International Journal of Geotechnical Engineering

ISSN: 1938-6362 (Print) 1939-7879 (Online) Journal homepage: https://www.tandfonline.com/loi/yjge20

Simplified approach for soil-spring stiffness

prediction of pile group

Lassaad Hazzar, Mourad Karray & Admir Pasic

To cite this article: Lassaad Hazzar, Mourad Karray & Admir Pasic (2019): Simplified approach for
soil-spring stiffness prediction of pile group, International Journal of Geotechnical Engineering, DOI:
10.1080/19386362.2019.1612576

To link to this article: https://doi.org/10.1080/19386362.2019.1612576

Published online: 05 May 2019.

Submit your article to this journal

Article views: 7

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=yjge20
INTERNATIONAL JOURNAL OF GEOTECHNICAL ENGINEERING
https://doi.org/10.1080/19386362.2019.1612576

Simplified approach for soil-spring stiffness prediction of pile group

Lassaad Hazzara, Mourad Karrayb and Admir Pasicc
a
Expertise Division, Hydro-Quebec, Montreal, Quebec, Canada; bDepartment of Civil Engineering, Faculty of Engineering, Sherbrooke University,
Sherbrooke, Quebec, Canada; cBridges and Structures Department, Stantec Inc. Bridge and Structures, Montreal, Quebec, Canada

ABSTRACT ARTICLE HISTORY

For deep foundations, several approaches have been developed to represent the effect of soil by using Received 26 September 2018
a spring with given stiffness. Typical stiffness values have been proposed for different types and Accepted 24 April 2019
densities of soil. These values generally ignore the effect of the depth and particularly the degradation KEYWORDS
of soil rigidity with strain. In this paper, a series of 3D numerical analyses are conducted to compute the Pile group; soil–spring;
equivalent soil–spring stiffness as a function of pile distortion for a group of pile located under the stiffness; multi-loading;
central pier of bridge structure above the river Sault–au–Mouton (Longue–Rive, Quebec) and subjected lateral capacity; p–y curve;
to multi–loading conditions. In these simulations, the degradation of the shear modulus is incorporated distortion; 3D modelling
in order to account for soil non-linearity. This equivalent stiffness of the soil-springs varies depending
on the distortion and may be adapted in an iterative process according to the pile deformation at each
depth. The procedure can be compared to the linear equivalent method used in dynamic analysis
where an equivalent shear modulus is adapted according to the shear strain. In fact, this type of linear
equivalent spring can very useful for structural engineers who want to incorporate the effect of the
lateral capacity of soil in their models.

Introduction predict the load–displacement relationship for piles under

lateral loads. Kumar, Lalvani, and Omar (2006) proposed
The behaviour of laterally loaded piles is a complex soil-
K_hMax approach, which is defined as a low strain modulus
structure interaction problem that has received a considerable
of horizontal subgrade reaction. Chin, Sew, and Chung (2010)
amount of attention over the last four decades mainly in the
presented the results and interpretation of a lateral load test
high stakes field of offshore engineering and earthquake geo-
on a fully instrumented spun pile in soft ground for the land
technical engineering. The primary function of a single pile or
viaduct section of a high-speed train project. The test results
a pile group is to transfer the external loads from the super-
were further analyzed to determine the soil-spring stiffness
structure to the surrounding soil medium without causing
profile along the pile depth. Generally, interpreted subgrade
excessive deflections. In practice, analysis of laterally loaded
reaction profile had shown a peak at about 5–6-m depth for
piles is carried out using beams on non-linear Winkler springs
all loading except for initial loading up to 20kN. This was due
model (often known as p–y method) due to its simplicity in
to the horizontal soil pressures not fully developed.
modelling layered soils. In this approach, soil–pile interaction
Until now, the contributions from the latest experimental
is characterized by p–y curves where p is the pressure on the
tests have been very limited and the most current codes still
soil that causes a relative pile deformation of y. A number of
propose approaches such as Broms’s method for the design of
methods used to construct the p–y curves (e.g. Matlock 1970;
piles under lateral loads (CGS 2013). This empirical method,
Cox, Reese, and Grubbs 1974; Reese, Cox, and Kooper 1974;
criticized in previous numerical studies (Hazzar et al. 2013), is
Reese and Van Impe 2001). Subsequently, many approaches
not entirely appropriate for the design of complex structures
(e.g. Terzaghi 1955; Vesic 1961; Broms 1964a, 1964b; Ménard,
such as bridges, which can only be realistically modelled by
Bourdon, and Gambin 1971; Davisson 1970; Matlock 1970;
sophisticated, three-dimensional (3D) numerical analysis. In
Poulos 1971; Gilbert 1995) have been developed to determine
this 3D modelling, the soil behaviour should be simulated by
the soil–spring stiffness, and typical values have been proposed
a series of springs surrounding the pile elements. Therefore,
for different types of soil and densities. However, the most
several questions were frequently asked by structural engi-
typical values (except Terzaghi 1955) generally do not explicitly
neers, about the appropriate values of the soil–spring stiffness
consider the effect of depth and the degradation of rigidity
to simulate the soil around the piles during static and seismic
parameters (G or E) as a function of strain.
loading (i.e. lateral loads), and the group reduction factors
By analysing 22 full-scale lateral load tests on piles with
that must apply to reflect the pile group effects.
different materials embedded in different types of soils,
The objective of this paper is to provide engineers with
Mwindo (1992) developed some empirical relations to esti-
a representative soil–spring stiffness in which the effect of
mate the strain-dependent soil–spring stiffness. These rela-
depth and the degradation of rigidity parameters as a function
tions were further analysed by Prakash and Kumar (1996) to
of strain is implicitly taken into account and a step-by-step

CONTACT Lassaad Hazzar hazzar.lassaad@hydro.qc.ca Hydro-Quebec, Montreal, Quebec H2X 4P5, Canada
© 2019 Informa UK Limited, trading as Taylor & Francis Group
2 L. HAZZAR ET AL.

design procedure to analyse piles under several loads. A series of
3D finite difference FD analyses using FLAC3D (Itasca 2009) are
carried out to predict the soil–spring stiffness and the reduction
factors of the piles group located under the central pier of the
Sault–au–Mouton river bridge (Longue–Rive, Quebec–Canada)
and subjected to vertical, lateral, overturning, and torsion loads.
This soil–spring stiffness vary depending on the pile distortion
and may be adapted in an iterative process according to the pile
deformation at each depth.
(1) West abutment: top of the footing directly below the
ground surface at the elevation of 38 m;
(2) Central pier: above the sole directly under the surface
of the riverbed elevation at 27 m;
(3) East abutment: above the sole to 38 m elevation in the
approach embankment.
A footing that is supported by a group of six drilled shafts
(Figure 1) supports the central pier (the part of the founda-
tion, which is the subject of this paper).

Group piles under multiloads

Description of the bridge project
The project consists of a bridge structure above the river
Sault–au–Mouton. The bridge deck contains two spans, each
75 m in length and 14.0 m in width. According to the
architectural plan of the bridge, the footings will be posi-
tioned at the following elevations:
Soil properties
A total of three boreholes with standard penetration test (SPT),
identified as F29 to F31, were made in 2004 to investigate the
soil conditions and to determine the rock depth around the
central pier (Figure 2). During the SPT tests, the soil deposit is
constituted of two sandy layers with groundwater at the sur-
face. Figure 2 shows the profiles of the blow count (N)

Figure 1. Bridge P-15,705 in Longue–Rive, 3D model of central pier.

N1 (SPT) Fine content (%)

0 20 40 60 80 100 120 0 20 40 60 80
0

-2

Footing base
-4

-6
Depth (m)

-8

-10

-12

-14

-16 F-29-04
F-30-04
a) F-31-04
-18
b)

Figure 2. Soil stratigraphy in the site of central pier: (a) the profile of the blow count corrected for the effective stress and energy level used in the SPT (N1) and (b)
the fine content profile.

corrected for the effective stress and energy level (N1) and the
fine content. The analysis of the soil behaviour around piles
subjected to lateral load requires the definition of soil stiffness
(shear modulus G(γ)ρ and Gmax) and soil resistance (undrained
shear strength cu, or friction angle ϕ) parameters. At very low
strain, Gmax is estimated from the shear–wave velocity (VS) and
the mass density (ρ) by the correlation:
Gmax ¼ ρ  VS2
The shear–wave velocity (VS) of each sandy layer is related
to the stress–normalized shear–wave velocity (VS1) for the
effective vertical stress (σv’) by the following equation:

0 0:25
σv
VS ¼ VS1 (2)
100
According to Karray and Éthier (2012), VS1 is related to the
blow count corrected for the effective stress and energy level
used in the SPT (N1), and the mean grain size (D50), as follows:
VS1 ¼ 108N10:25 D0:18
50
Table 1 summarizes the properties of these soil layers.
Piles properties
A total of six drilled shaft piles were installed. The piles have
a diameter (D) of 2.0 m and a length (L) of 16 m. A config-
uration of two by three piles has been proposed with spacing of
5.0 m by 10.0 m (Figure 3). Mechanical characteristics include
Table 1. Soil properties after geotechnical tests.
Depth (m) 0.0–6.0
Unit weight, ρ (kN/m3) 21.00
Angle of friction, ϕ (°) 35.00
Shear–wave velocity, VS (m/s) 200.00
Poisson’ ratio, ν (drained condition) 0.33
Figure 3. Pile group configuration: (a) the loading scenarios overlaid on the pile cap and (b) plan view of the pile group.
INTERNATIONAL JOURNAL OF GEOTECHNICAL ENGINEERING 3
a Poisson’s ratio (νp) of 0.20 and a modulus of elasticity (Ep) of
25 GPa.
Types of loads applied to the bridge
The bridge is located in seismic performance zone 3: referring
to the specification of the CAN/CSA–S6–06 (2006) code, this
structure is classified in the category of Lifeline Bridge.
(1) Subsequently, a seismic analysis with multi-mode spectral
method is considered and three loading cases have been
analyzed (Table 2):
● Case 1: applied forces without taking into account the
seismic excitation;
● Case 2: applied forces taking into account a 100% trans-
versal seismic excitation and a 30% longitudinal seismic
excitation;
● Case 3: applied forces taking into account the 100%
longitudinal seismic excitation and the 30% transversal
(3) seismic excitation.
Figure 3a shows the loading scenarios overlaid on the pile
group cap.
Prediction of soil–spring stiffness
Hypothesis
The soil structure interaction is a complex problem which must
be treated with caution. Figure 4 shows the problem under
consideration, where the soil beneath the foundation can be
6.0–20.0 replaced by three springs of impedance to reproduce the lateral,
21.00 vertical and rotation movements. In the case of Sault–au–
40.00 Mouton river bridge, the piles are deeply anchored to the
230.00
0.33 rock, and it is justified to use fixed nodes without introducing
4 L. HAZZAR ET AL.

Table 2. Values of applied forces. The Winkler (1867) method, also known as the subgrade
Case 1 2 3 reaction method, currently appears to be the most widely
Lateral load, Px (kN) 123 2.280 7.010 used in a design of laterally loaded piles. This method was
Lateral load, Py (kN) 1.950 3.150 1.340
Axial load, Pz (kN) 39.152 28.100 27.800
first introduced to analyse the response of beams on an elastic
Moment, Mx (kN.m) 30.000 53.010 22.200 subgrade by characterizing the soil as a series of independent
Moment, My (MN.m) 1.800 120 350 linearly–elastic soil springs. Since then, this concept has been
Torque, Mz (kN.m) 1.900 37.100 113.500
extensively employed for the laterally loaded pile problem.
The term of subgrade reaction indicates the lateral pres-
sure (p) per unit area of the surface of the contact between
a loaded beam or slab and the subgrade on which it rests and
Soil reaction
on to which it transfers the loads. The spring coefficient (k)
on walls known as the coefficient of subgrade reaction corresponds to
the ratio between the soil lateral pressure (p) at any given
point of the surface of contact and the lateral deflection (y)
produced by the load application at that point:
Soil
p
k¼ (5)
y
KH
The soil–spring stiffness or the modulus of subgrade reaction
(KS) is related to the pile diameter (D) by the following equation:
KV
KR
KS ¼ k  D (6)
Although this modulus of subgrade reaction is simply defined
(Eq. 6), it has been proven that it is very difficult to evaluate
Figure 4. Soil–pile interaction problem schema. (Hazzar 2014). This is because it cannot be measured in
laboratory tests but must be back-calculated from full-scale
impedance springs. For the soil around the piles, the problem is field tests. Several approaches (Hazzar 2014) have been devel-
different and the soil can be replaced with springs and dampers oped to predict KS. It has been shown that this parameter is
to reflect its stiffness and attenuation around the piles. For variable not only with the soil type and mechanical proper-
harmonic excitation of frequency (ω) the dynamic–stiffness ties, but also with stress level and the geometry of the pile.
coefficient is written as (Wolf 1997): It is evident that the Winkler model (Figure 5) cannot fully
capture the 3D aspect of soils. The fundamental assumption
K ðωÞ ¼ KS ðk þ ia0 cÞ (4) on which the technique of constructing a p–y curve is based is
the similarity between the load-deformation pattern of pile
With the dimensionless frequency a0 = ωD/VS (shear–wave head and the stress-strain behaviour of the interacting soil
velocity VS). In this complex variable notation, KS represents from carefully chosen element testing (e.g. triaxial tests).
the static soil–spring stiffness, k the spring coefficient (para-
meters to define in the next subsection), and c the correspond-
Modelling procedure
ing damping coefficient. In the current bridge project, the a0
value is of the order of 0.05 for the natural frequency of the To evaluate the soil-spring stiffness of piles group located
structure (f= 0.8 Hz). For the first three to four vibration modes, under the central pier of the Sault–au–Mouton river bridge
theoretically, a0 should not exceed 0.2. According to Kramer,
Satari, and Kilian (1990), the frequency dependence of a single
pile is usually not very important and especially at low values of
a0. A 3D numerical study performed by Massioud et al. (2011)
reaches the same conclusion. Therefore, it is possible to con-
sider a value of k equals to 1 and to use the values of soil–spring
stiffness (KS) in a multimodal spectral analysis. The soil damp-
ing ratio corresponds to that proposed for the structure.

Coefficient of subgrade reaction

With the absence of liquefaction problem and low pore pres-
sure generation during earthquake (Pasic and Karray 2016), the
problem of soil structure interaction can be reduced to
a problem of subgrade reaction modulus (or soil–spring stiff-
ness) evaluation. The assessment of subgrade reaction modulus
Figure 5. Single pile under lateral loading: (a) real vertical pile, (b) Winkler
can be treated in different ways that can lead to different results. idealization, and (c) p–y curves for lateral Winkler springs.

(Longue–Rive, Quebec–Canada) and subjected to vertical,
lateral, overturning, and torsion loads, a modelling procedure
is proposed to establish initially, using 3D numerical simula-
tions, the p–y curves for a single pile. The group of six piles
(2x3) is then modelled to determine the group effect. The p–y
curves were determined considering the piles and soil proper-
ties, taking into account the degradation of shear modulus (G
(γ)) as a function of the distortion (γ) of the soil around the
piles. These p–y curves are constructed to define the equiva-
lents springs stiffness, which take into account the effect of
soil degradation.
Finite differences modelling
Finite difference grid and boundary conditions
The 3D FD program FLAC3D (Itasca Consulting Group
2009) was employed to study the lateral displacement,
lateral resistance and stiffness of a single pile and pile
group. Figure 6 shows the 3D finite difference grids used
in the numerical analyses. A grid generator subroutine was
implemented using the FISH built–in programming lan-
guage providing the possibility of grid refinement and
geometry variation. The bottom elevation and the lateral
sides of the computational domain were taken far enough
from the group to avoid any significant boundary effect. In
order to make sure that the zones size has no effect on the
response of the characteristic piles, trial analyses have been
carried out to optimize mesh discretization. More specifi-
cally, the response of each characteristic pile has been
established carrying out an analysis in which only this
pile was activated. The responses of these piles have been
compared and mesh re-adjustment was performed as
necessary.
(a)
Figure 6. Finite difference grid for the 2 × 3 pile group with footing.
INTERNATIONAL JOURNAL OF GEOTECHNICAL ENGINEERING 5
Soil model
The Mohr–Coulomb’s law (M.C.) constitutive relation was used
to model soil layers and define shear failure of a soil mass based
on its state of stress. This elastic-perfectly plastic model requires
five basic input parameters, namely bulk modulus (Kmax), shear
modulus (Gmax), mass density (ρ), undrained shear strength (cu),
and friction angle (ϕ). As discussed earlier, Gmax is estimated from
equation (1) and Kmax is related to Gmax and the Poisson’s ratio
(ν) by the following equation:
2ð1 þ νÞ
Kmax ¼ Gmax (7)
ð3  6νÞ
Therefore, the material properties adopted in the analyses
for the two layers (referred to the properties deduced from
geotechnical tests) are presented in Table 3. Standard elastic/
plastic constitutive laws, such as M.C., can also produce the
reduction of shear–modulus that can be evaluated from
degradation curves: consider the M.C. model with
a constant shear modulus (Gmax) and a constant yield stress
(τm) corresponding to a shear strain or distortion (γ). In
pre-failure phase, the secant shear modulus (G) is assumed
to be equal to Gmax. For a cyclic excitation that involves
failure, G/Gmax is given by:
G τm γ
¼ ¼ m (8)
Gmax γGmax γ
In this present numerical analysis, the shear modulus–reduc-
tion curve that relates the ratio G/Gmax to γ is incorporated in
order to account for soil non-linearity. The upper limit of the
shear–modulus degradation curve proposed by Seed and
Idriss (1970) for sandy soil, is adopted (Figure 7).
(b)
6 L. HAZZAR ET AL.

Table 3. Soil parameters according to M.C. FnðtþΔtÞ ¼ kn un A þ σ n A (10)

Depth (m) ρ (kg/m3) Gmax (MPa) Kmax (MPa) cu (kPa) ϕ (°)
0.0–6.0 2100 54 117 0 35 ðtþΔtÞ ðtÞ ðtþ0:5ΔtÞ
0–20.0 2100 126 273 0 40 Fsi ¼ Fsi þ ks Δusi þ σ si A (11)
where Fn and Fsi are the normal and shear force, respectively,
Pile model kn and ks the normal and shear stiffness, respectively,ΔΔusi the
incremental relative shear displacement vector, un the absolute
The pile was modelled as linear-elastic material. Three para-
normal penetration of the interface node into the target face, σn
meters were required to define the pile material behaviour.
the additional normal stress added due to interface stress
These parameters are the elastic bulk modulus (Kp), the
initialization, and σsi the additional shear stress vector due to
elastic shear modulus (Gp), and the mass density (ρp).
interface stress initialization.
In many cases, particularly when linear elastic analysis is
performed, values for interface stiffness are assigned to
Soil–pile interface model
simulate the nonlinear behaviour of a physical system. In
The interface elements were modelled by the linear Coulomb the present nonlinear analysis, the value for the interface
shear–strength criterion that limits the shear force acting at stiffness should be high enough, in comparison with the
an interface node. The shear–strength criterion that limits the surrounding soil, to minimize the contribution of those
shear force acting at an interface node for sandy soils is given elements to the accumulated displacements. According to
by equation (8). the results of numerical analyses, a value of 106 kPa/m for
both kn and ks was sufficient to ensure that no additional
Fs max ¼ ðFn  p1 AÞtgϕi (9) deflections were attributed to the pile due to the deforma-
tion of the interface springs. The use of considerably higher
Separation is able to cause a significant increase in displace- values is tempting as it could be considered as more appro-
ments (Poulos and Davis 1980) and therefore the interface priate, but in that case, the solution convergence would be
elements are allowed to separate if tension develops and very slow.
exceeds the tension limit of the interface. Once a gap is
formed between the soil–pile interfaces, the shear and normal
Soil lateral pressure
forces are set to zero.
The normal and shear forces at the interface nodes are The soil lateral pressure (p) can also be calculated by sum-
determined by the following equations: ming the forces in the relevant direction acting on the soil–

30
b) Damping
25

20
Damping ratio (%)

15

10

5 Limits for sand

Seed and Idriss (1970)
0
1

0.8 Limits for sand

Seed and Idriss (1970)

0.6
G/Gma x

0.4

0.2

a) Shear modulus, G
0
0.0001 0.001 0.01 0.1 1
Shear strain, γ (%)
Figure 7. Adopted curve of shear modulus degradation for sandy soils (Seed and Idriss 1970).
 INTERNATIONAL JOURNAL OF GEOTECHNICAL ENGINEERING 7

where,
xi
nx;i ¼ cos θ ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffi
xi þ yi2
2

yi
ny;i ¼ sin θ ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffi
xi þ yi2
2

σi is the normal stress at the interface node at point Q, τxy,i

is the shear stress at the interface node at point Q, xi is the x–
coordinate of the interface node at point Q, yi is y–coordinate
of the interface node at point Q, and Ai corresponds to the
representative area of interface node.

Model validation
Before describing the numerical results, it should be noted
Figure 8. Schematic plot of soil–pile system with interface between them that the applicability of the adopted model was already cali-
(Hazzar 2014). brated and validated by predicting the pile response in several
published pile loading tests, where a single and group of piles
pile interface nodes at the same depth. A schematic of the are subjected to lateral loads (Hazzar, Hussien, and Karray
soil–pile system is presented in Figure 8. Each interface node 2017). In Hazzar, Hussien, and Karray (2017), a full scale
is associated with a normal force and a shear force (Hazzar lateral load test of a 3 × 5 pile group, performed at the Salt
2014). Lake City International Airport (Snyder 2004; Tobita et al.
The x–component of the lateral soil–pile pressure is 2008), is validated in terms of lateral load–deflection curves
summed over all the interface nodes to calculate the lateral and bending moment distribution for both single pile and pile
soil pressure p per unit length along the pile at a particular group. Figure 9 shows the computed and measured average
pile section and is expressed as: load per pile versus deflection. The computed results slightly
overestimated the lateral load carrying capacity of the piles 5,
X
n    6 and 8 whereas the computed loads of the two other piles (10
p¼ σ i nx;i þ τ xy;i ny;i  Ai (12)
i¼1
and 15) agreed well with the measured ones. Consequently,
the analysis model used for this study has a reasonable

Figure 9. Laterally loaded pile group: computed and measured lateral load for piles 1, 6, 8, 10, and 11 versus lateral deflection (Hazzar, Hussien, and Karray 2017).
8 L. HAZZAR ET AL.

applicability to capture the essential behaviour of the pile According to the definition of modulus of subgrade reac-
groups located under the central pier of the Sault–au– tion or soil–spring stiffness in the second section, the varia-
Mouton river bridge and subjected to multi–loading tion of the stiffness of equivalent springs KSeq with the pile
conditions. distortion γ is plotted at several depths in Figure 11. Figure 11
indicates that the stiffness’s of soil–springs are not constants
as they have already mentioned the most methods considered
Numerical results and discussions in practice (Hazzar 2014). Table 4 shows the soil–spring
stiffness predicted by the current numerical analysis and
On the basis of the constitutive model parameters
those given by several methods adopted in the practice.
described previously, the response of the laterally loaded
According to Table 4, these methods give reasonable values
single pile is presented in terms of p–y curves and stiff-
but do not explicitly consider the effect of depth and the pile
ness of equivalent soil–spring versus pile distortion. While
distortion.
the pile group response is presented in terms of lateral
deflection profile of each pile and its corresponding lateral
soil pressure profile as well as the pile-soil-pile interaction
Piles group effects
(group effects).
The lateral response of the 2 × 3 pile group was studied
with FLAC3D for the several loads conditions (Table 3).
Stiffness of equivalent soil–spring Figure 12 shows the profiles of lateral deflection and
lateral soil reaction for the loading case 1. It is clear that
The most widely used non-linear analysis for laterally
the behaviour is not the same for each pile. The group
loaded piles is the p–y curves, which are obtained based
results were compared with the lateral response of a single
the methods of prediction of lateral deflection (y) and
pile. Consequently, group reduction factors were deter-
lateral pressure (p) described in last sections. Figure 10
mined for the equivalent soil–spring stiffness along the
shows the numerical p–y at six depths along the pile. The
depth (Table 5). These reduction factors increase with
lateral soil capacity (pmax) increases with the depth. This
depth and its values depend on the location of each pile
increase can be explained by a decrease in the pile lateral
in the group.
deflection with the depth.

400
pmax (5m) = 350 kN

pmax (4m) = 300 kN

Soil lateral pressure, p (kN)

300

200

100

a) depth = 4 m b) depth = 5 m
0
700
pmax (7 m) = 600 kN
600
Soil lateral pressure, p (kN)

pmax(6 m) = 490 kN
500

400

300

200

100
c) depth = 6 m d) depth = 7 m
0
700
pmax (7 m) = 610 kN pmax (7m) = 640 kN
600
Soil lateral pressure, p (kN)

500

400

300

200

100
e) depth = 8 m f) depth = 9 m
0
0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80
Lateral deflection, y (mm) Lateral deflection, y (mm)

Figure 10. p–y curves between 4 and 9 m of depth.

 INTERNATIONAL JOURNAL OF GEOTECHNICAL ENGINEERING 9

50
depth = 4 m
45 depth = 5 m
depth = 6 m
40

Equivalent stiffness, KSeq (KN/mm)

depth = 7 m
35

30

25

20

15

10

0
50
depth = 8 m
45
depth = 9 m
depth = 10 m
40
Equivalent stiffness, K Seq (KN/mm)

depth = 11 m
35

30

25

20

15

10

0
0.001 0.01 0.1 1
Pile distorsion, γ (%)

Figure 11. Stiffness of equivalent soil–spring versus pile distortion between 4 and 11 m of depth.

Table 4. Modulus of subgrade reaction KS (kN/mm): Comparisons between the proposed to adopt the soil-spring stiffness (soil around the
current analysis and several methods.
piles) according to the pile distortion (i.e. pile lateral deflec-
Depth Current Broms Ménard, Bourdon, and Poulos Gilbert
(m) analysis (1964b) Gambin (1971) (1971) (1995)
tion at surface). This methodology can be applied in two
0.0–6.0 4 to 33 22.0 25.0 11.8 26.96
steps. In the first step, the designer calculates the pile distor-
6.0–20.0 2 to 40 22.0 32.0 27.5 30.9 tion for given soil-spring stiffness’s defined at several depths.
These stiffness’s can be chosen arbitrarily (e.g. using Broms
method) in a structural analysis (including piles) so that the
Proposed methodology maximum pile distortion does not exceed the value pre-
To help the designers to consider the soil effect in their scribed by the used design code. The second step consists of
structural analysis (see Introduction), a methodology is reassess, using Figure 11, the soil-spring stiffness’s according

Lateral deflection (mm) Lateral soil reaction (kN)

-2 -1 0 1 2 3 4 -100 -80 -60 -40 -20 0 20 40

-4

-6

-8

-10
Depth (m)

-12
Pile 1
-14
Pile 2
Pile 3
-16 Pile 4
Pile 5
-18
Pile 6
a) b)
-20

Figure 12. Lateral response of pile group under the condition of loading corresponding to the case 1: (a) lateral deflection profiles and (b) lateral soil reaction profiles.
10 L. HAZZAR ET AL.

Table 5. Reduction factors to reflect the group effects. References

Factor of reduc- Factor of reduction/ Factor of reduction/
tion/longitudinal transversal direction/ transversal direction/ Broms, B. B. 1964a. “Lateral Resistance of Piles in Cohesive Soils.”
Depth (m) direction middle piles back piles Journal of Soil Mechanics & Foundations Div, ASCE 90 (2): 27–64.
4 0.49 0.13 0.45 Broms, B. B. 1964b. “Lateral Resistance of Piles in Cohesionless Soils.”
5 0.48 0.14 0.38 Journal of Soil Mechanics & Foundations Div, ASCE 90 (SM3):
6 0.51 0.10 0.39 123–156.
7 0.51 0.09 0.42 CGS. 2013. Canadian Foundation Engineering Manual. 4th ed.
8 0.52 0.12 0.49 B. C. Richmond. Richmond, British Colombia, Canada: Canadian
9 0.53 0.21 0.62 Geotechnical Society.
10 0.53 0.34 0.70 Chin, I. T. Y., I. D. G. S. Sew, and I. F. C. Chung. 2010. Interpretation of
11 0.53 0.51 0.90
Subgrade Reaction from Lateral Load Tests on Spun Piles in Soft
12 0.58 0.73 1.00
13 0.72 0.95 1.00 Ground. Kuala Lumpur, Malaysia: G&P Geotechnics Sdn Bhd.
14 0.88 >1.00 >1.00 Cox, W. R., L. C. Reese, and B. R. Grubbs (1974). “Field Testing of
15 >1.00 >1.00 >1.00 Laterally Loaded Piles in Sand.” Proceedings of the 6th Annual
16 >1.00 >1.00 >1.00 Offshore Technology Conference, Houston/Texas,pp. 2079.
17 >1.00 >1.00 >1.00 Davisson, M. T. 1970. Lateral Load Capacity of Piles. Highway Research.
18 >1.00 >1.00 >1.00 Record 333, 104–112.
19 >1.00 >1.00 >1.00 Gilbert, C. 1995. “Une nouvelle approche des calculs d‘interaction sol-
20 >1.00 >1.00 >1.00 structure.” Revue Française de Géotechnique 7: 3–9. doi:10.1051/geo-
tech/1995072003.
Hazzar, L. (2014). “Analyse numérique de la réponse des pieux sous
to the pile distortion calculated in the first step. This second sollicitations latérales.” Ph.D. thesis. Sherbrooke University,
step is iteratively repeated until the difference between the Sherbrooke, Quebec/Canada, 214 p.
pile distortion calculated during the structural analysis and Hazzar, L., M. Karray, M. Bouassida, and M. N. Hussien. 2013.
the one used in the precedent step becomes relatively small. “Ultimate Lateral Resistance of Piles in Cohesive Soil.” DFI Journal
- the Journal of the Deep Foundations Institute 7 (1): 59–68.
To consider the pile group effects, for each pile in the 2 × 3 doi:10.1179/dfi.2013.005.
pile group, a group reduction factor is estimated (Table 5) for Hazzar, L., M. N. Hussien, and M. Karray. 2017. “On the Behaviour of
several depth. At given depth, this factor is multiplied by its Pile Groups under Combined Lateral and Vertical Loading.” Ocean
corresponding KSeq (the final value predicted in the second Engineering 131: 174–185. doi:10.1016/j.oceaneng.2017.01.006.
step). It should be mentioned that the group factors (Table 5) Itasca Consulting Group. (2009). FLAC3D: Fast lagrangian analysis of
continua in 3-dimensions 4.1, manual. Itasca, Minneapolis.
correspond to the loading case 1 and others factors have been Karray, M., and Y. Éthier. 2012. “Reply to the Discussion by P.K.
proposed for the loading cases 2 and 3. Robertson on ‘Influence of Particle Size on the Correlation between
Shear Wave Velocity and Cone Tip Resistance.” Canadian
Geotechnical Journal 49 (1): 124‒128.
Conclusions Kramer, S. L., R. Satari, and A. P. Kilian (1990). “Evaluation of in Situ
Strength of a Peat Deposit from Laterally Loaded Pile Test Results.”
For prediction the soil–spring stiffness of pile groups located Transportation Research No. 1278, Washington, D.C., pp. 103–109.
under the central pier of the Sault–au–Mouton river bridge doi:10.1099/00221287-136-2-327
Kumar, S., L. Lalvani, and M. Omar. 2006. “Nonlinear Response of
(Longue–Rive, Quebec–Canada) and subjected to lateral load, Single Piles in Sand Subjected to Lateral Loads Using K_hMax
vertical load, and moments, a series of rigorous 3D numerical Approach.” Geotechnical and Geological Engineering 24: 163–181.
analysis based on finite difference technique has been per- doi:10.1007/s10706-004-2760-4.
formed. The Mohr Coulomb criterion is used to model the Matlock, H. 1970. “Correlations for Design of Laterally Loaded Piles in
soil parameters and recognized shear–modulus degradation Soft Clay.” Journal of Soil Mechanics & Foundations Div, ASCE SM5:
63–91.
curve given by Seed and Idriss (1970) to adjust the value of Ménard, L., G. Bourdon, and M. Gambin. 1971. “Méthode générale de
shear modulus of soil. calcul d‘un rideau ou d‘un pieu sollicité horizontalement en fonction
The results obtained are interpreted rationally to conclude, des résultats pressiométriques.” Sols–Soils 6 (22/23): 16–29.
initially, that the stiffness of equivalent springs varies depend- Messioud, S., D. Dias, U. S. Okyay, and B. Sbartai (2011). “Impédances
ing on the pile distortion and may be adapted in an iterative dynamiques de fondations sur groupe de pieux.” Proceedings of the
19th Civil Engineering University Meeting, Tlemcen/
process according to the pile deformation at each depth. This Algeria,29–31 May, pp. 343–352.
process can be compared to the linear equivalent method Mwindo, J. M. (1992). “Strain Dependent Soil Modulus of Horizontal
used in dynamic analysis where an equivalent shear modulus Subgrade Reaction”. M.Sc. thesis. University of Missouri, Rolla/USA.
is adapted according to the shear distortion. This type of Pasic, A., and M. Karray (2016). “Conception du pont sur la route 138 à
linear equivalent spring is very useful for incorporating the Longue-Rive, avec l’attention particulière sur le comportement du sol
exposé aux efforts sismiques.” Presentation in the 23rd Symposium on
effect of soil lateral resistance in modelling. Secondly, the the Quebec research progress in engineering structures, Quebec/
reduction factors reflect the group effects depending upon Canada, 3 May.
the depth and spacing between the piles in the group. Poulos, H. G. 1971. “Behaviour of Laterally Loaded Piles: I-Single Piles.”
Journal of Soil Mechanics & Foundations Div, ASCE 5 (97): 711–731.
Poulos, H. G., and E. H. Davis. 1980. Pile Foundation Analysis and
Design. New York: John Wiley & Sons .
Disclosure statement Prakash, S., and S. Kumar. 1996. “Nonlinear Lateral Pile Deflection
No potential conflict of interest was reported by the authors. Prediction in Sands.” Journal of Geotechnical Engineering 122 (2):
130–138.

Reese, L. C., and W. F. Van Impe. 2001. Single Piles and Pile Groups
under Lateral Loading, 508. Rotterdam: A.A. Balkema.
Reese, L. C., W. R. Cox, and F. D. Kooper (1974). “Analysis of Laterally
Loaded Piles in Sands.” Proceedings of the 6th Annual Offshore
Technology Conference, Houston/Texas,OTC 2080, pp. 437–483.
Seed, H. B., and I. M. Idriss (1970). “Soil Moduli and Damping Factor for
Dynamic Response Analyses”. Report No. EERC 70-10, Earthquake
Engineering Research Center, University of California, Berkley, CA.
Snyder, J. L. (2004). “Full-Scale Lateral-Load Tests of a 3x5-Pile Group in Soft
Clays and Silts.” M.Sc. thesis. Brigham Young University, Provo, Utah.
Terzaghi, K. 1955. “Evaluation of Coefficients of Subgrade Reaction.”
Géotechnique 5: 297–326. doi:10.1680/geot.1955.5.4.297.
INTERNATIONAL JOURNAL OF GEOTECHNICAL ENGINEERING 11
Tobita, T., G. C. Kang, S. Iai, and K. M. Rollins (2008). “Analysis of
Statnamic Behaviour of Full-Scale Pile Group in Soft Clays and Silts.”
In: Proceedings of Geotechnical Earthquake and Engineering Soil
Dynamics IV Congress ASCE, GSP181.
Vesic, A. (1961). “Design of Pile Foundations. National Cooperative
Highway Research Program Synthesis of Highway Practice.” Report
No. 42, Transportation Research Board, Washington, D.C., 68 p.
Winkler, E. 1867. Die lehre von der elasticita et und festigkeit. Prag,
Dominicus.
Wolf, J. P. 1997. “Spring–Dashpot–Mass Models for Foundation
Vibrations.” Earthquake Engineering & Structural Dynamics 26:
931–949. doi:10.1002/(ISSN)1096-9845.