffness matrix of the ith element


expressed in a simplified form as

123
Author's personal copy
Indian Geotech J

2 3
K11 K12 K13 K14 hyperbolic relationship for developing p-y curve proposed
6 K21 K22 K23 K24 7 by Georgiadis et al. [5] is as follows:
½Ku i ¼ 6
4 K31 K32 K33 K34 5
7
y
p¼1 y ð11Þ
K241 K42 K43 K44 3 k þ pu
F1 F6 F2 F4
EI 6 F6 F3 F4 F5 7 where k isqthe ffiffiffiffiffiffiffi modulus of subgrade reaction expressed as
¼ 36 7 ð10Þ 4 E
h 4 F2 F4 F1 F6 5 k ¼ 0:65 EEI sd
12
1l2 (Vesic [26]); Es is Young’s modulus of
s

F4 F5 F6 F3 soil, l is the Poisson’s ratio of soil, and d is the diameter of


pile.
The elements of the stiffness matrix (Eq. 10) have been
derived in this present study and are presented in the
Validation of the Proposed Methodology
following expressions:
 
4k3 e4k þ 2sin2k:e2k  1 Case Study 1
F1 ¼ ;
/
  Lateral load vs deflection result obtained from the present
8k3 ek sink  cosk þ e2k cosk þ e2k sink
F2 ¼  study has been compared with the field test result reported
w
  by Comodromos [27] as illustrated in Fig. 2. Comodromos
2kh 2sin2ke  e4k þ 1
2 2k
[27] has reported the lateral load test response of a single
F3 ¼  ;
/ pile of 1.0 m diameter and 52.0 m length installed at a
 
8k2 h e3k sink  ek sink bridge site. The soil profile of the site consists of a 36.0-m-
F4 ¼
n thick layer of soft silty clay underlain by a medium-stiff
k 2
  clay layer of 12.0 m thickness, followed by a layer of dense
4ke h cosk þ sink  e2k cosk þ e2k sink
F5 ¼ ; sandy gravel. Hazaar et al. [28] simulated the field test
w
  carried out by Comodromos [27] using 3D numerical
2k2 h e4k  2e2k cos2k þ 1 analysis (Fig. 2). In the present study, analysis has been
F6 ¼ 
/ done using the developed stiffness matrix considering the
 4k  
where / ¼ e  4e þ 2cos2ke2k þ 1 , w ¼ e4k  6e2k
2k soil profile and pile properties from the literature and a
þ4cos2 ke2k þ 1Þ, n ¼ 2e2k  e4k þ 4e2k sin2 k  1 . good agreement between the results obtained from the
present study and the reported field and numerical results
Force–Displacement Relationship of Embedded Pile has been found.

The nodal forces {F} and displacement fDg of the Case Study 2
embedded pile are related as fF g ¼ ½K fDg; where the
global stiffness matrix ½K  for the embedded soil–pile Hsiung [19] considered a steel pipe pile of length 20 m
system is assembled from the stiffness matrix ½k of indi- embedded in homogeneous clay with a coefficient of sub-
vidual element. grade reaction of 40,000 kN/m3. The pile is having outer

Soil Nonlinearity 900


800
Different researchers have suggested various methods for 700
Lateral Load (kN)

generation of p-y curves. Matlock [1] and Georgiadis et al. 600


[5] developed p-y relationship for soft clay. Dewaikar and 500
Patil [7] proposed a hyperbolic model to develop p-y 400
curves for clay; Reese et al. [2] have proposed p-y curves 300 Present Analysis
for stiff clay. Reese et al. [8] have developed p-y rela- 200 Comodromos (2003)
tionship for sand. Guidelines to develop p-y curves for both 100 Hazzar et al. (2017)
clayey and sandy soil are available in API [9] and DNV 0
0 50 100
[10]. Nonlinear p-y curves proposed by Georgiadis et al.
[5] have been used in this study to determine the stiffness Lateral Deflection (mm)
of soil. Iterative procedure has been used to modify the Fig. 2 Comparison of the result obtained by present analysis with the
stiffness of soil considering a secant modulus approach. A field test data obtained by Comodromos [27] and numerical simula-
tion by Hazzar et al. [28]

123
Author's personal copy
Indian Geotech J

Table 1 Comparison of theoretical result (Hsiung [19]) with present study


Pile head fixity Load/moment Maximum deflection (mm)
Theoretical (Hsiung [19]) Present Study

Free 400 kN 16.7 16.4


1000 kNm 17.1 18.8

diameter = 610 mm, thickness = 9.53 mm and elastic Results and Discussion
modulus = 2.1  108 kN/m2. The results obtained from the
theoretical solution have been compared with the results The results obtained using the proposed numerical
obtained using the present study (Table 1), and a good methodology on variation of stiffness of pile embedded in
agreement has been achieved. homogenous soft clay deposit and improved thicknesses of
surface soil of the same deposit have been analysed and
Problem Considered discussed.

A single pile with different slenderness ratios and four Variation of Pile Head Stiffness with Normalized
different relative stiffnesses (Ep/Es = 2000, 3000, 4000 and Deflection (y/D) for Different Layer Improvement
5000) embedded in weak subsurface deposit (Fig. 3a) is Ratios (L1/L).
considered to study the influence of improvement of the
surficial deposit L1 (Fig. 3b) on the ground-line deflection Typical variations of pile stiffness with normalized
of pile head. Even though piles are generally constructed in deflection of pile head (y/D), for piles of slenderness ratio
group, in some cases there may be single pile or two piles L/D = 10, embedded in homogenous (L1/L = 0.0) soft clay
under any structural member, where the pile will be acting deposit and improved different thickness of surface soil of
as free-head pile. Further, the soil stiffness was increased the same deposit (i.e., L1/L = 0.1, 0.2, 0.3, 0.4 and 0.5) are
from Es (modulus of virgin deposit) to Es1 (Es1 = 2Es) for a presented in Fig. 4. It is found that, in general, stiffness of
depth of L1 from the ground surface. Generally, the sub- pile decreases with an increase in normalized deflection of
surface deposit consists of very soft-to-medium-stiff weak pile head (y/D), irrespective of pile stiffness factors (Ep/Es)
soil followed by comparatively stiffer layer. Hence, the top and slenderness ratios (L/D) of pile. Variation of pile head
layer has been improved to achieve a strength which is stiffness is predominant at lower values of pile head
easily attainable in site at a reasonable cost. Hence, Es1 deformation (up to y/D & 4% to 5% of pile diameter), and
= 2Es has been considered in this study. By changing the this rate of change of stiffness becomes very low at higher
thickness of soil improvement for different layer values of pile head deflections. For example, for piles with
improvement ratios (L1/L = 0.0, 0.05, 0.10, 0.15, 0.20, slenderness ratio L/D = 10 and layer improvement ratio
etc.), an attempt has been made to find out the optimum L1/L = 0.2, at normalized deflection of pile (y/D) 0.01,
depth of ground improvement required for piles of different 0.05 and 0.1, for relative stiffness Ep/Es = 5000 (Fig. 4a),
slenderness ratios (L/D = 10, 15, 20, 25, 30, 40 and 50). pile stiffnesses are 11,780 kN/m, 6510 kN/m and 4240 kN/
Any further improvement in soil beyond this optimum m, respectively. Hence, for an increase in normalized pile
depth will have negligible effect on pile head stiffness. head deformation (y/D) from 0.01 to 0.05, pile head stiff-
ness reduces 45%, whereas for an increase of 0.01 to 0.10,

Fig. 3 Schematic diagram of


the pile embedded in
homogeneous clay bed with and
without ground improvement

(a) (b)

123
Author's personal copy
Indian Geotech J

Pile Head Stiffness (kN/m)


Fig. 4 Pile head stiffness vs

Pile Head Stiffness (kN/m)


16000 Ep/Es = 5000 L1/L = 0.0 45000
normalized deflection for L/ Ep/Es = 2000 L1/L = 0.0
14000 L/D = 10 L1/L = 0.1 40000
D = 10, for pile stiffness factor L/D = 10 L1/L = 0.1
L1/L = 0.2 35000
values of a 5000 and b 2000 12000 L1/L = 0.2
L1/L = 0.3
10000 30000 L1/L = 0.3
L1/L = 0.4
25000 L1/L = 0.4
8000 L1/L = 0.5
20000 L1/L = 0.5
6000
15000
4000 10000
2000 5000
0 0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
Normalized Deflection Normalized Deflection
(a) (b)

Fig. 5 Pile head stiffness vs 20000 L1/L = 0.0 20000 Ep/Es = 5000 L1/L = 0.0
Ep/Es = 5000
Pile Head Stiffness (kN/m)

Pile Head Stiffness (kN/m)


normalized deflection for pile 18000 L/D = 15 L1/L = 0.05 18000 L/D = 20 L1/L = 0.05
stiffness factor 5000 for a L/ 16000 L1/L = 0.1 16000 L1/L = 0.10
L1/L = 0.15
D = 15, b L/D = 20, c L/ 14000 L1/L = 0.2 14000
L1/L = 0.20
D = 25 and d L/D = 50 12000 L1/L = 0.3 12000 L1/L = 0.25
10000 L1/L = 0.4 10000 L1/L = 0.30
8000 8000
6000 6000
4000 4000
2000 2000
0 0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
Normalized Deflection Normalized Deflection
(a) (b)
20000 L1/L = 0.0 20000 Ep/Es = 5000 L1/L = 0.0
Ep/Es = 5000
Pile Head Stiffness (kN/m)
Pile Head Stiffness (kN/m)

18000 L/D = 25 L1/L = 0.05 18000 L/D = 50 L1/L = 0.05


16000 L1/L = 0.10 16000 L1/L = 0.10
14000 L1/L = 0.15 14000 L1/L = 0.12
12000 L1/L = 0.20 12000
10000 L1/L = 0.25 10000
8000 8000
6000 6000
4000 4000
2000 2000
0 0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
Normalized Deflection Normalized Deflection

(c) (d)

this change is 64%, i.e., for an increase in normalized pile Variation of Pile Head Stiffness with Normalized
head deformation (y/D) between 5 and 10% of pile diam- Deflection (y/D) for Different Slenderness Ratios (L/
eter, pile head stiffness decreased by 19%. Similar type of D).
observation was found for Ep/Es = 2000 (Fig. 4b).
Figure 5 depicts the effect of layer improvement ratios Effect of slenderness ratios (L/D) on the variation of pile
(L1/L) on pile of different lengths (L/D = 15, 20, 25 and head stiffness for piles of different, relative stiffnesses (Ep/
50) for relative stiffness of pile Ep/Es = 5000. Optimum Es) embedded in homogeneous clay deposit is presented in
thickness of improved surficial soil (i.e., L1/L) decreases as Fig. 6. The plots exhibit that for piles in soft clay deposit
the slenderness ratio increases. At normalized deflection of [Ep/Es = 5000, Fig. 6a], pile head stiffness increases up to
pile head, y/D = 0.01, significant improvement in pile head slenderness ratio, L/D = 20, beyond which (L/D  20) the
stiffness was observed up to layer improvement ratio L1/ effect of pile length on lateral capacity of pile becomes
L  0.30, 0.25, 0.20 and 0.10 for pile slenderness ratios L/ insignificant at any value of normalized deflection (y/D),
D = 15, 20, 25 and 50, respectively. Any further increase whereas for pile in medium deposit [Ep/Es = 2000,
in improved surface soil thickness beyond these layer Fig. 6b], lateral pile capacity becomes independent of pile
improvement ratios is found to be insignificant. length at L/D [ 15.

123
Author's personal copy
Indian Geotech J

Fig. 6 Pile head stiffness vs

Pile Head Stiffness (kN/m)


12000 L/D = 10 L/D = 10

Pile Head Stiffness (kN/m)


25000
L1/L = 0.0 L/D = 15 L1/L = 0.0 L/D = 15
normalized deflection for L1/ Ep/Es = 5000 L/D = 20
10000 Ep/Es = 2000 L/D = 20
L = 0.0 for pile stiffness factor L/D = 25 20000
L/D = 25
values of a 5000 and b 2000 8000 L/D = 30 L/D = 30
L/D = 40 15000 L/D = 40
6000 L/D = 50 L/D = 50
10000
4000
5000
2000

0 0
0 0.02 0.04 0.06 0.08 0.1 0 0.05 0.1

Normalized Deflection Normalized Deflection

(a) (b)

Further, the effect of L/D ratio on variation of pile head Variation of Pile Head Stiffness with Normalized
stiffness with normalized deflection (y/D) for piles of different Deflection (y/D) for Different Pile Stiffness Factors
relative stiffnesses (Ep/Es) has also been studied for layer (Ep/Es).
improvement ratio, L1/L = 0.10 and 0.20, as presented in
Figs. 7 and 8, respectively. Plots for different slenderness The effect of Ep/Es ratio on variation of pile head stiffness
ratios (L/D) exhibit (Fig. 7) that the lateral capacity of pile with normalized deflection (y/D) for different layer
increases with the increased slenderness ratio (L/D), irre- improvement ratios (L1/L = 0.0 and 0.10) is portrayed in
spective of Ep/Es. For soft soil (Ep/Es = 5000) and layer Fig. 9 (a) and (b), respectively. The increase in Ep/Es ratio
improvement ratio L1/L = 0.10 (Fig. 7a), the increase in pile reduces lateral capacity of pile, whereas the increase in L1/
head stiffness is 31.6%, 7.5%, 4.9%, 3.7%, 3.3% and 2.2%, for L ratio improves the pile capacity. At any value of nor-
an increase in slenderness ratio (L/D) from 10 to 15, 15 to 20, malized deflection, the pile head stiffness is found to be
20 to 25, 25 to 30, 30 to 40 and 40 to 50, respectively, at increased with the decreased relative stiffness of pile. This
normalized deflection, y/D = 0.01; wherein for relative stiff- magnitude of pile head stiffness decreases sharply up to a
ness Ep/Es = 2000 (Fig. 7b), these changes were 17.4%, value of y/D  0.04–0.05 and becomes negligible at a
2.9%, 2.7%, 2.6%, 2.4% and 0.2%, respectively. Hence, it can higher value of pile head deflection. The rate of decrement
be concluded that with the decrease in pile stiffness factors is found to be higher as the Ep/Es decreases or soil stiffness
(Ep/Es), the percentage changes of pile head stiffness also increases.
reduce for all the slenderness ratios.
Figure 8 exhibits similar variation of pile head stiffness Variation of Pile Head Stiffness with Layer
with normalized deflection (y/D) for layer improvement Improvement Ratios (L1/L) for Different
ratio L1/L = 0.2. In general, pile head stiffness increases Normalized Deflection (y/D).
with the increased stiffness of soil, i.e. the decreased rel-
ative stiffness of pile (Ep/Es). However, at a particular Typical variation of pile head stiffness with layer
deflection of pile head, the percentage increase in pile head improvement ratio (L1/L) for different y/D on piles of
stiffness reduces with slenderness ratios irrespective of different lengths (L/D = 10, 20, 30 and 40) having Ep/Es-
layer improvement ratios (L1/L) and relative stiffness (Ep/ = 5000 is illustrated in Fig. 10. Generally, pile stiffness
Es). Considerable increase in pile head stiffness occurs up increases with the layer improvement ratios (L1/L). For any
to a length of pile, L/D  20, beyond which it is negligible. y/D, at lower values of L1/L, pile head stiffness increases

Fig. 7 Pile head stiffness vs


Pile Head Stiffness (kN/m)

20000 45000 L1/L = 0.10


Pile Head Stiffness (kN/m)

L1/L = 0.10 L/D = 10 L/D = 10


normalized deflection for L1/ 18000 Ep/Es = 5000 L/D = 15 40000 Ep/Es = 2000 L/D = 15
16000 L/D = 20 L/D = 20
L = 0.10 for Ep/Es of a 5000 L/D = 25 35000 L/D = 25
and b 2000 14000 L/D = 30 30000 L/D = 30
12000 L/D = 40 L/D = 40
L/D = 50 25000
10000 L/D = 50
20000
8000
15000
6000
4000 10000
2000 5000
0 0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
Normalized Deflection Normalized Deflection
(a) (b)

123
Author's personal copy
Indian Geotech J

Fig. 8 Pile head stiffness vs 20000 45000

Pile Head Stiffness (kN/m)


L/D = 10

Pile Head Stiffness (kN/m)


L1/L = 0.20 L/D = 10
18000 L1/L = 0.20
normalized deflection for L1/ Ep/Es = 5000
L/D = 15 40000 Ep/Es = 2000 L/D = 15
L = 0.20 for Ep/Es of a 5000 16000 L/D = 20 35000 L/D = 20
and b 2000 14000 L/D = 25 30000 L/D = 25
12000 L/D = 30 L/D = 30
25000
10000 L/D = 40 L/D = 40
20000
8000
15000
6000
4000 10000
2000 5000
0 0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
Normalized Deflection Normalized Deflection
(a) (b)

Fig. 9 Pile head stiffness vs

Pile Head Stiffness (kN/m)


25000 35000
Pile Head Stiffness (kN/m)

L/D = 10 Ep/Es = 5000 L/D = 10 Ep/Es = 5000


normalized deflection for L/ L1/L = 0.0 30000 L1/L = 0.10
20000 Ep/Es = 4000 Ep/Es = 4000
D = 10 and improvement ratio
Ep/Es = 3000 25000 Ep/Es = 3000
of a 0.0 and b 0.10
15000 Ep/Es = 2000 20000 Ep/Es = 2000

10000 15000
10000
5000
5000

0 0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
Normalized Deflection Normalized Deflection
(a) (b)

rapidly followed by negligible rate of increment at higher different codal provisions is generally 1% of pile diameter;
values of L1/L. However, the layer ratio up to which pile hence, the value of normalized deflection is chosen as y/
stiffness increases rapidly depends on the length of pile, i.e. D = 0.01. Whereas to study the failure of soil–pile system,
for L/D = 10 and 40, corresponding L1/L  0.30 and 0.10, a higher value of normalized deflection (i.e., y/D = 0.05)
respectively. has been chosen as at lower normalized deflection, soil
does not generally reach the failure state. It is found that
with the increase in slenderness ratios (L/D), the optimum
Determination of Optimum Layer Improvement value of layer improvement ratio (L1/L) reduces irrespec-
Ratio (L1/L). tive of the relative stiffness of pile (Ep/Es). For example, at
normalized deflection, y/D = 0.01, Ep/Es = 5000, 4000,
Variation of pile head stiffness with layer improvement 3000 and L/D = 30 to 50, the optimum value of L1/L lies
ratios (L1/L) for different lengths of pile of Ep/Es = 5000, between 0.08 and 0.15, whereas for Ep/Es = 2000, it varies
4000, 3000 and 2000 is plotted in Figs. 11 and 12 for between 0.07 and 0.10. Further, for L/D \ 30 and Ep/
deflection of pile head equal to 1% and 5% of pile diam- Es = 5000, 4000, 3000 this value ranges between 0.20 and
eter, respectively. Allowable lateral deflection as per 0.40, whereas between 0.18 and 0.35 for Ep/Es = 2000. For

Fig. 10 Pile head stiffness vs 18000 18000


Pile Head Stiffness (kN/m)

Pile Head Stiffness (kN/m)

layer improvement ratio for a L/ 16000 16000


D = 10 and b L/D = 40 14000 14000
12000 12000
10000 10000
y/D = 0.01 8000 y/D = 0.01
8000
y/D = 0.02 y/D = 0.02
6000 6000
y/D = 0.03 Ep/Es = 5000 y/D = 0.03
4000 Ep/Es = 5000 4000
y/D = 0.04 L/D = 40 y/D = 0.04
2000 L/D = 10 y/D = 0.05 2000
y/D = 0.05
0 0
0 0.2 0.4 0.6 0 0.1
L1/L L1/L
(a) (b)

123
Author's personal copy
Indian Geotech J

Fig. 11 Pile head stiffness vs 18000

Pile Head Stiffness (kN/m)

Pile Head Stiffness (kN/m)


L1/L for y/D = 0.01 and Ep/Es 16000 20000
of a 5000, b 4000, c 3000 and 14000
d 2000 12000 15000
L/D = 10 L/D = 10
10000
L/D = 15 L/D = 15
8000 10000
L/D = 20 L/D = 20
6000 L/D = 30 Ep/Es = 4000 L/D = 30
Ep/Es = 5000
4000 y/D = 0.01 L/D = 40 5000 y/D = 0.01 L/D = 40
2000 Es1 = 2Es Es1 = 2Es
L/D = 50 L/D = 50
0 0
0 0.2 0.4 0.6 0 0.2 0.4 0.6
L1/L L1/L
(a) (b)

25000 35000
Pile Head Stiffness (kN/m)

Pile Head Stiffness (kN/m)


30000
20000
25000 L/D = 10
L/D = 10
15000 L/D = 15 20000 L/D = 15
L/D = 20 15000 L/D = 20
10000
L/D = 30 10000 L/D = 30
Ep/Es = 2000
Ep/Es = 3000 L/D = 40
5000 L/D = 40 y/D = 0.01
y/D = 0.01 5000
L/D = 50 Es1 = 2Es L/D = 50
Es1 = 2Es
0
0 0 0.1 0.2 0.3 0.4 0.5
0 0.2 0.4 0.6
L1/L L1/L
(c) (d)

Pile Head Stiffness (kN/m)


Fig. 12 Pile Head stiffness vs 12000 14000
Pile Head Stiffness (kN/m)

L1/L for y/D = 0.05 and Ep/Es


10000 12000
a 5000, b 4000, c 3000 and
d 2000 10000
8000
8000
6000 L/D = 10 L/D = 10
L/D = 15 6000 L/D = 15
4000 L/D = 20 L/D = 20
4000 Ep/Es = 4000 L/D = 30
Ep/Es = 5000 L/D = 30
2000 y/D = 0.05 L/D = 40 2000 y/D = 0.05 L/D = 40
Es1 = 2Es L/D = 50 Es1 = 2Es L/D = 50
0 0
0 0.2 0.4 0.6 0 0.2 0.4 0.6
L1/L L1/L
(a) (b)
Pile Head Stiffness (kN/m)

16000
Pile Head Stiffness (kN/m)

14000 20000

12000
15000
10000
L/D = 10 L/D = 10
8000 10000 L/D = 15
L/D = 15
6000 L/D = 20 L/D = 20
Ep/Es = 3000 Ep/Es = 2000 L/D = 30
4000 L/D = 30 5000
y/D = 0.05 y/D = 0.05 L/D = 40
2000 L/D = 40 Es1 = 2Es
Es1 = 2Es L/D = 50
L/D = 50 0
0
0 0.2 0.4 0.6 0 0.1 0.2 0.3 0.4 0.5

L1/L L1/L
(c) (d)

y/D = 0.05 and L/D = 30 to 50, the optimum value of L1/ 0.40 for Ep/Es = 5000, 4000 and 3000 and between 0.20
L lies between 0.10 and 0.20 for Ep/Es = 5000, 4000 and and 0.35 for Ep/Es = 2000. Figures 11 and 12 can be used
3000, whereas for Ep/Es = 2000, it varies between 0.09 and as design charts to estimate the lateral pile capacity before
0.15. For L/D \ 30, this value ranges between 0.25 and and after ground improvement. A preliminary prediction of

123
Author's personal copy
Indian Geotech J

increase in lateral pile capacity can also be made by uti- Figs. 11 and 12 are plotted in Fig. 13, which exhibits the
lizing the design charts as described in the following relationship between optimum values of L1/L for different
section. slenderness ratios (L/D) and relative stiffness (Ep/Es) of
pile. A third-order polynomial is found to be suitable for all
Example of Use of Design Chart: the curves shown in Fig. 13a and b. The statistical models
along with the values of coefficient of determination (R2)
Problem 1: A pile of slenderness ratio L/D = 40 is are presented in Table 2. Total eight equations have been
embedded in a soft clay deposit with pile stiffness factor obtained for normalized pile head deflection of 0.01 and
Ep/Es = 5000. Determine the percentage increase in pile 0.05, using which the optimum depth of ground improve-
head stiffness for improving top clay layer with layer ment can be determined for different pile stiffness factors
improvement ratio L1/L = 0.05 and 0.10 at normalized pile (Ep/Es) and slenderness ratios (L/D). The ranges of the
head deflection (y/D) of 0.01.[Es1 = 2Es]. parameters used in developing the statistical models in
nondimensional form are as follows: cohesion C = 17 kPa
Solution From Fig. 11a, it is found that pile head stiffness
to 40 kPa; modulus of elasticity Es = 300 9 C and diam-
in virgin soil with pile stiffness factor Ep/Es = 5000 (soft
eter of pile D = 1.0 m.
clay) is K = 10,565 kN/m. For layer improvement ratio
(L1/L) of 0.05, pile head stiffness increases to K = 15,000
kN/m i.e. 42% increase in pile head stiffness was observed.
Summary and Conclusion
For further improvement, L1/L  0.1, K = 16,000 kN/m;
51% increase in capacity was observed.
The effect of ground improvement on lateral capacity of
Problem 2: A pile of slenderness ratio L/D = 30 is piles using developed simplified soil–pile stiffness matrix,
embedded in a soft clay deposit with pile stiffness factor expressed in terms of stiffness, embedded in soft and
Ep/Es = 5000. Determine the percentage increase in pile medium clay deposit has been investigated and presented.
head stiffness for improving top clay layer with layer Soil modulus value was increased to twice the initial value
improvement ratio L1/L = 0.05 and 0.10 at normalized pile up to different depths from ground surface (in terms of a
head deflection (y/D) of 0.01. If the lateral capacity to be percentage of pile length) to study the effect of improve-
achieved is 1.5 times the initial capacity, what is the ment ratio on pile head stiffness. Design charts (Figs. 11,
required layer improvement ratio (L1/L)? [Es1 = 2Es]. 12 and 13) have been proposed to estimate the lateral
capacity of pile pre- and post-ground improvement as well
Solution From Fig. 11a, it is found that pile head stiffness
as the effective depth of ground improvement. Equa-
in virgin soil with pile stiffness factor Ep/Es = 5000 is
tions have been developed for determining the optimum
K = 10,565 kN/m. For layer improvement ratio (L1/L) of
improvement ratios for pile head deflection equal to 1%
0.05, pile head stiffness increases to K = 13,000 kN/m; that
and 5% of pile diameter (i.e. y/D = 0.01 and 0.05). The
is, 23% increase in pile capacity was observed. For further
findings are summarized.
improvement, L1/L = 0.1, pile head stiffness K increases to
14,000 kN/m; that is, 33% increase in capacity was • The results obtained from the developed soil–pile
observed. Pile stiffness in virgin clay is about 10,565 kN/ stiffness matrix using the code developed in MATLAB
m. To achieve a lateral pile capacity equal to 1.5 times the have been validated against the reported field data,
initial capacity, the required pile stiffness is 15,800 kN/m. numerical results and theoretical study, and a good
Hence, from Fig. 11a the required layer improvement ratio agreement has been obtained.
L1/L  0.20. Similarly, for other pile slenderness ratios (L/ • In general, pile head stiffness reduces with the
D) and pile stiffness factors (Ep/Es), lateral pile capacity for increased normalized deflection (y/D), and this rate of
virgin as well as improved soil (for different layer reduction in stiffness is much more rapid at lower
improvement ratios) can be determined from the charts values of pile head deformation than at the higher
(Figs. 11 and 12). deformation values.
• In homogeneous deposit (L1/L = 0.0) of soft soil (Ep/
Es = 5000), for slenderness ratio L/D  20 and in
Statistical Analysis medium-stiff soil deposit (Ep/Es = 2000) for slender-
ness ratio, L/D  15, no improvement in pile head
An attempt has been made to estimate the values of opti- stiffness was observed.
mum depth of ground improvement through regression • For layer improvement ratio L1/L = 0.10, pile head
analysis based on the aforesaid theoretical findings. Opti- stiffness increases with the increase in pile slenderness
mum layer improvement ratios (L1/L) obtained from ratio for any relative stiffness (Ep/Es) of pile. For layer

123
Author's personal copy
Indian Geotech J

Fig. 13 L1/L vs L/D for 0.45 0.45


y/D = 0.01 y/D = 0.05
different pile stiffness factors 0.4 0.4
for a y/D = 0.01 and b y/ Ep/Es = 5000 0.35 Ep/Es = 5000
0.35

Optimum L1/L
Ep/Es = 4000

Optimum L1/L
D = 0.05 0.3 Ep/Es = 4000 0.3
Ep/Es = 3000 Ep/Es = 3000
0.25 0.25
Ep/Es = 2000 Ep/Es = 2000
0.2 0.2
0.15 0.15
0.1 0.1
0.05
0.05
0
0
0 20 40 60
0 20 40 60
L/D L/D

(a) (b)

Table 2 Statistical model


Ep/Es y/D Equation R2
   2  3
5000 0.01 L1
L ¼ 0:6543  0:0315 DL þ 0:0006 DL 4  106 DL 0.9970
L   2   3
0.05 L1
L ¼ 0:5190  0:012 D þ 2  105 DL þ106 DL 0.9981
L1
L  L 2  
6 L 3
4000 0.01 L ¼ 0:5903  0:0236 D þ 0:0003 D 10 D 0.9983
   2  3
0.05 L1
L ¼ 0:5795  0:0219 DL þ 0:0003 DL 2  106 DL 0.9985
L  L 2  3
3000 0.01 L1
L ¼ 0:7043  0:0413 D þ 0:001 D 8  106 DL 0.9994
L  L 2  3
0.05 L1
L ¼ 0:5795  0:0234 D þ 0:0004 D 3  106 DL 0.9980
L  L 2  3
2000 0.01 L1
L ¼ 0:6562  0:0387 D þ 0:0009 D 7  106 DL 0.9999
L  L 2  3
0.05 L1
L ¼ 0:5633  0:0249 D þ 0:0004 D 3  106 DL 0.9837

improvement ratio L1/L = 0.20, relative stiffness Ep/Es Funding No fund has been received from any funding agency for this
 3000 and slenderness ratio L/D  20, whereas for research work.
Ep/Es = 2000 and pile slenderness ratio L/D  15,
Availability of data and material Available from the corresponding
slight improvement in pile head stiffness was observed. author.
• Increase in pile stiffness factor reduces pile head
stiffness, thereby increasing pile head deflection for Compliance with ethical standards
the same applied lateral load.
Conflicts of interest The authors declare that they have no conflict of
• With the increase in layer improvement ratio (L1/L), the interest.
percentage increase in pile head stiffness gradually
reduces to a negligible value, indicating an optimum Code availability Available from the corresponding author.
layer improvement ratio (L1/L). Pile head stiffness
increases with the increased depth of soil improvement,
indicating higher lateral-load-carrying capacity of pile. References
For every Ep/Es ratio and L/D ratio, there is an optimum
1. Matlock H (1970) Correlations for design of laterally loaded piles
depth of ground improvement, beyond which there is a in soft clay. In: Proceedings of 2nd annual offshore technology
negligible increase in pile head stiffness. This optimum conference, OTC, Houston, Texas, vol 1, pp 577–594
value of layer improvement ratio decreases with 2. Reese LC, Cox WR, Koop FD (1975) Field testing and analysis
increasing slenderness ratio of pile. of laterally loaded piles in stiff clay. In: Proceedings of 7th
annual offshore technology conference, Paper No. OTC, vol
• Statistical models developed can be utilized to find out 2312, pp 671–690
the optimum depth of ground improvement for different 3. O’Neill MW, Gazioglu SM (1984) Evaluation of p–y relation-
slenderness ratios and pile stiffness factors. As prelim- ships in cohesive soils. In: Proceedings of analysis and design of
inary estimate, optimum depth of ground improvement pile foundations. ASCE Geotechnical Engineering Division,
pp 192–213
may be taken as 0.2 L to 0.3 L for piles with slenderness 4. Dunnavant TW, O’Neill MW (1989) Experimental p-y model for
ratio L/D \ 20 and 0.05 L to 0.1 L for L/D [ 20. submerged, stiff clay. J Geotech Eng 115(1):95–114
5. Georgiadis M, Anagnostopoulos C, Saflekou S (1992) Cyclic
lateral loading of piles in soft clay. Geotech Eng 23(1):47–60
6. Wu D, Broms BB, Choa V (1998) Design of laterally loaded piles
in cohesive soils using p-y curves. Soils Found 38(2):17–26

123
Author's personal copy
Indian Geotech J

7. Dewaikar DM, Patil PA (2006) A new hyperbolic p-y curve 18. Hsiung YM, Chen YL (1997) Simplified method for analyzing
model for laterally loaded piles in soft clay. In: Parsons RL (ed) laterally loaded single piles in clays. J Geotech Geoenviron Eng
Foundation analysis and design: innovative methods. ASCE, 123(11):1018–1029
Reston, pp 152–158 19. Hsiung YM (2003) Theoretical elastic-plastic solution for later-
8. Reese LC, Cox WR, Koop FD (1974) Analysis of laterally loaded ally loaded piles. J Geotech Geoenviron Eng 129(5):475–480
piles in sand. In: Proceedings of 6th Annual Offshore Technology 20. Poulos HG, Davis EH (1980) Pile foundation analysis and design.
Conference, OTC, Houston, Texas, pp 473–483 Wiley, New York
9. API RP-2A WSD (2000) Recommended Practice for Planning, 21. Hetenyi M (1946) Beams on elastic foundations. University of
Designing and Constructing Fixed Offshore Platforms—Working Michigan Press, Ann Arbor
Stresses Design, 21st edn. American Petroleum Institute, 22. Davisson MT, Gill HL (1963) Laterally loaded piles in a layered
Washington soil system. J Soil Mech Found Div 89(3):63–94
10. Veritas N (2002) Guidelines for design of wind turbines. Wind 23. Basu D, Salgado R, Prezzi M (2008) Analysis of laterally loaded
Energy Department, Ris National Laboratory, Roskilde, Det piles in multilayered soil deposits. Report FHWA/IN/JTRP-2007/
Norske Veritas 23, Joint Transportation Research Program, Purdue Univ, West
11. Brown DA, Shie CF (1990) Three dimensional finite element Lafayette, IN
model of laterally loaded piles. Comput Geotech 10(1):59–79 24. Ashour M, Norris G, Pilling P (1998) Lateral loading of a pile in
12. Trochanis AM, Bielak J, Christiano P (1991) Three-dimensional layered soil using the strain wedge model. J Geotech Geoenviron
nonlinear study of piles. J Geotech Eng 117(3):429–447 Eng 124(4):303–315
13. Yang Z, Jeremic B (2002) Numerical analysis of pile behaviour 25. Ashour M, Norris G (2000) Modeling lateral soil-pile response
under lateral loads in layered elastic–plastic soils. Int J Numer based on soil-pile interaction. J Geotech Geoenviron Eng
Anal Meth Geomech 26(14):1385–1406 126(5):420–428
14. Fan CC, Long JH (2005) Assessment of existing methods for 26. Vesic AB (1961) Beams on elastic subgrade and the Winkler’s
predicting soil response of laterally loaded piles in sand. Comput hypothesis. In: Proceedings 5th International Conference on
Geotech 32(4):274–289 SMFE 1961, vol 1, pp 845–851.
15. Mukhopadhyay M, Choudhury D, Phanikanth VS, Reddy GR 27. Comodromos EM (2003) Response prediction for horizontally
(2008) Pushover analysis of piles in stratified soil. In: Proceed- loaded pile groups. J Southeast Asian Geotech Soc
ings of 14th World Conference on Earthquake Engineering, vol. 34(2):123–133
14. International Association for Earthquake Engineering (IAEE), 28. Hazzar L, Hussien MN, Karray M (2017) Influence of vertical
Tokyo loads on lateral response of pile foundations in sands and clays.
16. Ahmadi MM, Ahmari S (2009) Finite-element modelling of lat- J Rock Mech Geotech Eng 9(2):291–304
erally loaded piles in clay. Proc Inst Civ Eng Geotech Eng
162(3):151–163 Publisher’s Note Springer Nature remains neutral with regard to
17. Kim Y, Jeong S (2011) Analysis of soil resistance on laterally jurisdictional claims in published maps and institutional affiliations.
loaded piles based on 3D soil–pile interaction. Comput Geotech
38(2):248–257

123

You might also like