Geotech Geol Eng

https://doi.org/10.1007/s10706-021-01871-3 (0123456789().,-volV)
( 01234567
89().,-volV)

ORIGINAL PAPER

Analysis of Concrete Piles Under Horizontal and Vertical

Simultaneous Loading with P-Y Method and Finite Element
Analysis
A. Ghasemipanah . R. Ziaie Moayed

Received: 31 July 2018 / Accepted: 28 May 2021

 The Author(s), under exclusive licence to Springer Nature Switzerland AG 2021

Abstract Pile foundations are often exposed to displacement is reduced compared to the situation
vertical and lateral loads. In common pile analysis, it with the absence of a vertical load.
is assumed that the effect of these two load types do
not depend on each other, and designing is usually Keywords RC piles  Pile head displacement 
done individually for each of the vertical and lateral Lateral and vertical loading  Load combination  P-Y
loads. The influence of vertical load on the response of method  AllPile
the pile lateral load can be extracted via finite element
analysis in a variety of soils. AllPile is a commercial
program that has been utilized in many executive tasks
and because of its ease of use in engineering, and in 1 Introduction
particular the analysis of different pile types under the
influence of vertical, lateral, and bending loads and When the upper layers of the soil are weak, in such a
their combinations, as a single pile or pile group. It is way that a shallow foundation could not be used to
one of the most popular and practical programs among distribute the load, deep foundations are used to
geotechnical engineers. The results of the single-pile transfer the load to the lower layers or the bedrock.
analysis under the lateral load and lateral load with The main application of deep foundations is to
vertical load in Allpile and Abaqus software are withstand vertical loads, but any structure is inevitably
presented to compare their results. Then for layered under the influence of horizontal loads. Horizontal
soil, a comparison has been made between the results loads can be caused by wind, ground stress, earth-
of Allpile analysis and numerical modeling with the quake forces, sea wave force, or berthing at seaports.
finite element method in Abaqus. The results generally Nowadays, in many practical works, the piles are
indicate that as the lateral load increases, the pile head analyzed independently by first subjecting them to a
displacement increases. Also, by increasing lateral vertical load to determine the bearing capacity and
load and applying a vertical load, the amount of their settlements, and then under the lateral load to
determine the flexural behavior. This approach is only
valid in a small lateral load (Moayed et al. 2008). In
A. Ghasemipanah (&)  R. Z. Moayed
addition to vertical load, piles are often under the
Civil Engineering Department, College of Engineering,
Imam Khomeini International University, Qazvin, Iran influence of lateral loads and bending moments. For
e-mail: [email protected] instance, in berthing structures, lateral forces are
R. Z. Moayed caused by the impact of berthing ships, pull from
e-mail: [email protected] mooring ropes and pressure of wind, current, wave and

123
 Geotech Geol Eng

floating ice, seismic force, active earth pressure, and analytical research studies (Poulos and Davis, 1980;
differential water pressure, in addition to self-weight Ramasamy, 1974; Yang and Jeremić, 2002), shows
of the structure and live load (Poulos and Davis, 1980). that in the presence of lateral load, by applying a
However, in coastal and offshore applications, lateral vertical load, lateral displacement increases. However,
loads are significantly more than 10–20% of the laboratory studies (Anagnostopoulos and Georgiadis,
vertical loads, and in some cases, studying the 1993; Jain et al. 1987) and field investigations
interaction effects of the vertical and lateral loads (Bartolomey, 1977; Karasev et al. 1977; McNulty,
combination required for systematic analysis is 1956; Sorochan and Bykov, 1976; Zhukov and Balov,
necessary. 1978) show that the lateral displacement of the pile
The design of a pile under vertical loads is under the vertical load is reduced. Goryunov (1973)
accomplished by solving the equilibrium equations carried out analytical research on a pile with a
in the direction of force, while in the piles with lateral diameter of 500 mm, a thickness of 24 mm, and a
loading the problem is solved by using non-linear length of 24 m used as a marine structure on three soils
differential equations. Investigating the ultimate bear- with different densities (dense soil, medium dense
ing capacity of a vertical pile under the influence of soil, and loose soil) are common in marine sediments.
lateral load and controlling the pile head displacement The study on the behavior of pile with lateral and
due to the interaction complexity between a semi-rigid combined loading is important. Begum (2010) con-
pile and an elastoplastic soil are important issues in ducted an experimental study of laterally loaded pile
geotechnical science (Reese and Van Impe, 2010). on a sloped surface by varied slope angles. This
Extensive research has been conducted to determine experimental study is also used to develop a graphical
the bearing capacity of the piles by researchers and nondimensional relationship between the lateral load
other investigators, including the methods proposed and maximum bending in a pile. Begum and
by Kédzi (1975), Meyerhof and Sastry (1978), Tom- Muthukkumaran (2009) conducted an experimental
linson and Woodward (2014), Vesic (1977), PIAA study of the laterally loaded pile on the sloped surface
Engineering, Poulos and Davis (1980) and the United by varied the slope angle (zero slopes, 1:1.5 and 1:2),
States Naval Forces Design Guidelines (Navy, 1982). L/D ratio (25, 30, and 35), and the relative densities of
Axial loads only cause displacements parallel to the the soil (30, 45 and 70%). Broms (1964a, b) developed
pile axis. While lateral loads can cause displacement solutions for the ultimate lateral resistance of a Pile
in any direction, the pile-soil system under the assuming the distribution of lateral soil pressure and
influence of the lateral load becomes a 3D problem. considering static of the problem and also two modes
In such cases, there is an urgent need to understand of failure and yielding of the soil along the length of
the pile behavior under the lateral load, and this the Pile were considered. Abbas et al. (2008) used to
requires a thorough investigation in this field. Many model a single Pile in layered soil by using PLAXIS-
researchers have discussed the need for an accurate 3D (FEM software) and compared the lateral load-
understanding of the pile’s behavior under the lateral carrying capacity between circular Pile and square
load, and have investigated in this field. Pile. Besides, an effect of slenderness ratio L/B is
A large number of studies have been carried out to carried out and also discussed the negative base
investigate the behavior of piles and pile groups under deflection of Pile. Judi et al. (2009) presented a series
pure lateral loads (Matlock and Reese, 1960; Poulos, of three-dimensional numerical analysis performed by
1971; Poulos and Davis, 1980). Besides, with the FLAC3D finite difference program on solid concrete
advent of latest generation computers, it is now piles with a circular section in clayey soils under a
possible to investigate the effects due to non-linearity combination of axial and lateral loading. Karthigeyan
and elastoplasticity of soil medium, asymmetric et al. (2006a, b) used a 3-dimensional finite element
loading on piles, etc. using 3-dimensional finite program GEOFEM3D, to analyze the combined
element analysis (Brown and Shie, 1991; Hong et al. loading on the pile in both sand and clayey soil. The
2003; Kimura et al. 1995; Muqtadir and Desai, 1986; initially considered pile was under horizontal loading
Trochanis et al. 1991; Wakai et al. 1995). Limited of 10 tons and vertical loading of 200 tons. The
information related to the behavior of the pile under analysis revealed that the amount of displacement
the combined vertical and lateral loads, based on decreased when soil density increases. Then, the pile

123
Geotech Geol Eng

was tested in each of the three soil types under a constant decreases the displacement of the pile head.
constant horizontal load of 10 tons and vertical loads Sorochan and Bykov (1976) also experimented an
of 0, 100, 200, and 300 tons, which resulted in in situ tests on a pile group with 4 drilling pile with a
displacements of 72, 82, 96, and 114 mm, respec- diameter of 600 mm and buried depth of 3 m, with the
tively. The analysis revealed that the lateral displace- distance between the piles equal to 3D (D equal to the
ment would increase under a constant lateral load with diameter of single piles). The thickness of the soil
the increasing vertical load. Lee et al. (1995) presented layer was 11 m and the values of u and C were 12 and
a simplified approach for the study of a row of piles 0.093 MPa, respectively. Lateral and vertical loads
used for slope stabilization in both homogeneous and were applied to the pile group, and the results are
non-homogeneous soil profiles. Begum and shown in Fig. 2. As shown in Fig. 2, displacement in
Muthukkumaran (2011) studied the behavior of a pile the presence of a vertical load of 800 kN is less than
due to the lateral load on a sloped surface and for this when there is no vertical load and it can be concluded
study, the pile is considered at the crest of the sloped that by applying the vertical load, the lateral displace-
surface. So in the field of analysis of pile due to ment decreases. Generally, it can be stated that the
combined load was not well explored. So in this paper, results of field and laboratory tests do not match the
the behavior of pile due to combined load is focused on analytical research results, and even though analytic
both and sloped grounds (1:1.5, 1:2 and 1:2.5). research expresses the increase of lateral displacement
Muthukkumaran et al. (2004) conducted the experi- by applying a vertical load, the in situ pile tests results
mental study on the aluminum pile in the sloped sandy represent a reduction in lateral displacement due to
soil surface with varying density. Muthukkumaran vertical loading. Poulos (1971) modeled pile-soil
et al. (2008) conducted the experimented study on the interactions using elastic continuum methods that
behavior of pile due to the varying surcharge load on a consider the soil to act as a 3D, linearly elastic,
sloped surface and this study was made in soil have homogeneous, isotropic, semiinfinite medium. Used
different relative densities (30, 45, and 70%). Karasev Mindlin’s equations to develop factors that account for
et al. (1977) have done in-situ tests on a pile with a additional displacements (ap) and rotations (ah) caused
diameter of 600 mm and a length of 3 m reinforces by interactions from adjacent piles. Poulos and Davies
with 8 bars of 16 mm thickness and located in the Kiev (1980) modified the elastic solution to account for
city. The desired pile is in a two-layer soil, the first nonlinearity using yield factors. The modulus of the
layer of which is a very stiff sandy loam with a subgrade reaction approach was extended to account
thickness of 6 m, and the second layer consists of loam for the soil nonlinearity. Anagnostopoulos and Geor-
with a thickness of 7 m. The pile is placed under giadis (1993) reported that the improved soil stresses
constant lateral load and the vertical loads are 0, 100, condition and the change in the plastic volume of the
200, and 400 kN, the results of which are given in soil under the vertical and lateral loads could not
Fig. 1 (Karasev et al. 1977). As shown in Fig. 1, by generally be calculated with conventional subgrade
increasing vertical load while lateral load remains
900

800

700
LATERAL LOAD (KN)

600

500

400

300

200
V=0 KN
100
V=800KN
0
0 10 20 30 40 50 60 70

LATERAL DISPLACEMENT (MM)

Fig. 1 The lateral displacement under normal load of 0 kN, (2) Fig. 2 Pile displacement by applying a lateral and vertical load
100 kN, (3) 200 kN and (4) 400 kN (Karasev et al. 1977) (Sorochan and Bykov, 1976)

123
 Geotech Geol Eng

reaction analytical methods and elastic halfspace shall

be calculated. Therefore, they presented a non-linear
three-dimensional finite element technique to analyze
this problem. Trochanis et al. (1991) studied the
behavior of a pile under the vertical and lateral loads
combination based on the non-linear three-dimen-
sional finite element method.
In this study, the piles loading test results in the site,
the finite element method, and the Allpile software
results have been investigated to determine the
accuracy and the Allpile software relations to evaluate
the pile behavior under lateral and vertical loads and
their composition and to be compared with finite
element method.

2 Mechanism of the Pile Behavior and its Fig. 3 Load transfer mechanism for piles under lateral load
Surrounding Soil under the Influence of Lateral (Basu et al. 2008)
Load
significant due to the high pressure, complete rotation
The piles are displaced when exposed to lateral loads. does not occur around a point.
If a vertical free-headed pile is considered, the applied In fact, at lower depths, the pile remains vertical
lateral load is firstly withstood by the upper layers of and displacement does not occur at the initial depths.
the soil, but if displacement increases, the pile head This displacement is shown in Fig. 5. Unlike the
will transfer the load to the lower layers of soil, and in previous state where the soil lateral resistance plays a
high displacements, the soil becomes plastic. The pile significant role in the pile bearing capacity, in this
on the surrounding soil (the soil that is along with the case, before the maximum lateral resistance is
applied load), which provides the lateral resistance of reached, the bending moment in the pile would reach
the pile to the displacement, creates compressive, the final state. Therefore, in long piles, the pile
shear stress, and strain. This general load transfer structure plays the controlling role in the bearing
mechanism is for lateral loads. The ultimate soil capacity of the pile. Like any other structure, in a pile
resistance is applied to the pile and the external forces under the lateral load, the ultimate bearing capacity of
are balanced. The soil resistance also provides the pile is not the only design criteria, but the
moment equilibrium applied to the pile (Fig. 3). displacement criteria also play the controlling role as
Assuming a short rigid free-headed pile (a pile with another design criteria, because it is always possible
the length to diameter ratio less than 10 or 12 in), it can that the pile head displacement under a load smaller
be seen that in lateral loading, the pile tends to rotate than the ultimate load would cause stresses beyond the
around a point of its length in such a way that with the allowable limit in the superstructure of the system
aid of this rotation, end of the pile passive force is used (Reese, 2001). In some references, lateral displace-
to counter the lateral force. When a rigid short pile ments in conventional structures are limited to 0.25
reaches its maximum load capacity, all soil resistance inches and in larger structures is limited to 0.5 in.
around the pile must be mobilized (Fig. 4). As shown (Matlock and Reese, 1960). Therefore, by choosing
in this figure, the mechanism of its behavior changes the ultimate load and controlling it with the pile
with the fixing of the pile head and the pile moves bearing capacity, the amount of pile head displace-
smoothly in the applied lateral load direction. In each ment under the ultimate load is controlled to ensure
of the two cases above, when the entire lateral that this criterion is checked with the necessary margin
resistance of the soil around the pile is mobilized, of confidence. Due to the involvement of many
the pile ultimate capacity has been reached. In long parameters in this issue, such as soil type, load type
piles, given that the soil’s resistance at lower depths is

123
Geotech Geol Eng

Fig. 4 The mechanism of

the behavior of a short pile
in lateral loading
(Tomlinson and Woodward,
2014)

Fig. 5 The mechanism of

the behavior of a long pile in
lateral loading (Tomlinson
and Woodward, 2014)

on the pile head, and the pile stiffness, it is not easy to the entire pile length. For instance, in piles used in
control this displacement. marine structures, that are used as an energy absorp-
The pile head condition is very influential in its tion structure, the bending moment curve, force, and
bearing capacity. The pile head can be either free or deformation changes along with the pile, are of great
fixed. As shown in Fig. 5, the pile head is capable of importance. One of the ways for reaching these curves
rotation in a free state, and in fact, there would be no is to assume elastic behavior for soil and pile
moment in its head. In a fixed state (Fig. 4), the pile (Tomlinson and Woodward, 2014). The elastic
head cannot rotate, and in addition to the sheer force, method was first introduced by Terzaghi in 1955.
the bending moment is also created. Terzaghi (1955) uses the subgrade reaction modulus
concept to study lateral piles and of course, due to the
non-linear soil behavior, the method is only applicable
3 Available Models for Investigating the Piles for small loads in the range of one third to half of the
Behavior under the Lateral Load ultimate bearing capacity, and this method could not
be used for critical loadings. In this method, the soil is
3.1 Elastic Method considered continuous, but its behavior is assumed to
be elastic. Poulos and Davis in 1980, as well as Poulos
In many cases, it is necessary to calculate the values of and Hull in 1989, investigated the effects of the pile’s
bending moments, shear force, and deformation over intervals in a pile group. The fracture modes for short

123
 Geotech Geol Eng

and long piles in the elastic method are shown in Endley et al. (1997) concerning P-Y curves, like
Figs. 6, 7, and 8, respectively. Duncan, presented a graphical method that is used to
Based on the method of load-deformation curves, evaluate the pile’s behavior in different types of soil.
proposed equations can be obtained by shearing and
bending along a pile (Evans and Duncan, 1982). The 3.2 Load–Displacement Method
variables of their equations are the pile diameter, the
soil elastic modulus, the pile inertia moment, the soil The P-Y method describes the relationship between
cohesion, and a series of numerical coefficients that the soil lateral displacement and the pile lateral force
are indicated in the tables. This method does not where p is the soil lateral resistance in the unit length
respond linearly like elastic methods. and y is the lateral displacement of the soil. Unlike the
Hence, the answers induced from numerical anal- shallow foundations where spring only indicates the
ysis and empirical experiments are closer. The state of the soil pressure, in the deep foundations under
advantages of this method are: (1) Designing will be the lateral load, the soil resistance is from the
very fast and does not require a computer and (2) In the compressive strength of the elements in front of the
proposed method, there is no need for trial and error. pile, the shear strength of the elements on both pile
The assumptions considered in extracting these graphs heads and possibly the low compressive strength of the
are: (1) The pile stiffness is assumed to be constant and elements behind the pile. Although soil compressive
(2) The resistance and soil specific strength parameters strength is much higher than shear strength, since all
are assumed constant. In this method, the following shear strength is mobilized under very small displace-
information can be extracted: ments, this resistance type is very important (Coduto,
2001).
• Displacement of the pile head on the soil surface
due to the applied force.
3.3 Subgrade Reaction Method
• Displacement of the pile head on the soil surface
due to the applied moment.
The subgrade reaction method (or the Winkler
• The value of the maximum bending moment in the
method) is one of the most common methods for
pile.
predicting the deformation and the pile bending
• Place of this bending moment.
moment. This method expresses the idea that the soil
reaction at any point is proportional to the displace-
ment of that point. In this method, the beam on elastic

Fig. 6 Deflection, slope,

bending moment, shearing
force and soil reaction
curves for elastic conditions
(Reese and Matlock, 1956)

123
Geotech Geol Eng

Fig. 7 Soil reaction and

curvature of bending
moments in short piles in
granular soils (Broms,
1964b, a)

Fig. 8 Soil reaction and

bending moments curve in
long piles in granular soils
(Broms, 1964b, a): a free
head pile and b fix head pile

subgrade model proposed by Winkler in 1876 is used. The load–displacement response of long single pile
In this model, the soil is considered as separate is usually predicted using subgrade reaction method.
springs, which only displaces when the load is applied This method has advantages that are relatively simple,
to them. In this method, the reaction module is and factors such as non-linear soil, subgrade reaction
considered for the soil. The pile is assumed to be like variations with depth, and soil layers are relatively
an elastic beam located on a subgrade. easy to combine. In the pile analysis under the
The constant value of the foundation spring can be influence of lateral loads, two stiffness parameters
related to the soil mass reaction coefficient (Terzaghi, are required: (1) Bending stiffness of the pile (Kimura
1955). Therefore, It is required that Eq. (1) be solved et al. 1995) and (2) the lateral stiffness of the soil, E, G,
and kh. If the elasticity theory is used, the pile stiffness
d4 y
Ep Ip þ kh yd ¼ 0 ð1Þ is expressed by the Young E modulus or the shear
dz4 modulus G. Anyway, the soil stiffness can be deter-
where y is the pile lateral deformation, kh the subgrade mined by the soil lateral displacement coefficient
reaction coefficient, d the pile diameter, z the pile (Eq. 2):
depth, p the soil reaction per unit length of the pile, and P
Ep and Ip, are the elasticity modulus and the pile inertia kh ¼ ð2Þ
y
moment, respectively.

123
 Geotech Geol Eng

where P is the soil reaction at a pile point per unit effective method and is almost the best one among
length of the pile (kN/m), y is the pile lateral existing methods.
deformation that is shown in Fig. 9 (Poulos, 1971). All of the numerical analyses in this section were
To solve this equation, initially it is necessary to carried out in three-dimensional space using the finite
determine the subgrade reaction coefficient (kh). There element program (ABAQUS). To analyze the actual
are three ways to determine this coefficient: (1) on configuration of the experiment, the dimensions and
lateral loading on a real-scale pile test (Reese and Cox, properties of different components (soil and pile) were
1969), (2) Plate load test and (3) Empirical equations chosen such that they represent those used in the actual
using other soil properties (Davisson, 1970; Skemp- test. The soil was discretized using 10-noded tetrahe-
ton, 1951). dron solid elements while the reinforcement was
modeled with 4-noded rectangular membrane ele-
3.4 Numerical Method ments that having negligible bending stiffness. The
total number of nodes and elements generating the
This method, in spite of other methods, can model soil given mesh was about 85,500 and 5500, respectively.
non-linearly and continuously and can also consider Boundary conditions were chosen such that displace-
the interaction between soil and pile. One of the most ment of the horizontal boundary was restricted in all
important features in this method, which other meth- directions, while vertical boundaries were restricted
ods do not have, is the ability to model the interaction horizontally and free to move in the vertical direction
between soil and pile with simultaneous consideration (smooth rigid). Also, the nodes at the base are fixed
of parameters. With the aid of this method, it is against displacements in both directions (rough rigid).
possible to observe many effects of the parameters on Four types of mesh global size are adapted as 0.034,
the soil load–displacement curve. In this method, even 0.024, 0.014, and 0.004 to find ultimate bearing
stress paths can be obtained at any depth of the pile and capacity piles. However, if a mesh is too dense, it will
at any point in the soil. Moreover, using this method, require a large amount of computer memory, and long-
each type of soil profile can be studied in a three- run times, especially for an approximate global size of
dimensional state. One of the reasons for this method’s 0.004 which is about 940 min. Ultimately, increasing
success in analyzing soil issues is the use of different the mesh density further than 0.034 (that used in the
soil behavioral models that can be used to study soil current study) produces only minor decreases in the
behavior with a wide range of strains, the effects of result (ultimate bearing capacity). In this case, an
loading speed, reduction of stiffness, etc. This is an increase from 7734 elements (current study) yields
only about 0.9 percent model accuracy in the ultimate

Fig. 9 Subgrade reaction

coefficient: (1) Actual
reaction of soil on the pile
(2) Modeling of the soil
reaction using elastic spring
after Poulos (1971)

123
Geotech Geol Eng

bearing capacity in comparison with the experimental software to calculate the behavior of a lateral-loaded
test. As the difference between experimental test and pile. This method was used by Wang and Reese,
numerical result is in a reasonable range, selection of (1993) to build the COM624S software. This program
this size of mesh (global size of 0.034) is not worrying, uses four non-linear differential equations for model-
because finer meshes not only haven’t a significant ing of piles under lateral loading in a two-dimensional
effect but also require a long run time. space. Allpile has properly restored this method in
Kimura et al. (1994) described the failure test of Windows and with proper graphic. There were also
cast-in-place concrete piles under lateral loading on studies to compare the results of Allpile and
single pile and pile group. This test was carried out on COM624S, which indicated a very good match
a single pile and a pile group with nine piles with a between their results. In the following, the computa-
diameter of 1.2 m in Japan. The purpose of this in-situ tional theory of the pile under lateral loading in Allpile
test was to investigate the interaction of the pile group would be mentioned. The four differential equations
and the total lateral resistance of the cast-in-place (Eq. 3–6) used in AllPile when applying lateral load
concrete piles. Cast-in-place piles were of reinforced are:
concrete type with a length of 30.4 m. Kimura et al.
d4 y d2 y
(1995) worked on numerical analysis of a single-pile EI þ Q  R  Pq ¼ 0 ð3Þ
and RC concrete pile group in the following year in an dz4 dz2
article. In this study, the pile lateral loading test was where Q is the axial compressive load on pile, y lateral
analyzed using the GPILE 3D finite element software. displacement of pile in depth Z, Z depth from the pile
The amount of lateral bearing capacity and lateral head, R soil reaction per unit length, E elasticity
displacement of the pile were checked, which indi- modulus of the pile, I the pile inertia moment and Pq
cates that the lateral displacement increases with distributed load along the pile length
increasing lateral load. Abdel-Rahman and Achmus,
(2006) did a numerical analysis of the pile under d3 y dy
EI þQ ¼P ð4Þ
lateral loading and vertical loading. According to the dz dz
results of piles numerical analysis in sand, a different dy
behavior under lateral and vertical loading can be ¼ St ð5Þ
dz
seen. Therefore, in the design of vertical piles, it is
necessary to consider the effect of both loads. d2 y
Iftekharuzzaman and Hawlader, (2013) investi- EI ¼M ð6Þ
dz2
gated a 21 m long steel pile with an external diameter
of 610 mm in sand under lateral loading using Abaqus Where P and M are shear and bending moments,
software. Also, piles numerical analysis was per- respectively. Allpile solves the non-linear differential
formed in the LPILE software, which is based on P-Y equations by Rees Finite Difference P-Y Method that
curves. The results of numerical analysis were com- analyzes the behavior of the soil-pile system against
pared in Abaqus and LPILE. In this comparison, it was lateral loading (shear and moment). The displacement,
concluded that finite element analysis using the Mohr– gradient, moment, and pile shear and the reaction force
Coulomb criterion can precisely model a pile under of the soil are shown in Fig. 10 (Allpile manual, 2007).
lateral loading. However, the LPILE software which is After entering data on the resistance and geometric
based on the P-Y method has some imperfections. parameters of the pile and soil and the amount of
lateral force applied in Allpile, this program has been
used to calculate the amount of displacement and
4 Numerical Model bending moment created in pile under lateral load and
reports the amount of displacement. Input parameters
4.1 Allpile Performance under Lateral Load for soil in Allpile include the SPT value (N1(SPT)),
specific gravity of the soil (c which is indicated by G in
Allpile is two-dimensional and commercially popular the software), angle of internal friction of soil particles
software that is easily available to engineers. Allpile
uses the computational method of the COM624S

123
 Geotech Geol Eng

Fig. 10 Deflection, slope, bending moment, shearing force and soil reaction curves in Allpile (Allpile manual, 2007)

(u), adhesion of soil particles (c), modulus of subgrade to being a factor of 1.2 times the experimental
reaction (k) and soil strain (e50), Which are obtained deflection when the interface was not active. This
from the experimental data and the software manual. means that for cast-in-place concrete piles some
bondage between the concrete and surround soil may
4.2 Soil-Pile Interface have occurred and thereby not allowing the slippage
normally observed along a smooth dry concrete
The interaction at the pile-soil surface can range from surface. Also construction methods for cast-in place
perfect contact where no relative sliding between soil piles are crude and this result in irregular rough sides
and pile occurs to perfect sliding conditions where no not allowing a continuous slip surface to exist.
friction develops along the shaft of the pile. Initially, it Therefore, it was concluded that cast-in-place piles
was assumed the soil and pile are both deformable were better represented by a perfectly bonded surface
bodies and can undergo finite sliding. A basic allowing no relative slippage to occur. However, for
Coulomb frictional model was used to govern the impact piles an interface model would be required to
interaction between the pile and sand surfaces. How- allow relative motion. This would suggest that the
ever, it was determined through the verification stage pile-soil interaction at the pile’s surface is dependent
that the interface characteristics were dependent on the on the placement type and can affect a model’s
placement method of the pile. Piles that are impact accuracy significantly.
driven (ID) tend to exhibit degrees of relative motion
between the pile and soil once the system is under
load. It was also determined through the FEM analysis 5 Validation of the Numerical Model
that cast-in-place (CIP) piles tend to have much less
relative sliding than their ID counterpart. The pre- In order to check the numerical modeling results with
dicted deflection for the CIP pile was approximately the actual results, three in-situ pile loading tests are
10 times greater than that of the experimental pile, available in the technical literature, which consists of
while the interface was activated. The model reduced long and short piles in sand and clay. The results of the

123
Geotech Geol Eng

in-situ pile loading tests were compared with the

results obtained from the Abaqus and Allpile soft-
wares. In the following, the details of tests and the
proposed modeling is provided. Two methods that can
be used to perform back analysis are deterministic and
probabilistic method. Deterministic method determi-
nes a unique set of parameter such as c and u by
considering the factor of safety equals to unity.
However, in order to deal with the uncertainties in
back analysis, probabilistic method offers a better
approach to analyze a multiple sets of parameters
simultaneously. The outcomes are numerous combi-
nations of parameters that result in slope failure.
Nevertheless, the results of probabilistic method are
realistic if the input parameters were correctly statis-
tically characterized. In this paper, a deterministic
approach of back analysis is used.

5.1 Case Study 1 (Tiwan)

For the model verification used in the present study,

the results of the full-scale loading tests in a project in
Taiwan, presented by Huang et al. (2001), were used.
In order to optimize the piles design of rail transport Fig. 11 The pile used in the Huang loading test (Huang et al.
systems in Taiwan, full-scale loading tests were 2001)
performed on single-pile and pile groups. The piles
execution was two types: Cast-in-place and driven
piles. In this study, the results of the lateral loading test
on a single-pile B7 were considered. The B7 pile is a
RC pile with a diameter of 1.5 m and a 34.9 m length.
Figure 11 shows a pile scheme used in the loading test
performed by Huang et al. (2001). Concrete reinforc-
ing was done using 52 bars with a diameter of 32 mm,
placed in two loops inside the pile. A hydraulic jack
was used during the test to apply the lateral load on the
pile head. Soil used in this model is silty sand (Fig. 12)
and its parameters are given in Table 1.
The parameters for concrete and steel used in
Table 2 are also given. In this study, it has been
assumed that the soil is completely submerged and has
0
a specific buoyancy weight of c ¼ 9kN=m3 . In Fig. 12 Soil model intended to simulate the Houang loading
test (Huang et al. 2001)
evaluating the model’s accuracy, in addition to the
studies by Huang, the numerical analysis results of
As shown, the 3D finite element method exhibits more
Conte et al. (2013) is used. Conte et al. (2013) used a
horizontal displacement values than the values of in-
three-dimensional modeling in Abaqus to analyze the
situ pile tests, but the P-Y method exhibits less
value of Huang’s work numerically. Comparison of
displacement.
the field test (Huang et al. 2001), Abaqus (Moayed
et al. 2008) and Allpile results is provided in Fig. 13.

123
 Geotech Geol Eng

Table 1 Geotechnical properties of soil considered in Huang 3500

model (Huang et al. 2001)

3000
Layer number E (MPa) m0 c’ (kPa) u0 ð Þ w0 ð Þ
2500

LATERAL LOAD (KN)

1 80.8 0.3 0 33 0
2 150.02 0.3 0 34 0 2000

3 150.02 0.3 0 28 0
1500
4 228.02 0.3 0 33 0
Huang et al. 2000
5 228.02 0.3 0 28 0 1000
Conte et al.2012 (FEM)
6 228.02 0.3 0 30 0
500 Abaqus (Taheri et al.2015)

AllPile
0
0 20 40 60 80 100 120 140
5.2 Case Study 2 (Japan) LATERAL DISPLACEMENT(MM)

Kimura et al. (1995) presented the installed in-situ RC Fig. 13 Verification of the model with the pile loading test
results at the Huang site (Huang et al. 2001)
pile behavior with a length of 4.30 and 1.2 m diameter
under the lateral load on the Kishiwada Bridge,
Hanshin Osaka highway. The soil is composed of a Table 3 Soil Parameters of Hanshin Osaka highway (Kimura
10-m layer of sand embankment on top and a mixture et al. 1995)
of alternate layers of clay and sand at the bottom. The Layer number E ðkPaÞ m cðkPaÞ uð Þ cðkPaÞ
soil profile is shown in Table 3. Comparison of the pile
tests (Kimura et al. 1995), Abaqus (Moayed et al. 1 28,000 0.333 0 33.9 19
2008) and Allpile results is shown in Fig. 14. In 2 25,500 0.333 0 34.4 21
Fig. 14, the horizontal displacement value in the two- 3 20,000 0.38 62.5 0 17
dimensional modeling (P-Y method) is more than the 4 44,000 0.333 0 37 21
displacement by the 3D finite element modeling and
in-situ test of the pile.
(2006a, b) also performed a three dimensional 3D
5.3 Case Study 3 (Kiev)
finite element analysis of cast-in-place concrete pile
under the lateral load with GEOFEM 3D, and
A full-scale pile test under the vertical and horizontal
compared based on the results of in-situ tests by
loads was performed to investigate the effect of these
Karasev et al. (1977). The comparison results of the
loads on the piles bearing capacity by Karasev et al.
pile tests, Abaqus (Moayed et al. 2008) and Allpile are
(1977). The diameter and length of the concrete piles
shown in Fig. 15. As shown in Fig. 15, the values
tested were 3 m and 600 mm, respectively, which can
obtained from Allpile (P-Y method) show higher
be called a short pile (rigid). The bed soil consists of
displacement values than Abaqus (3D finite element
two layers: the first layer of a very stiff sandy loam
method) and in-situ tests.
with a thickness of 6 m, and the second layer of sandy
clay with a lower stiffness and a thickness of 7 m. Soil
parameters are shown in Table 4. Karthigeyan et al.

Table 2 Concrete and steel properties used in modeling the Huang pile (Huang et al. 2001)
Concrete Properties Steel Properties
Ec ðMPaÞ mc fc ðMPaÞ fct ðMPaÞ Es ðMPaÞ ms fy ðMPaÞ

32,173 0.2 27.5 2.7 471 0.3 210

123
Geotech Geol Eng

1400 120

1200
LATERAL LOADING (KN)

100
1000

LATERAL LOADING (KN)

800 80

600
60
400 Kimiura et al. 1994
Abaqus (Taheri et al.2015)
200 40 Karaserv et al. 1977
AllPile
Karthigeyan et al. 2006
0
0 5 10 15 20 25
20 Abaqus (Taheri et al.2015)
LATERAL DISPLACEMENT (MM)
AllPile

Fig. 14 Comparison of Allpile results with field test and 0

0 5 10 15 20 25
numerical model
LATERAL DISPLACEMENT (MM)
6 Soil and Concrete Modeling in Abaqus Fig. 15 Verification of the model with the results of field test
(Karasev et al. 1977) and numerical data of Moayed et al. (2008)
The constructed soil model is shown in Fig. 16. In this and Karthigeyan et al. (2006a, b)
model, the circular environment is used, in which case
the pile elements are added. To increase the analysis
accuracy and the plastic strain calculation, smaller
areas are used around the pile. For the soil discretiza-
tion of cubic elements in 8 nodes of various dimen-
sions has been used and they have been determined to
show the changes in the stress and plasticity created in
the soil properly. Also, in the vertical direction, the
half a meter elements are used to the depth below the
pile and then one-meter elements are used. In mod-
eling the soil dimensions, along the horizon, 20 times
the diameter of the piles from its axis and in the
vertical direction, as far as the length of the pile, it was
distanced from its end.
By Huang et al. (2001), a series of penetration (SPT
and CPT) tests and flat dilatometer (DMT) tests were
performed before pile installation. Also, two penetra-
tion tests were accompanied with measuring the shear Fig. 16 Finite element mesh for the a Soil domain, b Pile and,
wave velocity (SCPT). In addition, laboratory tests c Reinforcement used in FE analysis
were conducted on the soil samples taken from the
boreholes. Down to the depth of 80 m, the soil can be of sandy silt (Fig. 12). These layers are characterized
generally categorized in the silty sand class with layers by the resistance at the tip (qc), which was about

Table 4 Soil parameters (Karasev et al. 1977)

Soil type Layer thickness E ðkPaÞ m cðkPaÞ uð Þ cðkN=m3 Þ

Very stiff sandy loam 6m 25,000 0.35 18 18 16.5

Sandy clay 7m 20,000 0.4 24 14 17.2

123
 Geotech Geol Eng

2–3 MPa. The mean values of qc in the sand were concrete behaves the same as elastic–plastic materials.
10 MPa. The dilatometer module (ED) shows a profile In comprehensive stresses, it is assumed that the
similar to qc that has a value of 10 MPa in the silty concrete is an elastoplastic material with the isotropic
layers and 40–60 MPa in the sand. The water level was hardening and the associated flow rule (Hibbit et al.
determined to be approximately 1 m above ground 2010; Hilleborg et al. 1976). For tensile stresses,
level. In addition, the CPT and DMT tests were another method is used to simulate the crack devel-
performed after the pile construction and before the opment phenomenon. The crack is developed when
loading test. Comparison of the results showed that the the stress state reached ‘‘crack detection’’ at the plastic
construction and implementation of the bored pile B7 level. This level allows defining the direction of the
had a small effect on the surrounding soil to a depth of crack. It is usually assumed that the direction of the
10 m. To calculate the initial stresses in soil layers crack is normal at this level. The plastic level is Mohr–
before the loading test, the coefficient of lateral resting Coulomb (Hibbit et al. 2010). When the direction of
pressure which was obtained by the value of stress the cracks is detected, the damaged elasticity theory is
index (KD) derived from dilatometry tests of Huang used to describe the behavior of the concrete after
et al. (2001), was calculated based on experimental failure with open cracks (Jirasek and Bazant, 2002).
equations of Marchetti (1997). Based on the KD values According to this theory, it is assumed that the
obtained from the DMT test in Huang et al. (2001) and material loses its strength by a softening mechanism
the equations provided by Marchetti (1997), k0 = 0.72 which is mainly a result of damage, i.e. open cracks are
was considered for all soil layers. In the present study, observed with loss of hardness in the materials, while
for modeling the short and long piles since the any permanent strain associated with cracking is
information obtained from field experiments is more ignored. The main advantage of the proposed behav-
reliable, a single-layer silty sand with the character- ioral model is that it only needs stress–strain relations
istics of the second layer of soil in Huang et al. (2001) of concrete in uniaxial conditions as an input param-
was considered and the k0 value were assumed 0.72. eter. The relationship between these two parameters,
Specifically, a linear elastic perfectly plastic model used in the present study, is shown in Fig. 17. These
with Mohr–Coulomb failure criterion and flow rule of relationships include components with a linear func-
non-associated type is considered for modelling the tion which are completely defined by the following
soil behavior. The soil parameters required by this parameters of the behavior model: elasticity modulus
constitutive model are the shear modulus, G, Poisson’s (Ec), compressive strength (fc) and tensile strength (fct)
ratio, m, shearing resistance angle, u, effective cohe- of concrete. As shown in Fig. 17, when the compres-
sion, c, and angle of dilatancy, w. sive stress is less than 0.3fc, the behavior of the
The pile is made of reinforced concrete and for its concrete is elastic. For higher stresses, plastic defor-
discretization like a soil, 8 element nodes were used. mation occurs due to hardening of materials. After a
These elements dimensions are chosen to be compat-
ible with the soil elements around them. Concrete and
steel in the pile are modeled non-linearly (Fig. 16a).
The selected structural model for concrete is based on
the theory of plasticity for compressive stresses and
crack mechanism for tensile stresses. This model is
suitable for reinforced concrete structures subjected to
monotonic loading. Concrete is assumed to behave
like elasto-plastic materials (Hillerborg et al. 1976). In
Figs. 16 and 16, the 8-node cubic element and created
pile model are shown. The behavioral model chosen
for the concrete based in the plasticity theory is related
to the compressive stresses and the fracture mechanics
is related to the tensile stresses. This model is useful
for the structures of reinforced concrete which are Fig. 17 Stress–strain relationship in this study for the concrete
influenced by monotonic loading. It was assumed that under uniaxial conditions

123
Geotech Geol Eng

softening strain is obtained, failure occurs in 1.81 Ec/fc 7 Result and Discussion
strain. For tensile stresses, there is a linear elastic
behavior until a linear tensile strength is achieved. By 7.1 The Behavior of a Long Pile under the Vertical
these conditions, materials are characterized by a and Lateral Loads Combination
softening strain behavior along with a strength that
decrease with the increase in strain for the approxi- 3D models have been studied using a finite element
mate calculation of the tension stiffening in the method with Abaqus software as well as 2D models
interaction between the pile and the reinforcement using piles for Allpile. The desired pile has a length of
(Conte et al. 2013). The stress–strain relations for steel 34.9 m and a diameter of 1.5 m, and its characteristics
bars are shown in Fig. 18. The behavior of the steel is are identical to those of the pile used in the Huang et al.
elastic perfectly plastic and is uniform for compres- (2001) research. In soil modeling, drained conditions
sion and tension. The parameters required for this were considered. The used soils are silty sand and the
behavioral model include the modulus of elasticity Es, parameters assumed for the soil is given in Table 5. In
the Poisson ratio ms and the yield strength of steel fy. order to examine the pile behavior under static lateral
In the friction contact, a penalty formulation has load and the vertical and lateral loads combination, the
been used. In this method, the friction coefficient (l) is pile is placed in a free-head position. The pile was
defined as l = tan (ul), where ul is the pile/soil analyzed under the lateral load for the three modes of
interface friction angle. The value of ul depends on ultimate load, allowable load and without vertical
surface roughness of the pile and effective angle of load. The ultimate vertical load of the pile was
internal friction, u0 . Poulos (1971) recommended the calculated by empirical relation. In addition, the
value of ul for piles in the range of 0.5u0 to 0.9u0 , maximum permissible load on the pile was 8800 kN,
where the lower values are for smooth steel piles. The which by considering the safety coefficient of 3,
value of l = 0.44 is used in this study. In the normal allowable load was approximately 3500 kN. The
contact between two surfaces, again the penalty Allpile results are shown in Fig. 19, according to
formulation has been used along with the hard contact which, in the absence of vertical load, increasing the
with the possibility of separation between the two lateral load would increase the pile free-head dis-
surfaces (Hibbit HD, 2010). In Allpile software, the placement. Moreover, in the case of a vertical constant
soil resistance is modeled as discontinuous nonlinear load of 8800 kN, when lateral load increases, pile free-
springs defining the properties empirically. Moreover, head displacement decreases in comparison with the
the pile/soil interface behavior cannot be modeled in case when there is no vertical load. As shown in
the p-y curve method (Iftekharuzzaman and Hawlader, Fig. 20, the Abaqus and finite element method results
2013). also show that with increasing lateral load, in the
absence of vertical load, the free-head displacement
would increase. In this case, the pile displacement is
high at the initial depth and approximately after the
depth of 7 or 8 m, the pile displacement is very close to
zero.
As can be seen, due to the fact that the soil’s
resistance at lower depths is significant because of the
high amount of surrounding pressure, the full rotation
is not around a point. In fact, at more depths, the pile

Table 5 Geotechnical properties of soil used in finite element

and 2D analysis
E ðMPaÞ m c ðkPaÞ uð Þ csat ðkN=m3 Þ k0

150.02 0.3 0 34 9 0.72

Fig. 18 Stress–strain relationship in this study for the steel bars

123
 Geotech Geol Eng

Fig. 19 Lateral DEFLECTION (MM)

displacement of long pile -10 0 10 20 30 40 50 60 70
modeling in Allpile

DEPTH (M)

Allpile (H=1000KN , V=8800KN)

Allpile (H=1000KN , V=0KN)
Allpile (H=1500KN , V=8800KN)
Allpile (H=1500KN , V=0KN)
Allpile (H=2100KN , V=8800KN)
Allpile (H=2100KN , V=0KN)

DEFLECTION (MM) Allpile and Abaqus. As can be seen in Fig. 21, with the
-10 0 10 20 30 40 50 60 70 80 lateral load and no vertical load, the results of the 2D
0
model are in good agreement with the 3D model.
5 Nevertheless, with the increase of the lateral load, the
accuracy and consistency of the results are reduced
10 and Allpile shows fewer values and makes it difficult
to cite Allpile results. However, with vertical load and
DEPTH (M)

15
lateral load simultaneously (Fig. 22), consistency of
20 the results of the two models is reduced and Allpile
shows more values than Abaqus.
25 Abaqus (H=1000KN , V=0KN)
Abaqus (H=1000KN , V=8800KN)
Abaqus (H=1500KN , V=0KN)
30
Abaqus (H=1500KN , V=8800KN)
Abaqus (H=2100KN , V=0KN) DEFLECTION (MM)
35 Abaqus (H=2100KN , V=8800KN)
-10 0 10 20 30 40 50 60 70 80
0

Fig. 20 Lateral displacement of long pile modeling in Abaqus

5

remains vertical and only deforms at the initial depth

10
(about one third of the top of the pile). According to
the Fig. 20, the displacement values of the pile head
DEPTH (M)

15
for 1000, 1500 and 2100 kN loads are 15, 34 and
69 mm, respectively. By increasing lateral load from 20

1000 to 1500 (50% increase), from 1500 to 2100 (40% Allpile (H=1000KN , V=0KN)
25
increase) and from 1000 to 2100 kN (more than twice), Allpile (H=1500KN , V=0KN)
Allpile (H=2100KN , V=0KN)
the pile head displacement was 2.26, 2 and 4.6 times 30 Abaqus (H=1000KN , V=0KN)
more, respectively. By applying a vertical constant Abaqus (H=1500KN , V=0KN)
load, this displacement would be reduced, and from 35 Abaqus (H=2100KN , V=0KN)

this point, it is consistent with the Allpile program.

Figures 21 and 22 are plotted to compare the results of Fig. 21 Comparison of the results of Allpile and Abaqus
software under lateral load in long piles

123
Geotech Geol Eng

DEFLECTION (MM) DEFLECTION (CM)

-10 0 10 20 30 40 50 60 -4 -2 0 2 4 6 8 10 12 14 16 18
0 0

5 1

10 2
DEPTH (M)

15

DEPTH (M)
3

20 4
Abaqus (H=1000KN , V=8800KN)
25 Allpile (H=1000KN , V=8800KN) 5 Abaqus (H=370KN , V=880KN)
Abaqus (H=1500KN , V=8800KN) Abaqus (H=370KN , V=0KN)
30 Abaqus (H=480KN , V=880KN)
Allpile (H=1500KN , V=8800KN) 6
Abaqus (H=480KN , V=0KN)
Abaqus (H=2100KN , V=8800KN)
Abaqus (H=560KN , V=880KN)
35 Allpile (H=2100KN , V=8800KN)
7 Abaqus (H=560KN , V=0KN)

Fig. 22 Comparison of the results of Allpile and Abaqus under

lateral and vertical loads combination in long piles Fig. 23 Lateral displacement of short pile modeling in Abaqus

7.2 Short Pile Behavior under the Vertical vertical load changes the stresses in the pile and the
and Lateral Loads Combination displacement resulting from the load and the reaction
of the soil would be smaller, thereby limiting the pile
As earlier mentioned, the short and long piles behavior lateral displacement. The vertical loads applied to the
is different under the lateral load. Accordingly, their pile were obtained by empirical relations with an
lateral displacement will also be different. In this ultimate load of 2640 kN and an allowable load of 880
section, behavior of pile is reviewed. The desired pile kN. Figure 24 also shows the results of a short pile
has a length of 7 m and a diameter of 1.5 m, and other with 2D model (Allpile), which shows no significant
specifications of pile and soil are considered to be the difference between the displacement values of the pile
same as a long pile. In Fig. 23, a short pile behavior is head in the presence or absence of vertical load, and
shown under pure lateral load and under vertical and these results are slightly far from reality.
lateral loads combination in the 3D model. As was As shown in Figs. 25 and 26, in the case of pure
expected, a short pile under the lateral load inside the lateral loading and in the case of vertical and lateral
soil was rotated around a point in its lower half, and loads combinations, the results of Allpile and Abaqus
from this point, there was a good consistency between
DEFLECTION (MM)
the behavior of piles and the Brahms method.
-20 -10 0 10 20 30 40 50
According to the investigations, it is necessary to 0
mention that if the excessive force is applied to the pile
head, the soil around the pile will fail and cause the 1

short pile to move inside the soil. This behavior

2
depends on the degree of the soil stiffness and the pile
inside it and can cause the failure of the pile structure
DEPTH (M)

3
or the soil. According to these curves, it is observed
that near the depth of 6 m, the deformation of the pile 4

is zero, which indicates the rotation of the pile at this

5 Allpile (H=370KN , V=880KN)
point. Allpile (H=370KN , V=0KN)
Also, with increase of the lateral load and the Allpile (H=480KN , V=880KN)
6
presence of vertical load in short piles with a 3D Allpile (H=480KN , V=0KN)
Allpile (H=560KN , V=880KN)
model, the amount of pile free-head displacement, 7
Allpile (H=560KN , V=0KN)
such as a long pile, is more than the case with lateral
load and without vertical load. The reason is that the Fig. 24 Lateral displacement of short pile modeling in Allpile

123
 Geotech Geol Eng

are not consistent with each other, and Allpile has a lot DEFLECTION (MM)
of displacement compared to Abaqus for pile free- -20 -10 0 10 20 30 40 50
0
head s and its results are very conservative. Therefore,
in modeling of short piles with Allpile, the design 1

would be conservative. In order to further evaluate the

2
results, the above results were compared with the

DEPTH (M)
results of modeling by Iftekharuzzaman and Hawla- 3
der, (2013). Iftekharuzzaman and Hawlader, (2013)
4
studied long steel pile with 21 m length and outer
diameter of 610 mm in sand under lateral load which 5 Abaqus (H=370KN , V=880KN)
was applied on a point with the distance of 0.3 m from Abaqus (H=480KN , V=880KN)

6 Abaqus (H=560KN , V=880KN)

pile head, by using Abaqus. Also, numerical analysis Allpile (H=370KN , V=880KN)
of the piles was performed in LPILE program, which is 7
Allpile (H=480KN , V=880KN)
Allpile (H=560KN , V=880KN)
based on P-Y curves. The pile is located in Texas and
the soil profile is shown in Fig. 27. The analysis results
Fig. 26 Comparison of the results of Allpile and Abaqus under
of Lpile and Abaqus programs are shown in Fig. 28. lateral and vertical loads combination in short piles
As can be seen, the results of a 3D finite element
modeling with Abaqus are in good agreement with the Due to the axial load, the lateral load-carrying capacity
results of the pile in-situ tests and in some cases show of the pile is increased in the sandy soil. This can be
higher values. However, the P-Y 2D modeling with explained as the heavy axial load acting on the pile
Lpile program shows more displacement and this head creates a fixity condition for the pile and hence
suggests that, in terms of overall outcomes, Allpile reduces the lateral deflection of the pile head. Also due
used in this study is consistent with Lpile. Figure 29 to the angle of internal friction (u), the lateral load-
also compares the displacement of the pile head due to carrying capacity gets increased. Compared to sloped
the application of various lateral forces in Lpile, ground, the effect of a settlement is less on the
Abaqus and in-situ tests in the pile site. The lines with horizontal ground in the combined loading condition.
a solid circle are in fact the pile head displacement The lateral load-carrying capacity of the pile in sandy
results in in-situ tests, which, as can be seen, are in soil due to combined loading on a sloped ground also
good agreement with the results of Abaqus shown with gets increased but it is until a certain axial load limit
black lines. But the results from Lpile shows with only. Beyond that, it gets decreased. This can be
dashed lines show higher values than the other two. explained as the effect of settlement is high in the
sloped ground than the horizontal ground. Due to axial
DEFLECTION (MM) load, there is no lateral displacement of the pile in the
-20 -10 0 10 20 30 40 50 horizontal ground, but it occurs on sloped ground. The
0
reason for this variation of the lateral capacity of the
1 pile is due to the lateral confinement of the soil around
the pile. It can be also explained as when the pile is on
2
horizontal ground, it is fully confined with the soil but
it is not the case on sloped ground. So due to the lateral
DEPTH (M)

3
load, the soil in the horizontal ground can offer more
4 resistance to the pile which is not present for the pile
Abaqus (H=370KN , V=0KN) on sloped ground.
5
Abaqus (H=480KN , V=0KN)
Abaqus (H=560KN , V=0KN)
6 Allpile (H=370KN , V=0KN)
Allpile (H=480KN , V=0KN) 8 Conclusions
7 Allpile (H=560KN , V=0KN)

In the present study, the short and long RC piles

Fig. 25 Comparison of the results of Allpile and Abaqus under behavior under lateral load and lateral and vertical
lateral load in short piles

123
Geotech Geol Eng

Fig. 27 Pile and soil profile

in Texas (Iftekharuzzaman
and Hawlader, 2013)

Fig. 28 Comparison of test results at the pile place, Abaqus and

Lpile software by Iftekharuzzaman and Hawlader, (2013)

loads combination in 2D and 3D models in Allpile and

Abaqus was analyzed and the following results were
obtained: Fig. 29 Lateral displacement of pile (Iftekharuzzaman and
Hawlader, 2013), solid lines: FE analysis; dashed line: LPILE;
1. When the vertical load is applied simultaneously solid circles: measured at ground line in pile load test
with the lateral load, the influence is felt only at
greater deflection levels, whereas the influence is load. In general, the influence of vertical load on
considerable at almost all the deflection levels the lateral response of the pile is less in the
when the vertical load is applied prior to the lateral

123
 Geotech Geol Eng

simultaneously applied vertical and lateral loads References

case.
2. The maximum Percentage Improvement in lateral Abbas JM (2008) Single pile simulation and analysis subjected
to lateral load. Electron J Geotech Eng 13:1–15
displacement is up to 1.3% in long piles and 1.5%
Abdel-Rahman K, Achmus M (2006) Numerical modeling of
in short piles for sandy soils in the simultaneously the combined axial and lateral loading of vertical piles, 6th
applied vertical and lateral cases. This is attributed European Conference on Numerical Methods in Geotech-
to the development of additional lateral soil nical Engineering, Graz, Austria, September
Allpile (2007) ‘‘Allpile Version 7 User’s Manual, Volume 2’’,
stresses in front of the pile and additional frictional
CivilTech Software, Bellevue, WA, USA, 2007
resistance developed along its length. Anagnostopoulos C, Georgiadis M (1993) Interaction of axial
3. In the two-dimensional modeling with Allpile, the and lateral pile responses. J Geotech Eng 119:793–798
results generally indicate that in sandy soils, as the Bartolomey A (1977) Experimental analysis of pile groups
under lateral loads, Proceedings of the Special Session 10
lateral load increases, the pile head displacement
of the Ninth International Conference on Soil Mechanics
increases. Also, by increasing lateral load and and Foundation Engineering, pp. 187–188
applying vertical load, the amount of displace- Basu D, Salgado R (2007) Elastic analysis of laterally loaded
ment is reduced compared to the situation with pile in multi-layered soil. Geomech Geongng Int J
2:183–196
absence of vertical load. For both long and short
Begam NA, Muthukkumaran K (2009) Experimental investi-
piles, the same behavior mode was visible in 3D gation on single model pile in sloping under lateral load. Int
modeling by finite element method in Abaqus. J Geotech Eng 3:133–146
4. In terms of consistency of the 2D and 3D model Begum NA (2010) Soil structure interaction of laterally loaded
pile and pile group on sloping. National Institute of
results, Allpile (2D model) has shown a good
Technology, Trichy
consistency with the Abaqus software in the case Broms B (1964a) The lateral resistance of piles in cohesive soils.
of long pile or when a lateral and vertical loads J Soil Mech Found Div ASCE 90:27–63
combination was applied. Although in the pres- Broms BB (1964b) Lateral resistance of piles in cohesionless
soils. J Soil Mech Found Div 90:123–158
ence of vertical load, this consistency is less than
Brown DA, Shie C-F (1991) Some numerical experiments with
the absence of it. a three dimensional finite element model of a laterally
5. However, in the case of short piles, the Allpile loaded pile. Comput Geotech 12:149–162
results are much more conservative, so that by Coduto D (2001) Foundation analysis and design. Prentice Hall
Conte E, Troncone A, Vena M (2013) Nonlinear three-dimen-
increasing the lateral load, the results consistency
sional analysis of reinforced concrete piles subjected to
with the Abaqus program is reduced and some- horizontal loading. Comput Geotech 49:123–133
times even double the results of the 3D analysis. Davisson M (1970) Lateral load capacity of piles. Highway
Moreover, the overall conclusion is that Allpile Research Record
Endley S, Yeung A, Vennalaganti K (1997) Lateral Loads on
does not provide accurate results under the
Long Piles and Piers, Proceeding of the Texas Sec-
combination of vertical and lateral loads and is tion ASCE, Volume Housyon, Texas, pp. 181–190
often conservative (in short piles especially). Duncan E (1982) Duncan JM. 1982. Simplified analysis of lat-
erally loaded piles. Report UCB/GT/82–04. University of
California
Goryunov B (1973) Analysis of piles subjected to the combined
Authors Contributions All authors contributed extensively to action of vertical and horizontal loads (discussion). Soil
the work presented in this paper. In other words, all authors Mech Found Eng 10:10–13
discussed the results and implications and commented on the Hibbit HD, Karlsson BI, Sorensen P (2010) ABAQUS theory
manuscript at all stages Mr. Ghasemipanah performed and user’s manual
numerical models, analysed data and wrote the paper. Dr. Hillerborg A, Modéer M, Petersson P-E (1976) Analysis of
Ziaie Moayed as a supervisor designed the project and devised crack formation and crack growth in concrete by means of
the project, the main conceptual ideas and proof outline. Mr. fracture mechanics and finite elements. Cem Concr Res
Badakhshan developed analytical equations, processed the 6:773–781
numerical data and extended the numerical model. Mr. Hong S, Lee F, Yong K (2003) Three-dimensional pile-soil
Hoseinzadeh drafted the manuscript and designed the interaction in soldier-piled excavations. Comput Geotech
figures as well as verified the numerical models. 30:81–107
Huang A-B, Hsueh C-K, O’Neill MW, Chern S, Chen C (2001)
Effects of construction on laterally loaded pile groups.
J Geotech Geoenviron Eng 127:385–397

123
Geotech Geol Eng

Iftekharuzzaman M, Hawlader BC (2013) Numerical modeling under lateral soil movement. In: 13th world conference on
of lateral response of long flexible piles in sand. Geotech earthquake engineering, Canada, 2147
Eng J SEAGS & AGSSEA 44:25–31 Muthukkumaran K, Sundaravadivelu R, Gandhi SR (2008)
Jain N, Ranjan G, Ramasamy G (1987) Effect of vertical load on Effect of slope on P-Y curves due to surcharge loading. Soil
flexural behavior of piles. Geotechnical Engineering Found Japan Geotech Soc 48(3):361–369
185–204 Navy U (1982) Foundations and earth structures design manual
Jirasek M, Bazant ZP (2002) Inelastic analysis of structures. 7.2. Dept. of the Navy, Navy Facilities Engineering
Wiley, London Command, Alexandria, Va
Judi A, Rabe K (2009) Three dimensional analysis of soil con- Poulos HG (1971) Behavior of laterally loaded piles: I-single
crete piles in clayey soils under lateral loading. In: 2nd piles. J Soil Mech Found Div ASCE 97(SM 5):711–731
international conference on new developments in soil Poulos HG, Davis EH (1980) Pile foundation analysis and
mechanics and geotechnical engineering, Near East design. Wiley, New York
University, Nicosia Ramasamy G (1974) Flexural behavior of axially and laterally
Karasev O, Talanov G, Benda S (1977) Investigation of the loaded individual piles and group of piles. Ph. D. thesis,
work of single situ-cast piles under different load combi- Indian Institute of Science, Bangalore, India
nations. Soil Mech Found Eng 14:173–177 Reese L (2001) Van Impe. WF, Single piles and pile groups
Karthigeyan S, Ramakrishna V, Rajagopal K (2006a) Influence under lateral loading, AA Balkema, Rotterdam
of vertical load on the lateral response of piles in sand. Reese L, Cox W (1969) Soil behavior from analysis of tests of
Comput Geotech 33:121–131 uninstrumented piles under lateral loading, Performance of
Karthigeyan S, Ramakrishna V, Rajagopal K (2006b) Numeri- deep foundations. ASTM International 160–176.
cal investigation of the effect of vertical load on the lateral Reese LC, Matlock H (1956) Non-dimensional solutions for
response of pile. J Geotech Geoenviron Eng ASCE laterally-loaded piles with soil modulus assumed propor-
133(5):512–521 tional to depth. Association of Drilled Shaft Contractors
Kédzi A (1975) Pile foundations-foundation engineering Reese LC, Van Impe WF (2010) Single piles and pile groups
handbook. Van Nostrand Reinhold under lateral loading. CRC Press
Kimura M, Adachi T, Kamei H, Zhang F (1995) 3-D finite Skempton A (1951) The bearing capacity of clays
element analyses of the ultimate behavior of laterally loa- Sorochan E, Bykov V (1976) Performance of groups of cast-in
ded cast-in-place concrete piles, Proceedings of the Fifth place piles subject to horizontal loading. Soil Mech Found
International Symposium on Numerical Models in Eng 13:157–161
Geomechanics, NUMOG V. AA Balkema, pp. 589–594 Terzaghi K (1955) Evalution of conefficients of subgrade
Kimura M, Kosa K, Morita Y (1994) Full-scale failure tests on reaction. Geotechnique 5:297–326
lateral loading cast-in-place concrete piles, Proc. 5th Int. Tomlinson M, Woodward J (2014) Pile design and construction
Conf. and Exhibition on Piling and Deep Foundation, practice. CRC Press
pp. 1–5.15 Trochanis AM, Bielak J, Christiano P (1991) Three-dimensional
Lee CY, Hull TS, Poulos HG (1995) Simplified pile-slope sta- nonlinear study of piles. J Geotech Eng 117:429–447
bility analysis. Comput Geotech 17:1–16 Version A (2010) 6.10, Hibbit Karlsson and Sorensen. Inc.:
Marchetti S (1997) The flat dilatometer: design applications. In: Pawtucket, RI
Proc III geotechnical engineering conference, keynote Vesic AS (1977) Design of pile foundations. NCHRP synthesis
lecture, Cairo University; p. 1–26 of highway practice
Matlock H, Reese LC (1960) Generalized solutions for laterally Wakai A, Ugai K, Gose S (1995) The 3-D FE analysis of model
loaded piles. J Soil Mech Found Div 86:63–94 group piles embedded in sand, Proc. of 5th Int. Symp. on
McNulty J (1956) Thrust loading on piles. J Soil Mech Found Num. Models in Geomechanics, pp. 613–618
Div 82:1–940 Wang ST, Reese LC (1993) COM624P: laterally loaded pile
Meyerhof G, Sastry V (1978) Bearing capacity of piles in lay- analysis program for the microcomputer, Version 2.0. US
ered soils. Part 1 Clay Overlying Sand. Can Geotech J Department of Transportation, Federal Highway Admin-
15:171–182 istration, Office of Technology Applications
Moayed RZ, Judi A, Rabe BK (2008) Lateral bearing capacity of Yang Z, Jeremić B (2002) Numerical analysis of pile behavior
piles in cohesive soils based on soils’ failure strength under lateral loads in layered elastic–plastic soils. Int J
control. EJGE 13:1–11 Numer Anal Meth Geomech 26:1385–1406
Muqtadir A, Desai CS (1986) Three-dimensional analysis of a Zhukov N, Balov I (1978) Investigation of the effect of a vertical
pile-group foundation. Int J Numer Anal Meth Geomech surcharge on horizontal displacements and resistance of
10:41–58 pile columns to horizontal loads. Soil Mech Found Eng
Muthukkumaran K, Begum NA (2011) Finite element analysis 15:16–22
of laterally loaded piles on sloping. Indian Geotech J
41(3):155–161
Publisher’s Note Springer Nature remains neutral with
Muthukkumaran K, Sundaravadivelu R, Gandhi SR (2004)
regard to jurisdictional claims in published maps and
Effect of sloping on single pile load deflection behavior
institutional affiliations.

123