Ain Shams Engineering Journal 13 (2022) 101734

Contents lists available at ScienceDirect

Ain Shams Engineering Journal

journal homepage: www.sciencedirect.com

Evaluating the lateral subgrade reaction of soil using horizontal pile load
test results
Dina M. Ors, Ahmed M. Ebid ⇑, Hisham A. Mahdi
Faculty of Engineering and Technology, Future University, Egypt

a r t i c l e i n f o a b s t r a c t

Article history: The aim of this research is to present a correlation between the measured coefficient of variation of lat-
Received 20 October 2021 eral subgrade reaction with depth (f) and both recommended values in codes and the elastic modulus of
Revised 26 January 2022 soil. In order to achieve that aim, four 3D FEM models were developed to simulate four full-scale, non-
Accepted 3 February 2022
working pile load tests constructed in different soil profile sites. The soil properties in the models (includ-
Available online 23 February 2022
ing the elastic modulus) were calibrated using experimental measurements. Besides that, both experi-
mental and recommended values of (f) were estimated. Comparing these values showed that the
Keywords:
experimental values are about (6 to 12) times the recommended ones for granular soils and increase
Laterally loaded piles
Pile load test
to about 30 times for intact rock. Finally, a non-linear regression was used to develop the correlation for-
Lateral subgrade reaction mula. The results showed perfect logarithmic relation between (Es) & experimental values of (f).
Nonlinear FEM modeling Ó 2022 THE AUTHORS. Published by Elsevier BV on behalf of Faculty of Engineering, Ain Shams Uni-
versity. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

1. Introduction capacity and the pile is classified as long or flexible pile. Most inter-
national design codes present closed form formulas to estimate the
Deep foundations have been widely used to support heavy ultimate lateral load for both types of piles in uniform soil profile.
structures especially in challenging zones with surface layers of They classified piles based on relative stiffness between pile and
weak soils. In most practical cases, piles are subjected to both ver- surrounding soil mass using a term called ‘‘Elastic length of pile”.
tical and lateral loading induced from different cases of loading. Most famous used formulas are ‘‘Hetenyi, 1946” for flexible piles,
Due to the complicated behavior of piled foundation under lateral ‘‘Duncan, 1995” for semi-rigid piles and ‘‘Barber, 1953” for rigid
loads especially in layered soil profiles, most design codes recom- piles [1].
mended field loading tests on non-working piles in project location
to validate the pile design before construction, besides field loading 2. Background
tests on working piles during construction.
Laterally loaded pile fail either by structural failure within pile Walsh 2005 investigated the lateral load response of the single
or by geotechnical failure in the surrounding soil mass, besides the pile in clay soil and compared this responses with the same pile
serviceability failure due to excessive lateral displacement. If the but in a group of fifteen piles acts as a group to resist the lateral
structural capacity of the pile is larger than the geotechnical one, load. Then, the tested single pile was modeled using LPILE Plus
the failure will occur in soil mass and controlled by its properties software and the group of piles was modeled using the GROUP
and the pile is classified as short or rigid pile. On the other hand, software to validate the finite element model against the experi-
if the geotechnical capacity is larger than the structural one, the mental results. Researchers validated the developed models
failure will occur within the pile and controlled by its flexural against experimentally tested piles concluding that the single pile
lateral resistance is larger than the average of the maximum load
carried by a single pile in the pile group. Also, the pile group in
⇑ Corresponding author.
sand resisted much higher loads than did the pile group in clay.
E-mail address: ahmed.abdelkhaleq@fue.edu.eg (A.M. Ebid).
However, the used software models the laterally loaded piles as
Peer review under responsibility of Ain Shams University.
line element and the soil as non-linear springs which neglect the
effect of concrete cracking and inaccuracy simulates the soil-pile
interaction [2].
Pender M. 2008, investigated the effect of the distribution of the
Production and hosting by Elsevier
soil profile modulus of elasticity on the pile head lateral stiffness.

https://doi.org/10.1016/j.asej.2022.101734
2090-4479/Ó 2022 THE AUTHORS. Published by Elsevier BV on behalf of Faculty of Engineering, Ain Shams University.
This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
D.M. Ors, A.M. Ebid and H.A. Mahdi Ain Shams Engineering Journal 13 (2022) 101734

Nomenclature
L Pile length
Abbreviations list: M External moment at pile head
c Cohesion strength of soil T Elastic length of pile
D Pile diameter Dmax Maximum lateral displacement of pile
Ec Elastic modulus of concrete a Friction angle between soil and pile
Es Elastic modulus of soil ecr Cracking strain of concrete
Est Elastic modulus of steel et Tensile strain of concrete
Ec I g Un-cracked flexural stiffness of pile est Strain of reinforcing steel
f The coefficient of variation of lateral subgrade reaction ey Yield strain of reinforcing steel
with depth g Ratio between concrete strain and the concrete cracking
f ctm Maximum tensile strength of concrete strain
f st Stress in reinforcing steel mc Poison’s ratio of concrete
fy Yield stress of reinforcing steel ms Poison’s ratio of soil
H Horizontal load at pile head U Internal friction angle of soil
Ie The effective moment of inertia of the pile Ws Dilation angle of soil
Ig The gross moment of inertia of the pile

The researcher calculated the pile initial lateral resistance three On the other hand, Ebid 2015, presented a simplified approach
times varying the distribution of the soil modulus through pile to consider cracking effect on the behavior of laterally loaded con-
depth as constant, linear and parabolic distribution. The results crete piles, he carried out a parametric study using a simplified lin-
proved that the modeling of soil with constant modulus requires ear and nonlinear FEM model for laterally loaded piles, and he
more advanced tests than the Standard Penetration Test and it suggested to use effective flexure stiffness (EsIe) of about 75% of
was recommended to use the work of Ashford et al 2001 who used the gross stiffness (EsIg) to account for the effect of concrete crack-
the shear wave velocity profile to get an improved indication of soil ing. Although he considered a nonlinear behavior for concrete pile
geotechnical properties including the soil stiffness through the to account for cracking effect, but the used linear springs to simu-
profile. However, he used a closed form equations of beam on elas- late soil mass which reduces the accuracy of his conclusions [1].
tic support which neglected the non-linear behavior of soil and the Aljanabi et al. 2020, tested a set of driven short piles under the
crack effect of concrete [3]. effect of the lateral load in clayey soil. The researchers investigated
Comodromos et al. 2009; developed a finite element model for a the effect of various parameters such as the degree of saturation,
group of piles in the aim to determination of their bearing capacity. effective unit weight, pile diameter and the embedded length.
The researchers used FLAC-3D software to simulate the lateral load The researchers found that the piles lateral bearing capacity
response of the proposed group of piles. It was found that the increased with the increase of the pile diameter and more affected
model is sensitive against the mesh size and the model overesti- by the pile embedded length. But the increase of the soil degree of
mates the bending stiffness of the simulated piles which recom- saturation decreases the pile lateral capacity. However, they
mends the need for a 3D nonlinear detailed model allowing for depended only on tiny scale laboratory tests without considering
cracking to get the accurate response when compared with the the size effect on soil properties especially the angle of internal
experimental results. Although the researchers developed a friction [7].
sophisticated elasto-plastic 3D FEM model with nonlinear soil- Keawsawasvong et al. 2020, employed a 3D finite element
pile interface elements and considered the concrete cracking effect, model taking the interface into consideration with no tension con-
But they only compared the model results with the field test dition to investigate the effect of various parameters on the lateral
results and didn’t present any comparisons with design code or response of the single pile in the undrained clay soil subjected to
earlier studies, besides that their proposed model is too compli- combined horizontal loading and moment. These parameters
cated to be used in practice [4]. included the pile length and overburden stress factors. The model
El Wakil and Nazir 2013, studied the behavior of laterally was capable to predict that the lateral capacity of the modeled
loaded barrettes in sandy layer using 28 small scale experimental piles and the overburden stress factor significantly affect the fail-
tests. They investigated the effect of different parameters on the ure mechanism in case of undrained clay as the overburden pres-
lateral displacement of pile head such as the aspect ratio of bar- sure increases; the elliptical rotated failure envelope expands its
rette section, the relative density of surrounding sand direction size. However, the authors needed experimental tests to support
and eccentricity of lateral load. The research indicated that both their conclusions which were based on the results of the finite ele-
relative density of sand and effective flexural stiffness of barrette ment model only [8].
(presented by cross section and loading direction) are the main fac- Shirgir et al. 2021, presented an analytical model for a pile in
tors affected the lateral displacement. However, the measured uniform and non-uniform soil profile under dynamic lateral loads.
results were greatly affected by scale factor, hence, they recom- For the uniform soil profile both of soil density and shear wave
mended a full scale experimental program [5]. velocity were constant with depth, on other hand, they had a para-
Farag 2014, presented a probabilistic approach based on (3D FEM- bolic distribution with depth in the non-uniform soil profile. The
Monte Carlo Simulation) technique to assess the reliability of later- developed analytical model was verified using two FEM models,
ally loaded pile in bi-layered soil profile in liquefaction condition. the first was simplified model where the soil was simulated as
He considered both pile and soil layers parameters as inputs with spring-dashpot element, while the second one was a full 3D model
assumed statistical characteristics based on engineering judgment, using solid elements for both soil mass and pile. The results of the
while the outputs were the drift and the bending moment in pile. It three models are close, however these results need to be verified
was concluded that pile diameter is the most effective variable [6]. experimentally [9].

2
D.M. Ors, A.M. Ebid and H.A. Mahdi Ain Shams Engineering Journal 13 (2022) 101734

Mahdi et al 2022, presented three (AI) models to predict both large diameter piles by drawing a virtual pile load-settlement
maximum drift and bending moment in a free head flexible pile curve and determine the allowable load using the allowable settle-
embedded in multi-layered soil profile. The utilized (AI) techniques ment. Each piles had a different embedded length that led to the
were (ANN), (GP) and (EPR). Although (ANN) model showed the difference in the pile depth/diameter ratio (L/D). Table 1 summa-
most accurate predictions, but (EPR) was the most practical model. rizes the piles geometry configurations and details of reinforce-
The research indicated that the both flexural stiffness of pile and ment. All tested piles were cast using same concrete grade
lateral subgrade reaction of top layer govern the behavior of the (C28/35) corresponding to cube characteristic compressive
pile [10]. strength 35 MPa. The steel of grade (ST50/60) was used for rein-
forcement cages including both main and transverse reinforce-
3. Objective ment. Steel bars diameters were 16 mm for stirrups and 32 mm
for main reinforcement. Tested piles were board through different
The previous literature review showed the need to more accu- soil profiles with different pile depth that ranged from 16 m to
rate estimation for the lateral subgrade reaction of soil considering 22 m. Borehole adjacent to each pile were investigated and shown
the nonlinear behavior of the soil and the cracking effect of con- in Fig. 9-c, Fig. 10-c, Fig. 11-c & Fig. 12-c.
crete using simple and practical formula. Hence, the aim of this Incremental Lateral load was applied at each pile head till 200%
research is to study the behavior of laterally loaded piles in of the design working load that is equivalent to 170 tons. Load set-
multi-layered soil profile using 3D FEM models where the soil pro- up is shown in Fig. 2. The lateral displacement at the pile head and
prieties are calibrated to match the experimental behavior of full in the same level of the applied lateral load was measured at all
scale pile load tests. The actual values of coefficient of variation load increments till the end of test. Fig. 9-b, Fig. 10-c, Fig. 11-c &
of lateral subgrade reaction with depth (f) were extracted from Fig. 12-c illustrate the measured load–displacement response of
the field tests and correlated to the calibrated elastic modulus of all tested piles. The experimental behavior approved that the
soil. Finally, the recommended values for (f) in design codes ware design load was safe and economic regarding the pile diameter,
evaluated in the light of the measured ones. Fig. 1 presents flow- embedded length, and reinforcement in the proposed soil profiles.
chart for the used methodology.
5. Modeling strategy
4. Reference experimental investigation
ABAQUS software is general finite element program that can be
Four non-working board piles from Cairo Monorail Project in used in a wide range of analyses extend to a nonlinear complex
Egypt were considered in this study. These tests were carried out analysis type involving contact, plasticity, and large deformations
between July 2020 and January 2021. All of them have the same as in the models simulated through this study. This section shows
longitudinal and transverse reinforcement ratio were constructed the sequence of the model construction to accurately simulate the
at different locations where the soil profile varies. The piles were lateral response of the reinforced concrete piles and the geotechni-
designed considering the variation of soil profile as per Navfac cal behavior of the surrounding soil layers. The proposed rein-
DM-7.02 [11] which is a well kwon design manual from naval facil- forced concrete pile was modeled by half cylinder and the
ities engineering command, although it was issued in 1986 but it surrounding soil mass was modeled as half cuboids. Fig. 3 shows
still one of the main references in most international design codes that the model was simulated as the positive X-Y axes in plane
such as Uniform building Code (UBC), International Building Code and the negative Y-axis as the vertical direction. The modeled half
(IBC), American Association of State Highway and Transportation soil mass had dimensions of 15 m
25 m in plane and
Officials (AASHTO) and many others. It is recommended to design (Lpile + 20D) for the width, length, and depth respectively. These
dimensions of the soil mass were chosen to eliminate the effect
of the minimal FEM dimensions when compared with the real
cases and to guarantee the pile stability against the model size
effects. Fine mesh size was used near in pile part and soil mass
zones surrounding the piles as shown in Fig. 3. Model setup steps
are listed in Table 2.

5.1. Material model

5.1.1. Soil mass

For the simulation of the soil mass was modeled as, built-in
first-order reduced integration 8-node brick element C3D8R was
used to represent the concrete in the finite element model [12].
The soil parameters used in the developed finite element model
are selected within the recommended ranges mentioned in the
Geotechnical Interpret Report for each borehole to fit the experi-
mental response. The soil elastic behavior was modeled as linear
elastic behavior in terms of elastic modulus (Es) and the Poisson’s
ratio (ts).
ABAQUS/CAE has a special built in Mohr-Coulomb plasticity
model that could be used for soil modeling. The classical Mohr-
Coulomb yield criterion has a straight line in the meridional plane
and an irregular hexagonal section in the deviatoric plane [13] as
shown in Fig. 4. However, the ABAQUS provides his Mohr-
Coulomb model with completely smooth flow potential instead
Fig. 1. Flowchart for the used methodology. of the classical hexagonal pyramid, the flow potential is a hyper-
3
D.M. Ors, A.M. Ebid and H.A. Mahdi Ain Shams Engineering Journal 13 (2022) 101734

Table 1
Piles geometry and Reinforcement.

Pile Pile Length (L) Diameter (D) Aspect Ratio RFT

ID (m) (m)
L/D Longitudinal Transverse
6OC00 16.00 2.00 8.00 52 T 32 20 T 16 /m
NCC17 18.40 2.00 9.20
NCC19 18.65 2.00 9.33
NCC22 22.00 2.00 11.00

Fig. 2. Loading system layout.

Fig. 3. Details of the finite element model.

4
D.M. Ors, A.M. Ebid and H.A. Mahdi Ain Shams Engineering Journal 13 (2022) 101734

Table 2
Model setup steps.

Element type Soil C3D8R See Section 3.1.1

Concrete C3D8R See Section 3.1.2
RFT T3D2 See Section 3.1.3
Material models Soil Mohr-columnb See Section 3.1.1
Concrete Drucker-Pruger See Section 3.1.2
RFT Perfect elastic–plastic See Section 3.1.3
Section assignment Soil Solid-homogeneous
Concrete Solid-homogeneous
RFT Truss
Assembly Assemble all parts as the actual pile test (embedded RFT in the pile bored See Fig. 3
in the soil mass)
Interaction Soil layers Tie constraint See Section 3.2.1
Pile-soil Interaction (Normal/Tangential) See Section 3.2.2
Concrete-RFT Embedded host See Section 3.2.3
Boundary conditions Symmetry Z-SYMM (U3 = UR1 = UR2 = 0)
Bottom Hinges (U1 = U2 = U3 = 0)
Loading steps Geostatic Model developed and validated by (Wang X. et al.,2021) is used for loading
Gravity load steps simulation:
Lateral Load 1-Deactivate the contact interaction (or set the friction coefficient to zero) at
the first step of the simulation and use the automatic method to run the
geostatic stress field procedure for each part of the model.
2-Activate the contact interaction (or restore the friction coefficient) at the
second step and run a static analysis (lateral loading on pile upper surface).

bola in the meridional plane, and it uses the smooth deviatoric sec- crack concrete model, brittle crack concrete model, and concrete
tion proposed by Menétrey and Willam. Abaqus Mohr-Coulomb damaged plasticity (CDP) model. The concrete damaged plasticity
model is shown in Fig. 5 [13]. (CDP) model was selected to model the plastic behavior of concrete
The soil plasticity behavior was defined using the Mohr- in the present study as this technique is capable of representing
Coulomb yield criterion in terms of the soil friction angle (U), dila- the complete inelastic behavior of concrete in both compression
tion angle (W) and cohesion (C). [13] Table 3 summarizes the final and tension [15]. CDP model requires definition of the uniaxial ten-
geotechnical input values for each soil layer after calibration sile and compressive behavior as well as a set of five parameters to
against the experimental test results. completely describe the plastic behavior of concrete. The plasticity
parameters are the flow potential eccentricity, the ratio of initial
5.1.2. Concrete biaxial compressive yield stress to initial uniaxial compressive
For the simulation of the reinforced concrete piles, built-in first- yield stress, and the dilation angle in degrees. These parameters
order reduced integration 8-node brick element C3D8R was used were set as the default values recommended by ABAQUS as 0.1,
to represent the concrete in the finite element model. The concrete 1.16, 0.66, 0, and 30° respectively [13 16].
material model was defined as elastic behavior followed by the Eq. (2) represents the nonlinear uniaxial compressive behavior
plastic part as shown in Fig. 6. Elastic behavior of concrete was of the unconfined concrete provided by the EN-1992. As the mod-
modeled as linear and isotropic in terms of Elastic modulus (Ec) eled piles had the same compressive strength; the constants in Eq.
and concrete Poisson’s ratio (mc) of 0.20. Elastic modulus of con- (2) were the same in all models as follows; concrete strain at peak
crete could be estimated according to EN-1992 using Eq. (1) where stress (ec1 ) equals to 2.15‰ and the equation constant (k) of 2.112.
fcm is the cylinder characteristic concrete strength. In this research,
(Ec) values were selected between 30 and 35 GPa to fit the exper-
kg  g2
imental test results [14]. rc ¼ f cm : ð2Þ
1 þ ðk  2Þg
0:30
Ec ¼ 22ð0:1f cm Þ ð1Þ
Where; g ¼ ec =ec1 , ec1 ðstrainatpeakstressÞ ¼ 0:7ðf cm Þ0:31 and
The plastic portion of the concrete material model can be sim- k ¼ 1:05Ecm :jec1 j=f cm
ulated in ABAQUS software using one of the three crack models
for conventional reinforced concrete elements which are; smeared

Fig. 4. Pyramidal surface in principal stress space and cross-section in the equi- Fig. 5. Menétrey and Willam flow potential in the deviatoric stress plane [13].
pressure plane [19].

5
D.M. Ors, A.M. Ebid and H.A. Mahdi Ain Shams Engineering Journal 13 (2022) 101734

Table 3
Model Input Geotechnical Properties of Soil layers.

Pile ID Soil Layer Depth c Elastic Plastic Interface

(m) (t/m2)
Es t c U Ws a Coeff. of friction
(t/m2) (t/m2) (o) ( o) ( o) tan(a)
6OC00 Fill (Double sleeve) 1.75 1.7 1500 0.33 0 27 0 18 0.000
layer 1-Cemented Sand 12.5 2.2 40 000 0.20 5 38 4 25 0.473
layer 2-Weak Sandstone 10.0 2.0 200 000 0.20 60 38 4 25 0.473
layer 3-Cemented Sand 12.5 2.2 40 000 0.20 5 38 4 25 0.473
NCC17 Fill (Double sleeve) 3.0 1.7 1500 0.33 0 27 0 18 0.000
layer 1-Intact Limestone 3.0 2.4 3 000 000 0.15 50 35 4 25 0.431
layer 2-Hard Clay 3.0 2.2 20 000 0.25 30 30 2 20 0.364
layer 3-Dense Sand 8.0 2.0 7000 0.30 5 38 4 25 0.473
layer 4-Intact Limestone 21.0 2.4 2 000 000 0.15 50 20 0 13 0.237
NCC19 Fill (Double sleeve) 1.75 1.7 1500 0.33 0 27 0 18 0.000
layer 1-Silty Sand 11.3 2.0 30 000 0.20 15 35 4 23 0.431
layer 2-Cemented Sand 10.0 2.2 60 000 0.20 20 35 4 23 0.431
layer 3-Stiff Clay 15.0 2.1 5500 0.30 30 20 2 13 0.237
NCC22 Fill (Double sleeve) 1.75 1.7 1500 0.33 0 27 0 18 0.000
layer 1-Crushed Stone 9.0 2.2 100 000 0.20 10 35 4 23 0.431
layer 2-Hard Clay 6.5 2.2 20 000 0.25 30 20 2 13 0.237
layer 3-Crushed Stone 23.5 2.2 200 000 0.20 15 35 4 23 0.431

Fig. 7. Steel Stress-Strain Relations Interaction.

was defined assuming no hardening especially in case of mono-

Fig. 6. Uniaxial stress–strain curve of concrete (Xiao, Chen, Zhou, Leng, & Xia,
tonic loading as in this study as shown in Fig. 7. Eqs. (4) and (5)
2017).
show the defined behavior of steel reinforcement in all models.
f st ¼ Est  est est 6 ey ð4Þ
The tensile strength of the concrete f ctm was calculated as a
function of the concrete compressive strength, thus the tensile f st ¼ f y est P ey ð5Þ
strength of concrete in all models was 2.80 MPa according to EN-
1992. [14] The non-linear post-peak portion of the tensile behavior
5.2. Interaction model
of concrete was modeled according to the modified Wang and Hsu
formula shown in Eq. (3). The equation presents the softening of
5.2.1. Interface between soil layers
concrete after the maximum tensile strength is reached; the power
The interaction between soil layers was considered as tie con-
(n) represents the rate of weakening that ranges from 0.50 to 1.50.
straint with no relative displacements at the interface. Thus, the
[16] Concrete cracking strain ecr was calculated as the ratio
slip between soil layers was ignored and considered to have the
between concrete tensile strength and the concrete elastic modu-
same translational and rotational degrees of freedom during all
lus and equivalent to 0.09‰.
loading steps.

n
ecr
rt ¼ f ctm if et > ecr ð3Þ
et 5.2.2. Interface between soil and pile
Concrete pile and surrounding soil mass are not perfectly
bonded during all loading stages; so, the interaction at the inter-
5.1.3. Steel reinforcement face between soil and concrete pile along pile length and at bottom
Reinforcement bars were modeled as discrete truss element in surface was defined using the default constraint enforcement
space T3D2. Steel is a homogeneous isotropic material that can method. Bearing between concrete elements and soil at the inter-
be accurately defined as perfect elastic–plastic in both tension face was defined as the non-tension behavior by hard contact pres-
and compression. [16 17] Steel used in all piles for both main rein- sure over-closure having finite sliding with the surface to surface
forcement and stirrups is steel Grade (ST50/60). Steel elastic as the discretization method. The Penalty method with coefficient
behavior was modeled as linear isotropic with 210 GPa elastic of friction was chosen to simulate the tangential contact between
modulus and Poisson’s ratio (mst) of 0.30. The plastic behavior concrete elements [18].
6
D.M. Ors, A.M. Ebid and H.A. Mahdi Ain Shams Engineering Journal 13 (2022) 101734

The coefficient of friction between soil layers and concrete pile increased gradually in each step increment till failure occurs in soil
was calculated according to the geotechnical properties of each soil or concrete pile which depends on the pile diameter and length as
layer in terms of the soil angle of friction developed at soil-pile well as the surrounding soil properties.
interface (a) that is equivalent to two thirds the soil internal angle
of friction (U). The defined coefficient of friction in all modeled pile
5.5. Validation of the finite element model
in this study is calculated from Eq. (6) [19]. Table 3 summarizes the
coefficient of friction used between pile and each soil layer.
Based on the experimentally monitored lateral load–displace-
Coeff :offriction ¼ tanðaÞ ð6Þ ment behavior, the developed finite element model is validated
in terms of load–displacement response, failure mode, expected
stresses and strains at failure. The load–displacement behavior of
5.2.3. Interface between concrete and reinforcement bars both FEM and the experimentally monitored behavior are compa-
The reinforcement truss elements were embedded in host con- rable. As the lateral displacement corresponding to the maximum
tinuum solid elements as shown in Fig. 3. Embedding means that applied load in the experimental tests was captured with error ran-
the translational degrees of freedom at the nodes of the embedded ged from (1.09 % to 9.75%) as shown in Table 4. The developed
element are eliminated and become constrained to the corre- modeling approach proved the design parameters and insured
sponding interpolated values in the host continuum element. the reliability of the developed FEM which has been successfully
utilized for assessing the safety and lateral capacity of pile.
5.3. Boundary conditions To investigate the failure criteria of the proposed set of the rein-
forced concrete piles, the previously developed and validated finite
The model of the laterally loaded piles is set by using the com- element model was submitted till failure. The FEM was capable to
mon pattern for the finite element models in geotechnical engi- capture the failure mode and the load/displacement at failure was
neering as the transitional degrees of freedom (ux, uy, uz) were concluded. Modes of failure are presented in terms of maximum
all constrained for all the nodes at the bottom plane of the modeled stresses and strains carried by the concrete pile and the main rein-
soil mass. Also, horizontal displacements were restrained for all forcement as well as the surrounding soil. All results at failure are
the vertical side planes. All rotational and transitional degrees of shown in Figs. 9, 10, 11 and 12 on the deformed shape of each ele-
freedom were permitted for the top plane as it was treated as ment at failure. All piles showed plastic strains that increased grad-
the free surface in the model where the load will be applied. Con- ually till the maximum tensile strength of concrete (280 MPa)
sidering finite element model symmetry about X-Y plane; a sym- which is clear in all failure modes, then the stresses were released
metry boundary condition in ABAQUS (Z-SYMM) was used to
reduce the analysis time. This symmetry boundary condition
restrains the transitional degree of freedom (uz) and the rotational Table 4
degree of freedom (uRX, uRY). Comparison between experimental and the FEM.

Pile Lateral displacement at 200% %error

5.4. Loading protocol ID of working load
Experimental FEM DFEM Dexp:
Dexp: :100%
Usually, the first step in modeling is to define the geostatic step (mm) (mm)
that includes the initial stresses and gravity loads [12]. Fig. 8 shows 6OC00 5.20 5.10 1.96%
the results of the first step of the geostatic loading. Then the main NCC17 2.25 2.30 +2.22%
loading step was defined to apply the static lateral loading as a sur- NCC19 7.28 7.20 1.09%
NCC22 4.00 4.39 9.75 %
face traction shear pressure to the top pile surface. The load was

Fig. 8. Geostatic step results of the FEM model.

7
D.M. Ors, A.M. Ebid and H.A. Mahdi Ain Shams Engineering Journal 13 (2022) 101734

Fig. 9. Pile 6OC00.

and transferred to the steel reinforcement which showed maxi- the maximum tensile strength of 50000 t/m2 at failure. As
mum strains at the zone of maximum curvature. Main reinforce- expected, the soil mass was affected by the lateral loading at the
ment yielded in tension as stresses in all modeled piled exceeded most upper layers only which experienced high stresses and

8
D.M. Ors, A.M. Ebid and H.A. Mahdi Ain Shams Engineering Journal 13 (2022) 101734

Fig. 10. Pile NCC17.

strains as shown in Figs. 9, 10, 11 and 12. The maximum principal soils and (9C) for cohesive soils, accordingly, the depth of soil lay-
stresses at failure were about 100 t/m2 for sand layers, 150 t/m2 for ers that showed plastic behavior was about 2.5 m in sand, 2.0 m in
crushed stone and 300 t/m2 for intact limestone. As per Barber, crushed stone and 1.0 m in intact limestone. These calculated
1953, the ultimate lateral soil capacity is (3 c h Kp) for granular depths matches the showed results in Figs. 9, 10, 11 and 12.

9
D.M. Ors, A.M. Ebid and H.A. Mahdi Ain Shams Engineering Journal 13 (2022) 101734

Fig. 11. Pile NCC19.

6. Discussion lateral displacement and elastic length of a free head long pile
respectively.
Previous analysis shows that the four tested piles failed due to
2:4H:T 3 þ 1:55M:T 2
structural failure and accordingly, they are long or flexible piles. Dmax ¼ ð7Þ
Navfac DM-7.02 recommended Eq. (7), 8 to calculate the maximum Ec :Ig

10
D.M. Ors, A.M. Ebid and H.A. Mahdi Ain Shams Engineering Journal 13 (2022) 101734

Fig. 12. Pile NCC22.

sffiffiffiffiffiffiffiffiffiffi
5 Ec :I g
(C28/35), the pile flexural stiffness (Ec :Ig ) is 2.4E + 6 (t.m2). The elastic
T¼ ð8Þ length of pile for each test could be calculated using Eq.7 using hori-
f
zontal load (H) of 170 ton, external moment (M) of (1.75 m
170
Where M, H, T, and f are the external moment, horizontal load, elas- ton) and hence the (f) values using Eq. (8) as shown in Table 5.
tic length of pile and the coefficient of variation of lateral subgrade Comparing the recommended values of (f) from Fig. 13 with the
reaction with depth respectively. For 2 m diameter pile with concrete experimental values as showed in Table 5 indicates that experi-

11
D.M. Ors, A.M. Ebid and H.A. Mahdi Ain Shams Engineering Journal 13 (2022) 101734

Table 5
Comparison between experimental and design (f) values.

Pile Exp. Dmax T f exp. f Navfac Es upper Description of upper layer

ID (mm) (m) (t/m3) (t/m3) (t/m2)
6OC00 5.20 2.80 14 000 1500 40 000 Dense Sand, N30 = 38
NCC17 2.25 2.05 67 500 2200* 3 000 000 Intact Limestone
NCC19 7.28 3.15 7 500 1350 30 000 Dense Sand, N30 = 35
NCC22 4.00 2.55 23 000 1900 100 000 Crushed Limestone
*
2200 t/m3 is the upper limit of (n) values for granular soils as per Navfac DM-7.02 [11].

effect of concrete, the yielding in reinforcement bars, the non-

tension, rough interface between pile and soil and finally the
nonlinear response of surrounding soil mass.
2. The results of the developed FEM models illustrated that the
behavior of long or flexible laterally loaded piles is controlled
by the properties of the upper soil layer down to a depth equals
to the elastic length of the pile (T). Below that depth, the soil
properties have minor effect of the pile behavior.
3. Comparing the measured and design values of the coefficient of
variation of lateral subgrade reaction with depth (f) showed
that measured values are about (6 to 12) times the recom-
mended values in design codes. Also, it indicated that measured
(f) values for intact rock are about 30 times the recommended
upper boundary of (f) for granular soils.
4. The developed correlation between the coefficient of variation
of lateral subgrade reaction with depth (f) and the correspond-
ing elastic modulus of soil showed a perfect logarithmic
relation.

8. Present study limitations and future research work

Fig. 13. Recommended values for (f) as per Navfac DM-7.02 [8].
This research focuses on the simulation of the lateral response
of the piles with aspect ratio (L/D) ranges from 9 to 11. Thus, fur-
ther studies should be conducted using validated methodology to
study the behavior of vertically loaded piles in multi-layered soil
profile with different aspect ratios.

Declaration of Competing Interest

The authors declare that they have no known competing finan-

cial interests or personal relationships that could have appeared
to influence the work reported in this paper.

References

[1] Ebid AM. ‘‘Simplified Approach to Consider Cracking Effect on the Behavior of
Laterally Loaded RC Piles. Int J Innov Res Sci Eng Technol (IJIRSET) 2015;4
(10):9589–98.
[2] Walsh JM. Full-Scale Lateral Load Test of a 3x5 Pile Group in Sand. Provo,
Fig. 14. Relation between Elastic modulus (Es) and experimental (f) value. Utah: Brigham Young University; 2005.
[3] Pender M. Effect of soil profile modulus distribution on pile head lateral
mental values are (6 to 12) times the design ones excluding test stiffness. In: 18th NZGS Geotechnical Symposium on Soil-Structure
Interaction., Ed. CY Chin, Auckland, 2008.
NCC17 where the design value of (f) is considered as the upper [4] Comodromos EM, Papadopoulou MC, Rentzeperis IK. Effect of Cracking on the
limit of granulate soil while the actual upper layer was intact Lime- Response of Pile Test under Horizontal Loading. J Geotech Geoenviron Eng
stone. Fig. 14 illustrates the relation between the selected Elastic 2009:1275–84.
[5] El Wakil AZ, Nazir AK, Wakil El. Behavior of laterally loaded small scale
modulus of upper layer from FEM models and the corresponding
barrettes in sand. Ain Shams Eng J 2013;4(3):343–50.
experimental (f) values. Eq. (9) presents the best fitting curve for [6] Farag R. Probabilistic pseudostatic analysis of pile in laterally spreading
this relation. ground: two layer soil profile. Ain Shams Eng J 2014;5(2):343–54.
[7] Aljanabi HA, Bazaz JB. Laboratory Study of Lateral Load Capacity and Behaviour
fexp: ¼ 12750lnðEsÞ  123000 ð9Þ of Short Piles in clayey soil. IOP Conf. Ser.: Mater. Sci. Eng. 2020;671
(1):012120. doi: https://doi.org/10.1088/1757-899X/671/1/012120.
[8] Ukritchon SKB. Failure Modes of Laterally Loaded Piles Under Combined
Horizontal Load and Moment Considering Overburden Stress Factors. Geotech
7. Conclusions Geol Eng 2020:4253–67.
[9] Shirgir V, Ghanbari A, Massumi A, Shirgir Vahid. Soil-Pile-Structure Interaction
1. The developed nonlinear FEM models were successfully cap- Effects in Alluvium with Non-constant Shear Modulus in Depth. Transp
Infrastructure Geotechnol 2021;8(2):254–78.
tured the behavior of the laterally loaded piles in multi-
[10] Mahdi HA, Ebid AM, Onyelowe KC, Nwobia LI. Predicting the behaviour of
layered soil profile. They were capable to simulate the cracking laterally loaded flexible free head pile in layered soil using different AI (EPR,

12
D.M. Ors, A.M. Ebid and H.A. Mahdi Ain Shams Engineering Journal 13 (2022) 101734

ANN and GP) techniques. In: Multiscale and Multidisciplinary Modeling, Ahmed M. Ebid Associated Professor, Future University,
Experiments and Design, 2022. Cairo, Egypt
[11] Structures, NAVFAC DM7-02 Foundations and Earth; Design Manula 7.02,
1986.
[12] Wang X-Y, Yuan Y-L, Hu C-M, Mei Y. Research on the Geostatic Stress Field
Procedure under complex Conditions. Adv Civ Eng 2021;2021:1–15.
[13] ABAQUS, ABAQUS analysis user’s manual, Version 6.10, Dassault Systèmes,
2020.
[14] Eurocode2, Design of concrete structures. Part 1-1: general rules and rules for
buildings, Brussels, 2004.
[15] Xiao Y, Chen Z, Zhou J, Leng Y, Xia R. Concrete plastic-damage factor for finite
element analysis: Concept, simulation, and experiment. Adv. Mech. Eng.
2017:1–10.
[16] Fathi DM, Okail HO, Mahdi HA, Abdelrahman AA. Cyclic load behavior of
precast self-centering hammar head bridge piers. Housing Building Res. Center
J. (HBRC) 2020:113–41.
[17] Canakci H. Pile foundation design using Microsoft Excel. Comput. Appl. Eng. Hisham Arafat Professor, Future University, Cairo,
Educ. 2007;15(4):355–66. Egypt
[18] Ors DMF, Okail HO, Zaher AH. Modeling of shear deficient beams by the mixed
smeared/discrete cracking approach. Housing Building National Res. Center
(HBRC) J. 2016;12(2):123–36.
[19] Labuz JF, Zang A. Mohr–Coulomb Failure Criterion. Rock Mech. Rock Eng.
2012;45(6):975–9.

Dina M. Ors Assistant Professor, Future University,

Cairo, Egypt.

13