HOSTED BY Available online at www.sciencedirect.

com

ScienceDirect
Soils and Foundations 58 (2018) 85–96
www.elsevier.com/locate/sandf

On predicting displacement-dependent earth pressure for

laterally loaded piles
Pengpeng Ni a, Linhui Song b, Guoxiong Mei c,⇑, Yanlin Zhao c,⇑
a
School of Civil and Environmental Engineering, Nanyang Technological University, Singapore 639798, Singapore
b
School of Physical and Mathematical Sciences, Nanjing Tech University, Nanjing 211800, China
c
Key Laboratory of Disaster Prevention and Structural Safety of Ministry of Education, College of Civil Engineering and Architecture,
Guangxi University, Nanning 530004, China

Received 16 March 2017; received in revised form 28 August 2017; accepted 4 October 2017
Available online 6 December 2017

Abstract

This paper presents the derivation of a depth-dependent soil displacement model for laterally loaded piles for use in the calculation of
displacement-dependent earth pressure. A set of fourth-order differential equations are proposed to compute the pile deflection profile
along the pile length. The radial displacement of the soil due to pile movement can be evaluated based on the geometric compatibility
requirements. The soil displacement pattern is then used in the earth pressure model to provide the pattern of earth pressure distributed
around the pile circumference. The experimental data of the pile response, in terms of the p-y curves reported in the literature, are
employed for a comparison with calculations from the proposed approach and other analytical models. The advantages of the developed
calculation framework have been demonstrated, namely, that it can accurately reproduce the experimental measurements of soil reac-
tions acting on a pile at different depths and that the influence of the pile installation can be taken into account. An illustrative example
of cantilever sheet piles is finally provided to show the ability of the proposed method to analyse complicated problems.
Ó 2017 Production and hosting by Elsevier B.V. on behalf of The Japanese Geotechnical Society.

Keywords: Pile; Lateral pressure; Displacement; Soil-pile interaction; p-y curves

1. Introduction during earthquakes, and inclined slope terrain. Fig. 1(a)

Pile foundations have generally been used in a variety of
infrastructures, such as high-rise buildings, hydraulic
structures, highways, railways, bridges and nuclear power
stations. During the design of these structures, a pile can
be loaded axially to transfer loads from the superstructure
to the surrounding soil (Ni et al., 2017a, 2017c, 2017d). In
addition to axial loads, a pile can experience relative pile-
soil movement in the horizontal plane due to lateral loads
induced by wind, ground vibration and lateral spreading
Peer review under responsibility of The Japanese Geotechnical Society.
⇑ Corresponding authors.
E-mail addresses: meiguox@163.com (G. Mei), zhaoyanlin@gxu.edu.
cn (Y. Zhao).
presents the typical behaviour of a laterally loaded pile,
where the pile resists the forces (or moments) at the head.
The soil-pile interaction mechanism (p-y curve, where the
soil reaction is calculated by the integration from soil pres-
sure) is dependent on the stiffness of both the pile and the
soil. On many occasions, the pile design, in terms of geom-
etry (i.e., cross-section shape, dimension and length), pile
type (i.e., pile end fixity) and spatial variation (i.e., spac-
ing), is based on an analysis of the effect of lateral loads.
The analysis of laterally loaded piles can be categorized
into two tasks: one is to discretize the soil medium and the
pile structure using finite elements and to solve the interac-
tion problem based on continuum mechanics; the other is
to measure the p-y curves experimentally and to match

https://doi.org/10.1016/j.sandf.2017.11.007
0038-0806/Ó 2017 Production and hosting by Elsevier B.V. on behalf of The Japanese Geotechnical Society.
86 P. Ni et al. / Soils and Foundations 58 (2018) 85–96

(a) Load (b)

Superstructure
y pa
p
O

z pp

p p0
(c)
Direction of pile motion

pp

pa p0

ya O yp y
Move away from the soil Push into the soil

Fig. 1. Schematics of (a) laterally loaded pile, (b) soil pressure around circumference of pile and (c) displacement-dependent p-y curve proposed by Mei
et al. (2009).

these measurements empirically with analytical solutions.
The continuum-based approach is attractive since it can
explicitly consider the material nonlinearity, the variation
in stiffness of both the pile and the soil and the soil stratig-
raphy. For instance, Bransby and Springman (1999) inves-
tigated the influence of pile spacing and soil nonlinearity on
the load transfer mechanism (p-y function) using plane
strain numerical analyses; Pan et al. (2002) evaluated the
response of a single pile with differing flexural stiffness sub-
jected to lateral soil movements using three-dimensional
finite element models; Yang and Jeremić (2002) conducted
calculations for a single pile under lateral loads in layered
elastic–plastic soils numerically. A finite element tool has
also been developed to analyse the behaviour of sheet pile
walls subjected to surcharge loads (Georgiadis and
Anagnostopoulos, 1998) and of single piles in sloping
grounds (Georgiadis and Georgiadis, 2010, 2012).
Extensive experimental studies have been performed to
facilitate an understanding of the pile response due to lat-
eral loads. Gabr et al. (1994) proposed an interpretation
of dilatometer tests to derive the p-y curves for piles in clay.
Reduced model-scale (Kim et al., 2004) and large-scale lab-
oratory tests (Zhu et al., 2011) were conducted to investi-
gate the performance of a single pile under lateral static
loads, and the testing results were compared with the exist-
ing p-y models. It was seen that pile groups can lower the
load capacity more than individual piles (Rollins et al.,
1998; Rollins and Sparks, 2002). Studies on single piles rep-
resent an upper bound solution to the problem by provid-
ing conservative estimations at the design stage. The
influence of cyclic lateral loading on the pile response has
been explored experimentally in centrifuge (Abdoun and
Dobry, 2002; Abdoun et al., 2003) and in large-scale tests
(Dunnavant and O’Neill, 1989). Brown et al. (1988) indi-
cated that there was a negligible effect of the load cycles
on the group efficiency and that cyclic loads could only
densify the surrounding soil. However, the current p-y
approaches could potentially overestimate the pile head
deflection for monopiles (Hokmabadi et al., 2012), since
the influence of the pile length and the diameter (in the
order of 40 m and 2 m, respectively) was not included.
A suitable length for a pile is often required in order to
provide an adequate ultimate axial load capacity (to resist
differential settlement). Velez et al. (1983) defined an active
length beyond which the pile would behave the same as an
infinitely long pile in the lateral direction (i.e., fully flexi-
ble). In practice, lateral loads are usually not sufficient to
deform the entire length of a pile and the relative pile-soil
displacement below the active length can be disregarded
(Goit et al., 2014). This raises questions as to the current
understanding of p-y curves (Kondner, 1963; Matlock,
1970; Reese et al., 1974; O’Neil and Murchison, 1983;
Gazioglu and O’Neill, 1984; Murchison and O’Neill,
1984; Norris, 1986; Ashour and Norris, 2000; Basu et al.,
2009), where the difference between rigid and flexible piles
is always ignored. Kim et al. (2009) have reported in their
experimental and numerical work that flexible piles could
behave quite differently in terms of bending moment and
lateral displacement compared to rigid piles. The deficiency
of empirical soil reaction models has also been observed for
other similar structures such as pipelines (Saiyar et al.,
2016). Due to lateral loads, a zone of passive soil reaction
 P. Ni et al. / Soils and Foundations 58 (2018) 85–96 87

(i.e., passive earth pressure pp) develops in front of the pile
movement and an active soil wedge (i.e., active earth pres-
sure pa) develops behind the pile (see Fig. 1(b)). In general,
pile deflection varies nonlinearly with depth. Earth pressure
is also a nonlinear function of soil displacement (Mei et al.,
2009; Ni et al., 2017b), as shown in Fig. 1(c). These nonlin-
earities create difficulty when analysing the soil-pile interac-
tion response.
In this paper, an analytical solution is proposed to solve
the soil-pile interaction force–displacement relations (p-y
curves) for laterally loaded piles. First, the magnitude of
pile deflection at different depths is determined using the
fourth-order differential equation, where a Taylor series
approximation is used to simplify the calculation based
on boundary conditions. The pattern of the soil displace-
ment around the circumference of a pile is assumed to
deform only in the radial direction. The cross-section of
the pile preserves a circular shape (i.e., the ovality is
neglected), where the pile movement corresponds to the
motion of a rigid body. Based on the geometric compatibil-
ity requirements, the distribution of the soil displacement is
then solved by pile deflections. It is assumed that pile
deflections along the pile length remain small, so the pile
responds elastically, and the Euler-Bernoulli beam theory
is applicable for describing the correlation between pile
deflection and soil pressure. Once the variation in pile
deflection with depth is obtained, the radial pattern of soil
displacement and pressure can be evaluated using the
displacement-dependent lateral earth pressure model of
Mei et al. (2009). The efficacy of the proposed method is
evaluated by comparing it against experimentally measured
p-y curves (Kim et al., 2004; Zhu et al., 2011) and analyses
using other existing p-y models (Reese et al., 1974; O’Neil
and Murchison, 1983; Wesselink et al., 1988; Kim et al.,
2004; Mei et al., 2009). In the end, an illustrative example
is given to show the application of the developed approach
to an evaluation of the distribution of pile deflection and
soil pressure along the pile length for cantilever sheet piles.
2. Depth-dependent pile deflection and soil pressure
Previous studies (Ashour and Norris, 2000; Basu et al.,
2009) on laterally loaded piles have featured the limitations
of complexity in the calculation where multi-parameters
are often involved. The calibration of these parameters
could hinder their application by practical engineers. In
this paper, an easily implemented approach is proposed
to analyse the distribution of pile deflection subjected to
lateral loads. The displacement-dependent earth pressure
model of Mei et al. (2009) is then used to calculate the earth
pressure for a given relative pile-soil movement.
2.1. Cantilever sheet piles
Based on the Euler-Bernoulli beam theory, a fourth-
order differential equation governs the equilibrium
between a pile’s resistance induced by flexure and the soil
reaction.


d 4y m
EI 4 ¼ p ¼  n  cz ð1Þ
dz 1 þ expðbyÞ
The equation is valid as long as the pile deflection, y,
remains small and the pile responds elastically along the
entire pile length. The flexural rigidity of the pile, EI, can
therefore link the depth-dependent pile deflection with
the associated displacement-dependent earth pressure.
The sign convention employed here is because the soil resis-
tance is in the same direction as the pile deflection for the
portion of the cantilever sheet pile that is above the dredge
line (see Fig. 2).
pffiffiffiffi
Assuming k ¼ 5 EIc , Eq. (1) can be written as


d 4y m
¼  n  k5 z ð2Þ
dz4 1 þ expðbyÞ
Following the Taylor series expansion, the solution
becomes

(a) (b)

Sand: Sand:
γ γ
φ φ
P P

Dredge line Active zone Dredge line Active zone

pa pa
L

Sand: Clay:
γ γsat
D

φ φ=0
Passive zone pp c
pp
Active zone Passive zone Active zone
pa pa

Fig. 2. Pressure distribution for sheet pile wall above water table penetrating (a) sandy soil and (b) clay (Das, 1999).
88 P. Ni et al. / Soils and Foundations 58 (2018) 85–96

X
1
The same calculation procedure using the Taylor series
y ¼ a0 þ a1 z þ a2 z 2 þ    þ ai z i ¼ ai z i ð3Þ expansion can be conducted and the pile deflection is then
i¼0
achieved.
Here, parameter i denotes the order of the Taylor series
expansion. Considering boundary conditions at the ground M 0 2 Q0 3
y ¼ y 0 þ u0 z þ z þ z
surface (z = 0), the coefficients can be solved as a0 = y0, 2EI 6EI
a1 = u0, 2a2 = M0/EI and 6a3 = Q0/EI (i.e., deflection of

k5  z 5 m
y0, rotation of u0, moment of M0 and shear of Q0). Thus,  n
24 1 þ exp ðby 0 Þ
X1
X1
ði þ 4Þ  ði þ 3Þ  ði þ 2Þ  ði þ 1Þ  aiþ4  zi mk5  z5

i¼0 ð j þ 4Þ  ð j þ 3Þ  ð j þ 2Þ  ð j þ 1Þ
" # j¼0
2 3
m
¼  P1   n  k5  z ð4Þ 1 1
1 þ exp b i¼0 ai zi 4  Pj   P 5 ð9Þ
1 þ exp b i¼0 ai zi 1 þ exp b
j1
ai zi
i¼0
When i = 0, Eq. (4) becomes

 It should be noted that the signs of the higher order
m
24  a4 ¼  n  k5  z terms in the Taylor series expansion of Eqs. (7) and (9)
1 þ exp ðba0 Þ

 are different for cantilever sheet piles and pile foundations.
m
¼  n  k5  z ð5Þ
1 þ exp ðby 0 Þ 2.3. Soil pressure
Therefore, the fifth order coefficient, a4, can be calcu-
lated as Both Eqs. (7) and (9) require the calculation of coeffi-

 cients using boundary conditions. For a regular pile or
k5  z m the pile segment of the cantilever sheet wall above the
a4 ¼  n ð6Þ
24 1 þ exp ðby 0 Þ dredge line, parameters y0, u0, M0 and Q0 need to be solved
at the ground surface, while the same calculation should be
Similarly, all higher order coefficients can be evaluated
performed at the dredge line for the cantilever sheet pile
from Eq. (4). Combining all terms in the Taylor series
below the excavation. The determined pile deflection pro-
expansion of Eq. (3), the variation in pile deflection with
file can subsequently be used to compute the soil displace-
depth can be obtained as
ment pattern around the pile circumference. Finally, the
M 0 2 Q0 3 distribution of the depth- and displacement-dependent
y ¼ y 0 þ u0 z þ z þ z
2EI 6EI earth pressure can be evaluated with the displacement-


k5  z 5 m dependent earth pressure model of Mei et al. (2009) for
þ n use in the design.
24 1 þ exp ðby 0 Þ
X1
mk5  z5
þ
j¼0
ðj þ 4Þ  ðj þ 3Þ  ðj þ 2Þ  ðj þ 1Þ 3. Soil displacement induced by lateral pile movement
2 3
1 1 Piles deform laterally due to external loads/moments at
4  Pj   P 5 ð7Þ the head, such as wind or relative pile-soil movement
1 þ exp b i¼0 ai zi 1 þ exp b j1 ai zi i¼0 induced by ground motion (e.g., landslides). The pile
where parameter j represents the higher-order terms of the deflection disturbs the surrounding soil, such that the pile
Taylor series expansion, which is a function of index i. It is pushed into the soil in the direction of the pile motion
should be emphasized that the last term in Eq. (7) repre- and is moved away from the soil in the opposite direction
sents all higher order terms in the Taylor series expansion of the pile motion. Therefore, it causes passive and active
(i.e., greater than the fifth order). zones in front of and behind the pile, respectively.
A simplification is proposed in this paper to calculate
the pattern of ground disturbance around the pile circum-
2.2. Pile foundations
ference, where only the radial soil displacement is
allowed. The pile movement is considered as a rigid body
For the segment of cantilever sheet pile that is below the
motion in the horizontal plane, so the complexity
dredge line (Fig. 2), or a regular pile foundation, the net
involved with the distortion of the pile cross-section is
earth pressure always acts in the opposite direction to that
ignored (i.e., cross-section remains circular). This is not
of the relative pile-soil movement. Eq. (2) can be rewritten
unreasonable since most piles are made of steel and con-
as

 crete and can essentially be regarded as rigid in the hori-
d 4y m zontal plane (i.e., no ovality), even though piles are
¼  n  k5 z ð8Þ
dz4 1 þ expðbyÞ flexible due to longitudinal bending.
 P. Ni et al. / Soils and Foundations 58 (2018) 85–96 89

3.1. Passive zone r  sin a1  r  tan a0 cos a1 ¼ y  tan a0 ð13Þ

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Considering the symmetry of the problem, one quarter In Eq. (11), the term sin a1 is replaced by 1  cos2 a1 ,
and the final azimuth, a1, is calculated by the re-
of the pile is analysed. Fig. 3(a) shows the calculation of
arrangement of Eq. (13), as follows:
the soil displacement pattern around a pile in the passive
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
zone. The cross-section is taken at depth z (Fig. 1(a)), r2 ð1 þ tan2 a0 Þ  y 2  tan2 a0  y  tan2 a0
where a pile experiences a lateral deflection of y (the pile cos a1 ¼ ð14Þ
rð1 þ tan2 a0 Þ
centre moves from the initial position at O to a new loca-
tion at O0 ). When point A is on the right side of point F, Substituting Eq. (13) into Eq. (11), the radial displace-
the azimuth, a0, is less than arctan(O0 H/OO0 ) = arctan(r/ ment is
y), where r indicates the radius of the pile. The radial dis- pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
r2 ð1 þ tan2 a0 Þ  y 2  tan2 a0  y  tan2 a0
placement of the surrounding soil, DS, can then be solved DS ¼
based on the geometric compatibility requirements. cos a0 ð1 þ tan2 a0 Þ
y
AE ¼ OC  OD ¼ r  cos a1 þ y  r  cos a0 ð10Þ þ r ð15Þ
cos a0
AE r  cos a1 þ y When azimuth a0 is greater than arctan(O0 H/OO0 ) =
DS ¼ ¼ r ð11Þ
cos a0 cos a0 arctan(r/y), the solution will correspond to the displace-
ment at point F.
The correlation between a0 and a1 is expressed as
follows: GF ¼ r  r  sin a0 ð16Þ
ðr þ DS Þ sin a0 ¼ r  sin a1 ð12Þ GF r
DS ¼ ¼ r ð17Þ
sin a0 sin a0
The relationship between pile radius r and pile displace-
ment y, at a given initial azimuth, a0, can then be written as Overall, soil displacement DS, induced by pile displace-
follows: ment y at specific depth z in the passive zone, can be
expressed as
8 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
>
> r2 ð1þtan2 a0 Þy 2 tan2 a0 ytan2 a0
>
> þ cosy a0  r;
>
> cos a0 ð1þtan2 a0 Þ
>
> h  i
< 
DS ¼ a 0 2 0 ; arctan r
y ð18Þ
>
> r  r;
>
>
>
>
sin a0
h  i
>
>
: a0 2 arctan r ; 90
y

3.2. Active zone

Similarly, the soil displacement pattern in the active

zone is schematically shown in Fig. 3(b), where a pile
moves a distance of y horizontally from the initial location
at O to reach a final location at O0 . Radial displacement DS
can be calculated as

AE ¼ OC  OD ¼ r  cos a0 þ y  r  cos a1 ð19Þ

AE y  r  cos a1
DS ¼ ¼rþ ð20Þ
cos a0 cos a0
The relationship between a0 and a1 is now written as

ðr  DS Þ sin a0 ¼ r  sin a1 ð21Þ

Substituting Eq. (20) into Eq. (21), the final azimuth, a1,
can then be solved.

r  tan a0 cos a1  r  sin a1 ¼ y  tan a0 ð22Þ

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
r2 ð1 þ tan2 a0 Þ  y 2  tan2 a0 þ y  tan2 a0
Fig. 3. Calculations of soil displacement due to lateral pile movement: (a) cos a1 ¼ ð23Þ
at passive zone and (b) at active zone. rð1 þ tan2 a0 Þ
90
The radial displacement becomes
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
r2 ð1 þ tan2 a0 Þ  y 2  tan2 a0 þ y  tan2 a0
DS ¼ 
cos a0 ð1 þ tan2 a0 Þ
y muth (the displacement approaches 0 at a0 = 90°). Funda-
þ þr
cos a0
3.3. Worked example
The proposed formulas for estimating the radial dis-
placement of the surrounding soil in both the passive and
active zones contain three parameters: the azimuth, a0,
the lateral pile movement at a specific depth, y, and the pile
radius, r. The derivation has limitations as the influence of
the soil parameters (i.e., relative density, friction angle,
cohesion and Poisson’s ratio) is overlooked. For example,
sand with different relative densities could behave differ-
ently, where dense sand will dilate and loose sand will con-
tract due to shearing. The volumetric deformation of soil is
not properly considered in this investigation; therefore, the
calculated displacement could be overestimated or underes-
timated for dense and loose compacted conditions, respec-
tively. Poisson’s ratio also affects the compaction of soil
subjected to shearing. The shear strength parameters of
the friction angle and cohesion could influence the earth
pressure around the pile, which may result in different dis-
placement patterns. The effect of relative pile-soil stiffness is
simply considered in the calculation of fourth-order differ-
ential equations, as mentioned previously, such that the
estimated pile deflection could vary at different depths.
Fig. 4 depicts a set of calculated soil displacement pat-
terns around the pile circumference, which are normalized
against the pile radius. The undeformed pile cross-section is
represented by a solid line, and the soil displacements are
shown in dashed and dashed-dotted lines for the soil in
the passive and active zones, respectively. Three levels of
pile movement are considered, namely, y/r = 0.1, 0.3 and
0.5, to illustrate the influence of the pile deflection on the
1
Active
Normalized soil displacement
At−rest
0.5
0
Pile displacement
−0.5
y/r = 0.1, 0.3, 0.5
−1
−1 −0.5 0 0.5
Normalized soil displacement
Fig. 4. Example of soil displacements induced by lateral pile movement.
P. Ni et al. / Soils and Foundations 58 (2018) 85–96
distribution of soil displacement/pressure. It is anticipated
that the maximum soil displacement occurs at a0 = 0° for
both passive and active conditions, and that the magnitude
of the soil disturbance decreases with an increase in the azi-
ð24Þ
mentally, the displacement fields around the pile depend on
the pile geometry and the soil properties. For practical con-
siderations, significant gaps and/or interpenetration could
occur at the soil-pile interface. Soils (e.g., soft clay or loose
sand) could flow around the laterally moving pile. The
derivation of the radial displacement pattern around the
pile cannot consider all these complex influences; instead,
it is a simple estimation of the soil-pile interaction sub-
jected to moderate lateral motion, where the influence of
the formation of soil-pile gaps is ignored (i.e., the soil is
perfectly bonded to the pile). This is a limitation of the cur-
rent derivation for use in analysing a pile experiencing a
significant amount of ground movement.
4. Displacement-dependent earth pressure model
The calculation of soil reactions due to lateral pile
movement is a displacement-dependent problem. Fig. 1(c)
illustrates the schematics of the variation in earth pressure
with the relative pile-soil displacement. When a pile moves
away from the soil, a soil wedge develops behind the pile.
At a displacement of ya, the soil wedge reaches a limit con-
dition and will fail at the minimum active earth pressure,
pa. On the other hand, when a pile is pushed into the soil,
a passive zone develops in front of the pile, and the maxi-
mum passive earth pressure, pp, is reached for a displace-
ment that is sufficiently large (e.g., yp) (Rollins et al.,
1998; Ashour and Norris, 2000). Beyond the active length,
there is negligible pile deflection due to lateral loads (Velez
et al., 1983; Goit et al., 2014), where the soil reaction is con-
sidered to be under the at-rest condition.
Mei et al. (2009) proposed a displacement-dependent
earth pressure model that can characterize the monotonic
increase and decrease in passive and active earth pressure
with soil displacement. This model has been successfully
applied to analyse retaining walls experiencing different
Passive
modes of rotation (Mei et al., 2009) and with compressible
geofoam inclusions (Ni et al., 2017b). In this paper, the Mei
et al. (2009) model is adopted to derive the p-y curve for
laterally loaded piles.


k k  4 k 0 cz
p¼   ð25Þ
1 þ expðbyÞ 2 2
where parameter k is a function of Rankine’s passive (kp)
and Jaky’s at-rest earth pressure coefficient (k0) in the form
of k = 4kp/k0  4. Parameter b considers the influence of
Rankine’s passive (kp), active (ka), Jaky’s at-rest earth pres-
sure coefficient (k0) and the displacement to mobilize the
1 1.5
full active pressure (ya) in the form of b = ln((kp  ka)/
(kp  2k0 + ka))/ya < 0. The unit weight of the soil is
denoted as c and the depth is expressed as z.
 P. Ni et al. / Soils and Foundations 58 (2018) 85–96
The displacement-dependent earth pressure model is re-
arranged as follows:


m
p¼  n  cz ð26Þ
1 þ expðbyÞ
using two parameters, namely, m = 2(kp  k0) and n = kp
 2k0, to facilitate the calculation of pile deflection with
depth considering the effects of the pile’s flexural stiffness.
The shear forces mobilized between the pile and the soil
should conventionally be combined with normal forces in
the calculation, although the contribution of shear forces
is less significant. In this investigation, only the contribu-
tion of normal forces is incorporated.
The characteristics of the displacement-dependent earth
pressure are in contrary to the assumption that the ultimate
soil reaction varies with the pile diameter (Broms, 1964a,
1964b). However, the displacement to mobilize the full soil
resistance is generally regarded as a function of the pile
diameter at a specific depth. For example, Ng et al.
(2004) indicated that the active soil wedge could be fully
developed at approximately 1% of the pile diameter for
dense sand. This is consistent with experimental (Fang
et al., 2002; Rollins and Sparks, 2002) and numerical
(Fan and Long, 2005) observations that for medium dense
to dense sand, a normalized displacement ya/d (i.e., d is the
pile diameter) of between 0.5% and 3% is required to mobi-
lize the active condition. For loose sand, the ya/d value
could be as large as 11 –15% (Fang et al., 2002; Rollins
and Sparks, 2002; Cubrinovski et al., 2006). Hansen
(1961) provided a range in ya/d values depending on the
properties of the surrounding soil, such as 3 –5% for stiff
to soft clay, 2–3% for dense sand, 3–5% for medium sand
and 7–10% for loose sand.
5. Comparison of calculated p-y curves
For evaluation, a series of large-scale laboratory tests
performed on steel pipe piles in low liquid limit silt (Zhu
et al., 2011) is considered in the following. The experimen-
tally measured p-y curves are compared to calculations
using the proposed approach as well as other analytical
methods. Similarly, the experimental data of Kim et al.
(2004) on pre-installed and driven piles in sand are used
to assess the ability of the developed analytical framework
to interpret the influence of the installation methods on the
pile response in terms of the p-y curves.
5.1. Steel pipe piles in silt
A large-scale soil tank, 15 m long by 5 m wide by 6 m
deep, was assembled and a series of laboratory tests was
conducted on steel pipe piles experiencing lateral static
and impact loading (Zhu et al., 2011). The lateral deflection
of the soil tank was restrained by steel plates with a thick-
ness of 0.005 m. Teflon lubricant was applied around the
sidewalls to minimize the boundary effect, where an infinite
width of soil volume was modelled. The tested pile featured
91
a length of 4.5 m, a diameter of d = 0.159 m and a bending
stiffness of EI = 1.417 MNm2. A low liquid limit silt was
used to backfill the trench, which had an effective internal
friction angle of 30°. The soil was fully saturated during
the tests, so the measured density of 1.93 g/mm3 indicated
an effective unit weight of 9.12 kN/m3. The tests were con-
ducted in the 1 g environment to examine the soil-pile
dynamic interaction models, and the scaling problems were
not directly considered. However, the flexural rigidity was
kept constant between the model pile and the correspond-
ing prototype-scale pile. The pile was pushed into the
ground using a hydraulic jack mounted on the reaction
frame. This pile installation method could densify the sur-
rounding soil, but there was no measurement of the soil
disturbance. Therefore, a lower bound displacement is esti-
mated as ya = 2.5%d to mobilize the full active soil pres-
sure in this investigation for the backfill material of silt
(Hansen, 1961).
The p-y curves, measured in the static load test at differ-
ent depths from the ground surface (i.e., z = 1d, 2d, 3d, 4d
and 5d) (Zhu et al., 2011), are plotted in Fig. 5. The results
analysed with the Reese et al. (1974) method are included.
Kim et al. (2004) developed a set of simple hyperbolic p-y
functions based on the Broms (1964a) ultimate soil-pile
reaction. The Kim et al. (2004) model requires the input
of only two parameters. Zhu et al. (2011) employed a con-
stant of horizontal subgrade reaction nh of 4828 kN/m3 and
a coefficient f of 14.2 to calculate the p-y curves using the
Kim et al. (2004) model, and the results are also given in
the figure. The proposed method provides evaluations of
the displacement-dependent p-y curves, assuming that the
driving mechanism of piles alters the ya value, but not
the friction angle.
It is clear that the best interpretation of the testing data
is provided by the new developed approach, as illustrated
by the thicker solid lines, while the Reese et al. (1974)
method constantly assesses much lower soil resistance
and the Kim et al. (2004) model always overestimates the
pile response. The measured p-y curves do not show any
15
Kim et al. 2004 Proposed method
Reese et al. 1974 Zhu et al. 2011
Soil reaction, p (kN/m)
10 z = 5d
5
Δz = 1d
z = 1d
0
0 2 4 6 8 10 12
Lateral displacement, y (mm)
Fig. 5. Proposed method compared with measured (Zhu et al., 2011) and
predicted (Reese et al., 1974; Kim et al., 2004) methods.
92 P. Ni et al. / Soils and Foundations 58 (2018) 85–96
sign of softening behaviour, but all the analytical calcula-
tions provide different degrees of softening. In particular,
the peak soil resistances have been mobilized at all depths
in the Reese et al. (1974) method. The conservatism, com-
pared to the experimental measurements, has been reduced
significantly from the evaluations using the Kim et al.
(2004) model to the proposed approach. It is interesting
that the performance in the analytical solutions gets better
at greater depths. This is probably induced by the loading
scheme which uses a hydraulic jack at the pile head, where
the friction force mobilized between the loading rig and the
soil is not properly taken into account in the tests. The ini-
tial earth pressure acting on the pile (at small lateral dis-
placement), measured from the experimental work of Zhu
et al. (2011), was always slightly higher than that in the cal-
culations from different analytical solutions. This is
believed to be partially induced by the set-up of earth pres-
sure cells, where the soil particles of greater size could be in
direct contact with the sensors which would introduce
higher pressures (i.e., nonuniform earth pressures) (Zhu
et al., 2011). As the pile deflected, the soil particles were
re-arranged to have uniform contact with the sensors.
5.2. Pre-installed and driven piles in sand
A series of model-scale laboratory tests was conducted
on piles embedded in Nak-Dong River sand in a strong
box of 0.76 m by 0.76 m in plane and 0.5 m deep (Kim
et al., 2004). The soil container was designed to have
enough clearance between the sidewalls and the pile (more
than 6 times the pile diameter), so that the boundary effect
would be negligible. Two types of piles were tested: pre-
installed piles using the pre-drilling and grouting technique
and driven piles. All these model piles had an equivalent
prototype diameter of d = 0.408 m and a bending stiffness
of EI = 45 GNm2 after scaling 34 times (Kim et al.,
2004). It should be emphasized that the adopted scaling
factor was a geometric scaling, and that the soil modulus
was not scaled. Nak-Dong River sand is classified as poorly
graded sand. The influence of the relative density of the
sand was evaluated by counterpart tests, where medium
dense and dense conditions were achieved by pouring the
sand from a height of 200 mm above the ground surface
at different dropping rates (air pluviation technique). The
measured dry densities were 13.13 kN/m3 and 13.72 kN/
m3 for the medium sand (relative density of Dr = 50%)
and the dense sand (Dr = 73%), respectively. The pre-
installed pile introduced minimal disturbance to the sur-
rounding soil by pre-drilling a hole for the pile, while the
driven pile caused more compaction in the sand. It should
also be noted that drainage was allowed, and that less
excess pore pressure was generated during the whole instal-
lation process of the pre-installed piles. For the driven
piles, the installation was completed in a short time inter-
val, and more excess pore pressure was generated (Kim
et al., 2004). Therefore, the soil is considered to be under
drained and undrained conditions for pre-installed and dri-
ven piles, respectively. Park and Jeong (2015) performed a
series of triaxial tests on Nak-Dong River sand of various
densities, and reported the drained and undrained friction
angles to be 25.3° and 11° for the sand at medium dense
and dense conditions, respectively, in the model tests of
Kim et al. (2004). A single value of displacement at the full
active condition is assumed as ya = 1.5%d for medium
dense sand (Fan and Long, 2005) for ease of the
calculation.
A comparison between the measured p-y curves (mark-
ers) of Kim et al. (2004) and the calculated responses at a
depth of three times the pile diameter (i.e., z = 3d) is illus-
trated in Fig. 6. Five load transfer models are employed:
the Reese et al. (1974) p-y curve for silica sand, the Amer-
ican Petroleum Institute (API) curve by O’Neil and
Murchison (1983) for silica sand, the Wesselink et al.
(1988) p-y curve for calcareous sand, the Kim et al.
(2004) hyperbolic p-y function and the proposed p-y rela-
tionship in this paper.
The analytical models for silica sand (Reese et al., 1974;
O’Neil and Murchison, 1983) provide a very stiff elastic
response, but a peak resistance is mobilized at small pile
deflection. The pile response is different when it is embed-
ded in Nak-Dong River sand, which has a sub-angular
shape, although it contains very large portions of silica
sand (Park and Jeong, 2015). The soil particles are believed
to be at a much tighter packing state due to the sub-angular
shape, which may further increase the shear strength of the
soil. This explains why the experimental data show a much
softer response and none of the piles (despite the different
installation methods) could reach the peak soil-pile interac-
tion force. The Wesselink et al. (1988) model was devel-
oped to estimate the pile behaviour in calcareous sand,
which is more representative of the sandy material used
in the model tests of Kim et al. (2004). However, the influ-
ence of the installation approaches was not taken into
account (Wesselink et al., 1988). It is not surprising that
their method provides a reasonable evaluation of the p-y
0.8
Measured Driven
(Kim et al. 2004)
Proposed
Soil reaction, p (kN/m)
Reese et al. 1974
0.6
Proposed method
0.4
API
Pre−installed
0.2 Wesselink et al. 1988
0
0 0.05 0.1 0.15 0.2 0.25
Normalized displacement by pile diameter, y/d
Fig. 6. Proposed method compared with measured (Kim et al., 2004) and
predicted (American Petroleum Institute (API) model, Reese et al., 1974;
Wesselink et al., 1988; Kim et al., 2004) methods.
 P. Ni et al. / Soils and Foundations 58 (2018) 85–96
curve for the pre-installed pile. The Kim et al. (2004) hyper-
bolic p-y function and the developed p-y response all give
comparable results to the laboratory testing data. The cal-
culation of Kim et al. (2004) is actually more like a curve-
fitting process, whereas the advantage of the new method in
this investigation is that it is more like a prediction of the
testing data. The effect of the installation methods can be
handled well by assuming different friction angles, relative
densities, unit weights or mobilized displacement at the full
active condition. The pile driving process can densify the
surrounding soil, such that the soil parameters for the
dense compacted condition should be used in the proposed
approach. Pre-installed piles using the pre-drilling and
grouting technique could introduce minimal disturbance
to the surrounding soil, and the soil parameters for the
medium sand should be employed in the calculation.
6. Illustrative example
In the following, an illustrative example will be given to
show the calculation process for cantilever sheet piles,
where three stages of excavation are considered. The pro-
posed analysis method can provide an evaluation of pile
deflection and the soil pressure profile along the pile length.
The novelty of this approach is that the earth pressure pat-
tern around the pile circumference at any specific depth can
be assessed. Cantilever sheet piles are analysed rather than
regular pile foundations because regular pile foundations
require that calculations be conducted for the pile sections
both above and below the dredge line. The analysis can
exhibit the efficacy of the proposed method in evaluating
complicated problems.
As shown in Fig. 7, the total excavation depth is 7.0 m,
and the foundation pit needs to be excavated in three stages
±0.0 m
Excavation
−3.0 m Dredge line Soil:
Unit weight:
−5.0 m Dredge line γ = 18.0 kN /m³
Friction angle:
φ = 25°
L = 15.0 m
−7.0 m Dredge line
Cohesion:
c = 15 kPa
D = 8.0 m
Cantilever sheet pile:
Spacing: b = 2.4 m
Flexural rigidity:
EpI p = 0.6 GN ·m²
d = 0.8 m
Fig. 7. Use of cantilever sheet piles to retain soil during excavation.
93
(e.g., 3.0 m, 5.0 m and 7.0 m). It is decided that cantilever
sheet piles be used to support the earth pressure during
the excavation. The soil has a unit weight of 18.0 kN/m3,
an internal friction angle of 25° and cohesion of 15 kPa.
The water table is assumed to be far below the pile tip,
so the investigated case corresponds to a sheet pile wall
in absence of a water table.
All these are enough to estimate the proper depth of the
embedment of sheet piles. The analyses of two cases are
introduced by Das (1999), where the sheet pile can pene-
trate a sandy soil or clay, as presented in Fig. 2
(a) and (b), respectively. The pile head deflects towards
the left (the pile moves away from the retained soil), so
an active zone is presented above the dredge line. Below
the dredge line, the tendency of the pile movement is
reduced towards the left. Therefore, a passive zone is
formed on the left side. Considering the force and the
moment equilibrium, an active zone is generated near the
pile tip. Based on the Das (1999) procedure, the theoretical
depth of the embedment, D, is evaluated as 9.4 m and 6.9 m
for the sheet pile wall penetrating a uniform soil layer of
sand and of clay, respectively. The D value for the analysed
case should lie within the suggested range. In this paper, an
embedment of D = 8.0 m is assumed for the following cal-
culation. The pile diameter is d = 0.8 m and the spacing is
b = 2.4 m. Flexural rigidity EI is then calculated as 0.6
GNm2. The displacement at the full active pressure of
ya = 3.75%d is used to characterize the surrounding soil
at medium density (Hansen, 1961).
The pile deflection profile is depicted in Fig. 8, where the
influence of the excavation can be seen. It is initially close
to a linear pattern and approaches a nonlinear form as the
construction proceeds. At the first stage of construction,
the excavation reaches a depth of 3 m, and the pile head
deflection is slightly above 10 mm. In the following, more
pile deflections can be observed. For example, the pile head
moves approximately 33 mm and 60 mm at the second (5
m) and the third excavation stages (7 m), respectively.
There is a linear relationship between the pile head move-
ment, yhead, and the excavation depth, zdredge.
y head ¼ 12:2  zdredge  26:3 ð27Þ
This equation can be used to indicate the maximum pile
deflection at different levels of excavation for similar
projects.
Based on the pile deflection profile, the soil reactions can
be estimated using the displacement-dependent earth pres-
sure model of Mei et al. (2009). Fig. 9(a) presents the cal-
culated results, along with the evaluations obtained from
Rankine’s theory for active and passive pressure and Jaky’s
theory for the at-rest condition. The Mei et al. (2009)
method provides calculations within the range evaluated
by Rankine’s and Jaky’s theories. In the active zone, the
computed earth pressure increases with depth and the
increasing rate is also raised. This matches the expectation
well, since the decreased pile deflection with depth reduces
the possibility to mobilize the full active pressure and the
94 P. Ni et al. / Soils and Foundations 58 (2018) 85–96

0 earth pressure approaches a greater value proximal to the

at-rest condition. In the passive zone, the computed earth
1 pressure increases with depth just below the dredge line,
2 which is subsequently reduced due to a decreased amount
of pile deflection near the pile tip. The net pressure on
3
Dredge line the pile is then calculated as given in Fig. 9(b), and it is
4 quite similar to the schematic pressure distribution illus-
trated in Fig. 2. The studied case is close to that investi-
5
Dredge line gated by Das (1999) for a sheet pile wall above the water
6 table penetrating a sandy soil, and the effectiveness of the
Depth, z (m)

current method is demonstrated.

7
Dredge line The proposed approach is advantageous in that it can be
8 used to evaluate the full profile of the pile deflection and
the soil pressure in both vertical and circumferential direc-
9
tions of the pile. Fig. 10 illustrates the changes in the earth
10 pressure pattern due to the lateral pile movement com-
pared to the at-rest condition for the pile section at a depth
11
of 7.5 m from the pile head (i.e., 0.5 m below the dredge
12 line at the final excavation) experiencing an approximate
Excavation at 3 m pile deflection of 20 mm. The at-rest earth pressure, p0, is
13
Excavation at 5 m 56.3 kPa, and it varies between the minimum active earth
14 Excavation at 7 m pressure of pmin = 32.3 kPa and the maximum passive
earth pressure of pmax = 80.4 kPa for the laterally loaded
15
60 55 50 45 40 35 30 25 20 15 10 5 0 pile. The redistributed pressure pattern is similar to the
Displacement, y (mm) derivation of Matlock (1970), where a passive zone with
a higher earth pressure develops in front of the pile
Fig. 8. Pile deflection profile during excavation.
movement.

(a) 0 (b) 0
1 Passive by Rakine Active by Rakine 1
Proposed method Proposed method
2 2
At−rest by Jaky At−rest by Jaky
3 3
4 4
5 5
6 6 Active zone
Depth, z (m)

Depth, z (m)

Passive zone Active zone Dredge line

7 7
Dredge line
8 8
Passive zone
9 9
10 10
11 11
12 12
13 13
14 14
Active zone
15 15
−400 −300 −200 −100 0 100 200 300 400 −100 −75 −50 −25 0 25 50 75 100
Pressure, p (kPa) Pressure, p (kPa)
Fig. 9. Soil pressure distribution: (a) at both sides of sheet pile wall and (b) net pressure.
 P. Ni et al. / Soils and Foundations 58 (2018) 85–96
(a) (b)
pmax = 80.4 kPa
O O
p0 = 56.3 kPa
Fig. 10. Soil pressure distribution around circumference of pile: (a) at-rest condition and (b) considering pile movement.
7. Conclusion
A set of approximate design equations was developed in
this paper to evaluate the performance of laterally loaded
piles. A set of fourth-order differential equations was pro-
posed to derive the lateral profile of the pile deflection.
The analyses used the Taylor series to evaluate the param-
eters at the boundaries. All these calculations were valid as
long as the pile responses remained in the elastic range, so
it was applicable to employ the Euler-Bernoulli beam the-
ory to describe the relationship between the pile deflection
and the soil pressure. A radial displacement mechanism
was assumed for the soil subjected to pile movement. The
ovality of the pile cross-section was ignored, which enabled
the pile to be analysed as a rigid body motion in a horizon-
tal plane. For a given pile deflection, the soil displacement
pattern was calculated based on the geometric compatibil-
ity requirements. The estimated pile deflection could then
be used to calculate the pattern of soil pressure acting on
the pile using the displacement-dependent lateral earth
pressure model of Mei et al. (2009).
Experimental measurements of the pile response, in
terms of the p-y curves, were used to assess the efficacy of
the proposed method. Compared to the measured p-y
curves of Zhu et al. (2011), the calculations in the current
study presented a better performance than those estimated
by either the analytical solution of Reese et al. (1974) or the
curve fitting model of Kim et al. (2004). The accuracy of
the calculation model was improved for the pile sections
at a greater depth, where the measurements were less influ-
enced by the loading device of the test. The effect of the pile
installation method has not been considered in most of the
existing analytical models (Reese et al., 1974; O’Neil and
Murchison, 1983; Wesselink et al., 1988), while the devel-
oped approach can take it into consideration by using dif-
ferent parameters, such as the friction angle and the
displacement corresponding to the full active state. A com-
95
pmin = 32.2 kPa
parison between the measured data and the curving fitting
model of Kim et al. (2004) and the current design frame-
work demonstrated its ability to provide reasonable evalu-
ations for piles installed using different techniques.
An illustrative example was included to demonstrate
how the proposed approach could be applied to analyse
cantilever sheet piles. Three stages of excavation were con-
sidered, where the induced pile deflection and the soil pres-
sure were calculated. The computed earth pressure was
generally within the range estimated by Rankine’s active
and passive theory and by Jaky’s at-rest theory. The nov-
elty of this developed design framework was illustrated
by an example of the pile section where the earth pressure
pattern was evaluated.
It should be emphasized that the proposed technique
has the following limitations: (a) the influence of the
shear-strain-strength characteristics (friction angle and
cohesion) and the deformation parameters (relative density
and Poisson’s ratio) of the soil is overlooked; (b) the forma-
tion of soil-pile gaps and the interpenetration of the pile
into the soil are not considered; (c) the impact of dynamic
loading during the pile installation is not taken into
account. Further investigations are needed to develop a
more rigorous analytical solution for the evaluation of lat-
erally loaded piles.
Acknowledgements
This work was supported by the National Science Fund
for Excellent Young Scholars (Grant No. 51322807), the
National Natural Science Foundation of China (Grant
No. 51578164), the Natural Science Foundation of
Guangxi Province (Grant No. 2016GXNSFGA380008),
and the Ministry of Education of China through the
Changjiang Scholars Program to Dr. Guoxiong Mei, and
the National Natural Science Foundation of China (Grant
No. 41672296) to Dr. Yanlin Zhao.
96 P. Ni et al. / Soils and Foundations 58 (2018) 85–96
References
Abdoun, T., Dobry, R., 2002. Evaluation of pile foundation response to
lateral spreading. Soil Dyn. Earthquake Eng. 22 (9), 1051–1058.
Abdoun, T., Dobry, R., O’Rourke, T.D., Goh, S., 2003. Pile response to
lateral spreads: centrifuge modeling. J. Geotech. Geoenviron. Eng. 129
(10), 869–878.
Ashour, M., Norris, G., 2000. Modeling lateral soil-pile response based on
soil-pile interaction. J. Geotech. Geoenviron. Eng. 126 (5), 420–428.
Basu, D., Salgado, R., Prezzi, M., 2009. A continuum-based model for
analysis of laterally loaded piles in layered soils. Geotechnique 59 (2),
127–140.
Bransby, M.F., Springman, S., 1999. Selection of load–transfer functions
for passive lateral loading of pile groups. Comput. Geotech. 24 (3),
155–184.
Broms, B.B., 1964a. Lateral resistance of piles in cohesionless soils. J. Soil
Mech. Found. Divis. 90 (3), 123–158.
Broms, B.B., 1964b. Lateral resistance of piles in cohesive soils. J. Soil
Mech. Found. Divis. 90 (2), 27–64.
Brown, D.A., Morrison, C., Reese, L.C., 1988. Lateral load behavior of
pile group in sand. J. Geotech. Eng. 114 (11), 1261–1276.
Cubrinovski, M., Kokusho, T., Ishihara, K., 2006. Interpretation from
large-scale shake table tests on piles undergoing lateral spreading in
liquefied soils. Soil Dyn. Earthquake Eng. 26 (2), 275–286.
Das, B.M., 1999. Principles of Foundation Engineering. PWS PUBLISH-
ING, Brooks/Cole Publishing Company, Pacific Grove, California.
Dunnavant, T.W., O’Neill, M.W., 1989. Experimental p-y Model for
Submerged, Stiff Clay. J. Geotech. Eng. 115 (1), 95–114.
Fan, C.-C., Long, J.H., 2005. Assessment of existing methods for
predicting soil response of laterally loaded piles in sand. Comput.
Geotech. 32 (4), 274–289.
Fang, Y.-S., Ho, Y.-C., Chen, T.-J., 2002. Passive earth pressure with
critical state concept. J. Geotech. Geoenviron. Eng. 128 (8), 651–659.
Gabr, M.A., Lunne, T., Powell, J., 1994. p-y analysis of laterally loaded
piles in clay using DMT. J. Geotech. Eng. 120 (5), 816–837.
Gazioglu, S.M., O’Neill, M.W., 1984. Evaluation of py relationships in
cohesive soils. In: Analysis and design of pile foundations. ASCE. pp.
192–213.
Georgiadis, K., Georgiadis, M., 2010. Undrained lateral pile response in
sloping ground. J. Geotech. Geoenviron. Eng. 136 (11), 1489–1500.
Georgiadis, K., Georgiadis, M., 2012. Development of p–y curves for
undrained response of piles near slopes. Comput. Geotech. 40, 53–61.
Georgiadis, M., Anagnostopoulos, C., 1998. Lateral pressure on sheet pile
walls due to strip load. J. Geotech. Geoenviron. Eng. 124 (1), 95–98.
Goit, C., Saitoh, M., Oikawa, H., Kawakami, H., 2014. Effects of soil
nonlinearity on the active length of piles embedded in cohesionless soil:
model studies. Acta Geotech. 9 (3), 455–467.
Hansen, J.B., 1961. The ultimate resistance of rigid piles against
transversal forces. Danish Geotechnical Institute, Bulletin, 12, Copen-
hagen, Denmark.
Hokmabadi, A., Fakher, A., Fatahi, B., 2012. Full scale lateral behaviour
of monopiles in granular marine soils. Mar. Struct. 29 (1), 198–210.
Kim, B.-T., Kim, N.-K., Lee, W.-J., Kim, Y.-S., 2004. Experimental load-
transfer curves of laterally loaded piles in Nak-Dong River sand. J.
Geotech. Geoenviron. Eng. 130 (4), 416–425.
Kim, Y., Jeong, S., Won, J., 2009. Effect of lateral rigidity of offshore piles
using proposed py curves in marine clay. Mar. Georesour. Geotechnol.
27 (1), 53–77.
Kondner, R.L., 1963. Hyperbolic stress-strain response: cohesive soils. J.
Soil Mech. Found. Divis. 89 (1), 115–144.
Matlock, H., 1970. Correlations for design of laterally loaded piles in soft
clay. Offshore Technol. Civil Eng. Hall Fame Papers Early Years, 77–
94.
Mei, G.X., Chen, Q.M., Song, L.H., 2009. Model for predicting
displacement-dependent lateral earth pressure. Can. Geotech. J. 46
(8), 969–975.
Murchison, J.M., O’Neill, M.W., 1984. Evaluation of py relationships in
cohesionless soils. In Analysis and design of pile foundations. ASCE.
pp. 174–191.
Ng, C.W., Simons, N.E., Menzies, B.K., 2004. A Short Course in Soil-
Structure Engineering of Deep Foundations, Excavations and Tun-
nels. Thomas Telford Publishing, Heron Quay, London.
Ni, P., Mangalathu, S., Mei, G.X., Zhao, Y.L., 2017a. Permeable piles: an
alternative to improve the performance of driven piles. Comput.
Geotech. 84, 78–87.
Ni, P., Mei, G.X., Zhao, Y.L., 2017b. Displacement-dependent earth
pressures on rigid retaining walls with compressible geofoam inclu-
sions: physical modeling and analytical solutions. Int. J. Geomech. 17
(6), 04016132.
Ni, P., Mei, G.X., Zhao, Y.L., 2017c. Numerical investigation of the uplift
performance of prestressed fiber-reinforced polymer floating piles.
Mar. Georesour. Geotechnol. 35 (6), 829–839.
Ni, P., Song, L.H., Mei, G.X., Zhao, Y.L., 2017d. Generalized nonlinear
softening load transfer model for axially loaded piles. Int. J. Geomech.
17 (8), 04017019.
Norris, G., 1986. Theoretically based BEF laterally loaded pile analysis.
In: Proc. 3rd Int. Conf. on Numerical Methods in Offshore Piling. pp.
361–386.
O’Neil, M., and Murchison, J. 1983. An Evaluation of p–y Relationships
in Sand. Report of the American Petroleum Institute.
Pan, J., Goh, A., Wong, K., Selby, A., 2002. Three-dimensional analysis
of single pile response to lateral soil movements. Int. J. Numer. Anal.
Meth. Geomech. 26 (8), 747–758.
Park, S.-S., Jeong, S.W., 2015. Effect of specimen size on undrained and
drained shear strength of sand. Mar. Georesour. Geotechnol. 33 (4),
361–366.
Reese, L.C., Cox, W.R., Koop, F.D., 1974. Analysis of laterally loaded
piles in sand. Offshore Technol. Civil Eng. Hall Fame Papers Early
Years, 95–105.
Rollins, K.M., Peterson, K.T., Weaver, T.J., 1998. Lateral load behavior
of full-scale pile group in clay. J. Geotech. Geoenviron. Eng. 124 (6),
468–478.
Rollins, K.M., Sparks, A., 2002. Lateral resistance of full-scale pile cap
with gravel backfill. J. Geotech. Geoenviron. Eng. 128 (9), 711–723.
Saiyar, M., Ni, P., Take, W.A., Moore, I.D., 2016. Response of pipelines
of differing flexural stiffness to normal faulting. Géotechnique 66 (4),
275–286.
Velez, A., Gazetas, G., Krishnan, R., 1983. Lateral dynamic response of
constrained-head piles. J. Geotech. Eng. 109 (8), 1063–1081.
Wesselink, B., Murff, J., Randolph, M., Nunez, I., Hyden, A., 1988.
Analysis of centrifuge model test data from laterally loaded piles in
calcareous sand. Eng. Calcareous Sedim. 1, 261–270.
Yang, Z., Jeremić, B., 2002. Numerical analysis of pile behaviour under
lateral loads in layered elastic–plastic soils. Int. J. Numer. Anal. Meth.
Geomech. 26 (14), 1385–1406.
Zhu, B., Chen, R.-P., Guo, J.-F., Kong, L.-G., Chen, Y.-M., 2011. Large-
scale modeling and theoretical investigation of lateral collisions on
elevated piles. J. Geotech. Geoenviron. Eng. 138 (4), 461–471.