Soil Dynamics and Earthquake Engineering 145 (2021) 106712

Contents lists available at ScienceDirect

Soil Dynamics and Earthquake Engineering

journal homepage: http://www.elsevier.com/locate/soildyn

A study on the effects of piled-raft foundations on the seismic response of a

high rise building resting on clayey soil
HamidReza Bolouri Bazaz a, Ali Akhtarpour b, *, Abbas Karamodin b
a
Faculty of Civil, Water and Environmental Engineering, Shahid Beheshti University, Tehran, Iran
b
Department of Civil Engineering, Faculty of Engineering, Ferdowsi University of Mashhad, Iran

A R T I C L E I N F O A B S T R A C T

Keywords: Piled-raft foundations are often used to transfer the structural load to the depth of the soil and tall structures can
High-rise buildings be constructed on medium to soft soils by utilizing a group of piles. The static effects of a group of piles are quite
Clay evident while their dynamic effects are quite unknown and therefore, not considered in the design of the
Piled-raft foundation
superstructure.
Excavation
Soil-structure interaction
In the present research, the performance of pile-raft foundations on the seismic response of a 20-story
benchmark building resting on soft clayey soil has been investigated using a 2D finite difference model,
namely FLAC. The study has been carried out by analyzing the seismic response of the structure with and without
pile-raft foundations, on the surface and inside an excavation.
The results show that when the structure is constructed on the surface, piled-raft foundations have a significant
beneficial effect on the seismic response of the superstructure, for instance a 40% decrease in the base shear. In
the case where the structure is located inside an excavation the use of piled-raft foundations are not noteworthy
and do not have a considerable effect.
The results of this study give an insight on the effects of pile-raft foundations on the seismic response of a
superstructure to overcome the current shortage of information in this area and to better predict the behavior of
tall structures in soft soil situations.

1. Introduction
Over the last decades, there has been a great improvement in seismic
soil-structure interaction studies. Rapid development of hardware and
software technologies have made it possible for researchers to investi­
gate the interaction effects of complicated soil and structure systems [1].
The outcome of considering soil-structure interaction can be beneficial
or detrimental based on the soil type, earthquake characteristics, foun­
dation and structure dynamic properties, etc. [2].
The seismic effects of soil structure interaction for tall structures is
evident, especially when it comes to clayey and soft soils [2]. This
phenomena can be more complex when dealing with piled-raft foun­
dations. To this date, the studies regarding this issue are quite limited
and there is no certain estimation of the response of the soil-structure
system under these circumstances [3].
When strip and mat foundations are inadequate for stress trans­
mission of building to soil, piled-raft foundations can be employed [4]. It
is quite evident that piles increase the bearing capacity of soils and are a
good solution when encountering soft soils in static problems, but the
seismic performance of them remain quite unknown in different situa­
tions. Several studies have been performed to analyze the seismic
behavior of soil-pile-structure interactive systems which show that
piled-raft foundations may have detrimental or beneficial effects on the
seismic response of the structure (i.e. [1,5–8]), for example application
of piled-raft foundations increased the seismic response of two short
buildings [2,9], while on the other hand it decreased the response of a
20-story building [2]. The various frequencies of every system may
direct the proximity of natural and predominant frequency of the whole
system, which leads to severe failure of the whole system [3].
Remarkable numerical and experimental works have been done in the
field of soil-pile-structure interaction. Experimental approaches such as
centrifuge and 1-g shaking table tests have been performed mostly on
liquefiable/dry sandy soil (e.g. Refs. [1,10–13]) rather than soft clays (e.
g. Refs. [14–19]) due to the difficulties of preparation of soft clay. It
should be noted that most of the performed experimental approaches
assume the superstructure as a simplified single degree of freedom

* Corresponding author.
E-mail address: akhtarpour@um.ac.ir (A. Akhtarpour).

https://doi.org/10.1016/j.soildyn.2021.106712
Received 27 March 2020; Received in revised form 26 February 2021; Accepted 1 March 2021
Available online 21 March 2021
0267-7261/© 2021 Elsevier Ltd. All rights reserved.
H. Bolouri Bazaz et al. Soil Dynamics and Earthquake Engineering 145 (2021) 106712

Table 2
The drained and undrained parameters of Bangkok soft clay for the hardening
soil model (CYSoil Model).
Parameter Drained Parameters Undrained Parameters

Pr 100 kPa 100 kPa

Kr 5.87 4.41
m 0.9975 0.9975
R 4.5 4.5
α 0.87 0.87
cp 0 0
ϕp 23.6 29
ϕ0 0 10
ψp 0 0
0.293 0.293
εfs

2. Development of numerical model

The finite difference software adopted in this study is FLAC2D which

is being used in engineering mechanics computations [24]. In the
following sections, the soil characteristics, adopted soil constitutive
Fig. 1. Soil profile at the location of Sukhumvit Station – after Likitlersuang
model, structural elements, interface elements and boundary conditions
et al. [30].
are explained.

oscillator or in simple terms, a lumped mass. Few studies have consid­

ered the effect of higher modes of the superstructure by actually
modelling the superstructure (i.e. [6,20]). Numerical studies, such as
finite element and finite difference time domain analysis have been
carried out to study the seismic soil-pile-structure interaction by
assuming nonlinearity of soil behavior (e.g. Refs. [2,5,15,21,22]). In
addition, dynamic p-y method or beam-on-dynamic Winkler-foundation
model is also a common numerical approach to investigate
soil-pile-structure interactions (e.g. Refs. [3,8]).
The effect of piled-raft foundations are not limited to building
structures and can be observed in different types of structures especially
when the P-Delta effect is substantial, e.g. river bridges, where the
application of piled-raft foundation plays an important role in mid-span
failure of river bridges [23].
In this study, a total of 35 two-dimensional finite difference analysis
were performed using FLAC2D software to investigate the seismic
response of a 20-story benchmark building resting on clayey soils under
different conditions subjected to seven far-field ground motions by uti­
lizing the direct method of soil-pile-structure system analysis. To
simulate soft clay behavior more accurately a capped hardening soil
model has been employed. Moreover, nonlinearity in the structural
members have been considered by assuming elastic-perfectly plastic
behavior for them. The effects of piled-raft foundations are investigated
in different situations by probing the maximum seismic response of the
structure.
The understandings and outcome from this research may provide a
worthful perception on inelastic seismic behavior of tall structures with
piled-raft foundations.
2.1. Soil characteristics and adopted constitutive model
Selecting a soil profile is one of the critical steps in the definition of a
problem, especially when soft soils are under consideration. When
studying the soil-structure interaction for tall buildings, a deep soil
profile is usually applied for the modelling procedure. Due to the effect
of gravity and consolidation of deeper layers, imposing soft soil prop­
erties entirely to the soil profile may not be realistic. Therefore, in this
study the soil profile of a case study has been employed to obtain more
logical results.
Bangkok soft clay has been thoroughly used for the investigation of
soft soil behavior [25]. Numerous experimental and field tests have been
performed to acquire Bangkok soil properties [26–28]. Accordingly, the
soil profile of a recent mega project in Bangkok is selected which consists
of soft, medium, stiff and hard clay with increasing depth. The case study
chosen is Sukhumvit station of the Bangkok Mass Rapid Transit (MRT)
Underground Railway which includes significant geotechnical works
and experimental data. The details of the mega project and aforemen­
tioned station can be found in previous studies (i.e. [25,29–32]). The soil
profile at the location of the aforesaid case study is shown in Fig. 1 and a
summary of the drained and undrained geotechnical properties, derived
from a wide collection of experimental tests [29,30], are presented in.
The constitutive model selected for this study is a combination of two
models: (1) elastic-perfectly plastic model also known as the Mohr-
Coulomb model and (2) capped-hardening soil model (see Table 1).
The capped-hardening soil model is employed for modelling soft soil
behavior (layers 2a & 2b) owing to the fact that for soft soils the plastic

Table 1
The drained and undrained properties of the soil profile.
Layer No. Soil Type Depth Density Elasticity Modulus Drained Parameters Undrained Parameters

c’ ϕ′ υ c ϕ Su υ
1 MG 0–2.5 18 8 1 25 0.2 1 25 – 0.3
2a BSC1 2.5–7.5 16.5 10 Capped Hardening Soil Model Parameters
2b BSC2 7.5–12 16.5 20.5
3 MC 12–14 17.5 27.5 10 25 0.2 – – 55 0.495
4 1stSC 14–20 19.5 40 25 26 0.2 – – 80 0.495
5 CS 20–21.5 19 53 1 27 0.2 1 27 – 0.25
6 2ndSC 21.5–26 20 72 25 26 0.2 – – 120 0.495
7 HC 26–60 20 240 40 24 0.2 – – 240 0.495

2
H. Bolouri Bazaz et al. Soil Dynamics and Earthquake Engineering 145 (2021) 106712

Fig. 3. Caparison of normalized shear modulus experimental results and Sig3

model – experimental results after Teachavorasinskun et al., 2002 [36].

Fig. 2. Shear wave velocity of Bangkok subsoil [28].

Table 3
Parameters of Sig3 model.
Parameter Value

a 1
B − 0.5
x0 − 0.67

deformation is imposed by confining pressure as well as shear stress.

Although this model is much more complicated than the Mohr-Coulomb
model, its parameters can be determined by common soil mechanics
tests. The calibration process of this model for drained and undrained
parameters of Bangkok soft clay has been performed by Bolouri Bazaz
et al. with the use of a series of triaxial and odometer tests [25]. Table 2
shows the parameters obtained for Bangkok soft clay (Layers 2a & 2b). It
is worth mentioning that the capped-hardening soil model behavior has
been implemented by taking advantage of CYSoil constitutive model in
FLAC2D software.
The seismic response of a soil-structure system is highly dependent
on the dynamic properties taken into consideration.
One of the most effective parameters is the maximum dynamic shear
modulus of soil also known as the small-strain shear modulus. There are
many different experimental and field tests which can determine the
shear modulus of soil, but since the dynamic shear modulus reduces by
the increase of shear strain, it is necessary to acquire this parameter by
small strain experiments such as bender element or resonant column
rather than triaxial test [33]. A common way to obtain the shear
modulus in very small strain situations is the use of in-situ tests such as
cross-hole and down-hole [34] where the maximum shear modulus can
be obtained by the values of shear wave velocity (Vs) and density (ρ) as:
Gmax = ρVs2 (1)
In this study, the dynamic shear modulus has been derived from the
shear wave velocity of the soil profile which is illustrated in Fig. 2 [28].
Another aspect which considerably affects the seismic response of a
soil-structure system is the strain-dependent modulus and damping
functions. In this study, the equivalent-linear method has been
employed for defining the previous functions which is in use for many
years to calculate the wave propagation in sites subjected to seismic
excitation [24,35]. A series of experimental tests has been performed by
Teachavorasinskun et al. for the determination of G/Gmax and damping
ratios of Bangkok clays [36]. For the implementation procedure, a
Fig. 4. Caparison of damping modulus experimental results and Sig3 model –
experimental results after Teachavorasinskun et al., 2002 [36].
FLAC2D sigmoidal model (namely Sig3) has been employed. Table 3
shows the parameters considered for the model and Fig. 3 & Fig. 4 show
the comparison of the average experimental results and model predic­
tion for normalized shear modulus and damping ratio respectively. The
figures clearly show that the prediction of the model for the shear
modulus reduction is accurate while for the damping ratio, it is not quite
precise. It can be observed that for low shear strains (in this case under
0.2%) the predicted model underestimates the damping ratio while for
higher shear strains it is over estimated. For geotechnical problems
where the shear strains during earthquake are considerably high (i.e.
crest of a dam), hysteretic damping leads to over damped material and is
not recommended. For the case of this study, shear strains of soil during
seismic motions have been monitored and are predominantly under
0.2%, therefore the damping ratio predictions of the model are accept­
able. Furthermore, in most calibration attempts of the damping ratio, a
deficiency is observed between the predicted damping ratio and the
experimental results in low cyclic shear strains [37]. To overcome this
issue it is recommended to add a local or Rayleigh damping in addition
to hysteretic damping [24]. Hence, a local damping of one percent has
also been considered to cover this matter.
2.2. Structural characteristics
In the present research, the typical 20-story benchmark building
designed for the SAC project has been taken into account. The structure

3
H. Bolouri Bazaz et al. Soil Dynamics and Earthquake Engineering 145 (2021) 106712

Fig. 5. The characteristics of the 20-story superstructure.

4
H. Bolouri Bazaz et al. Soil Dynamics and Earthquake Engineering 145 (2021) 106712

Table 4
The normal and shear coupling spring.
Parameter Shear Coupling Spring Normal Coupling Spring

Cohesion csoil .P 9csoil .D

Friction ϕsoil ϕsoil
Stiffness ⎛ [ ]
4 ⎞ ⎛ [ 4 ⎞
]
10 K + G 10 K + G
⎜ 3 ⎟ ⎜ 3 ⎟
⎝ ⎠P ⎝ ⎠D
Δzmin Δzmin
( )
Tension – csoil
D
tan(ϕsoil )

Table 5
The selected earthquakes and their specifications.
Earthquake Station Name Magnitude Fault PGA
Incident (Mw) Mechanism (m/s2)

San Fernando Pasadena - Old 6.61 Re 0.2

(1970) Seismo Lab
Morgan Hill Gilroy - Gavilan 6.19 SS 0.12
Fig. 6. The elastic-perfectly plastic behavior of structural elements. (1984) Coll
Morgan Hill Gilroy Array #1 6.19 Re 0.098
is a moment resisting steel frame with 5 and 6 bays in the N–S and E-W (1984)
Loma Prieta UCSC 6.93 Re/Ob 0.4
direction respectively with a length of 6.1 m per bay in both directions
(1989)
[38]. Since two dimensional analysis are performed in this research, the Northridge-01 LA – Wonderland 6.69 Re 0.16
shorter edge of the building (N–S direction, 5 bays) has been considered. (1994) Ave
This structure originally has 20 stories and two additional basements Northridge-01 Vasquez Rocks 6.69 Re 0.15
which are horizontally fixed at level 0.0 with roller supports. Therefore, (1994) Park
Iwate-Japan IWT010 6.9 Re 0.29
in this study the 20-story building was considered without the base­ (2008)
ments and fixed supports have been assumed for the columns. Addi­
tionally, a 2 m foundation has been taken into account following Re: Reverse, SS: Strike Slip, Re: Reverse Oblique.
common regulations for the superstructure in the interactive problems.
The layout of the structure is illustrated in Fig. 5. utilized for the superstructure to considerably reduce the calculation
The elastic-perfectly plastic model was chosen for the structural el­ time.
ements for the reason that the elastic model may not accurately predict
the hierarchy of the failure mechanisms for structures subjected to 2.3. Interface of soil and structural elements
seismic motions [39,40]. In the latest version of the finite difference
software (FLAC2D) utilized in this study, the possibility of considering Shear and normal coupling springs can be utilized for simulating the
this model for structural elements has been provided. Fig. 6 shows a interface between soil and structural elements (i.e. foundation, piles and
schematic of the behavior of the structural elements. retaining wall). A spring-slider system interface connects these elements
The plastic moment of each element (Mp) has been determined by: by dint of force and motion. The characteristics of the interface system
M p = ZFy (2) are defined as cohesive strength, frictional resistance, and stiffness in the
normal and shear direction. Moreover a tension strength for the normal
where Z is the section modulus and Fy is the yield stress [41]. In addi­ coupling springs can be considered to model a gap between pile and soil
tion, the p-Δ effect was considered by initiating large-strain mode in during ground motion ([24,42]). The definition of these parameters,
seismic analysis of FLAC2D software. In this mode, the grid point co­ derived from prior studies (i.e. [24,42–44]) are summarized in Table 4.
ordinates are updated every step according to the displacements calcu­ Where, K, G, csoil & ϕsoil are respectively the bulk modulus, shear
lated during the seismic excitation of the model [24]. modulus, cohesion and friction angle of the adjacent soil, P & D are the
Rayleigh damping was regarded for the structural system (C = α M + perimeter and diameter of the structural element and Δzmin is the
β K; where C, M & K are damping, mass and stiffness matrices and α & β smallest width of the neighboring zone.
are constants) [24]. By considering the first and fourth period of vi­
bration in modal analysis of the superstructure and 5% damping, the 2.4. Selected earthquake records
mass-proportional and stiffness-proportional damping constants, α & β,
are respectively calculated as 0.1435 and 0.0076. Rayleigh damping In this study seven earthquake records selected from stations located
significantly increases the calculation time by increasing the calculation on rocky grounds, have been adopted to minimize the amplification
process for every step and decreasing the minimum time step. This leads effects of soil. Table 5 presents a summary of the specifications of the
to calculation times over 100 times slower when considering Rayleigh earthquakes [45]. There are commonly three methods to utilize the
damping. To overcome this issue another type of damping can be taken earthquake records, (a) direct use, (b) PGA scaling and (c) code specific
into consideration. Since the maximum seismic response of the structure ground-motion. Since the earthquake record itself is an uncertainty in
is the main finding in this research and is almost caused in the beginning earthquake engineering, code specific ground motions are often used in
of an earthquake, choosing another kind of damping may lead to similar practice since it modifies the recorded time series to match a specific
results. In this study various values of local damping was examined and spectrum and therefore fewer accelerograms need to be used [35]. Ac­
the maximum responses of the superstructure were compared regarding cording to the fact that in soil structure systems the earthquake record is
to Rayleigh damping in fixed based situation. It was observed that a local applied at the bottom of the model, the earthquake records were scaled
damping of 5% is the best match to the aforementioned Rayleigh by Perturbation approach [35] to match the ASCE class (A) response
damping and returns only about 5% error in seismic structural response spectra which corresponds to hard rock. S1 and Ss were selected as 1.61
of the superstructure which is acceptable. Therefore local damping was and 1.15 based on a site in downtown Los Angeles where the benchmark

5
H. Bolouri Bazaz et al. Soil Dynamics and Earthquake Engineering 145 (2021) 106712

It is worth mentioning that the 20-story building on the surface

without piled-raft foundation cannot be studied due to the softness of the
soil which caused instability for the superstructure. Also, for the models
inside excavation, the retaining wall and the structure are not con­
nected. A summary of the models studied in this research are presented
in Table 6.
The characteristics of the retaining walls have been obtained
regarding FHWA–NHI–14-007 [48]. The piles utilized in models 3 & 5
are 1 m in diameter and 8.8 m deep and have a spacing of 6.1 m in both
directions, in other words it is assumed that the piles are constructed
under the columns of the superstructure. For the fourth and fifth model,
the 20-story building is constructed in a 14.5 m excavation. The
retaining wall is assumed a 19 m deep concrete reinforced tangent pile
wall with a diameter of 1.4 m which is not connected to the super­
structure. The schematic illustration of the aforementioned models are
presented in Fig. 9.

3. Results and discussions

Fig. 7. The scaled response spectra of the selected earthquakes.

The two dimensional soil-structure interaction models presented in

building has been originally designed according to them [46]. The
the previous section have been analyzed in the time domain, undergoing
response spectra of the aforementioned earthquake records are illus­
seven scaled earthquake records, by the direct method in finite differ­
trated in Fig. 7.
ence software FLAC2D. This section presents and discusses the results of
the superstructure concerning base shear, horizontal and vertical ac­
2.5. Studied models celeration, displacement, inter-story drifts and moments. The response
of the superstructure is expressed in terms of “Max” and “Mean” values
In this study, five models were considered to investigate the effects of which are the maximum and average of the maximum values obtained in
piled-raft foundation on the seismic response of high-rise structures. each earthquake, respectively. The results disclose the significant effects
Two of the models are considered fixed-base while the other three are of piled-raft foundations in soft soils for the superstructure.
interactive models. All of the models were simulated in two dimensional
finite difference program (FLAC2D). The geometric dimensions of the
3.1. Base shear
interactive models, based on previous researches (e.g. Refs. [42,47]),
were selected as 250 m width and 60 m deep. Additionally, the struc­
In the current study, the base shear is the maximum lateral force of
tural elements have been scaled in perpendicular direction to carry out
the ground level taken place in a seismic motion. Comparison of the base
proper 2D modelling [24,42].
shear gives a view on the shear force induced in a structure during an
The input seismic motion for the fixed-base models have been
earthquake. The mean and max values of base shear is illustrated in
perceived from two free field models (see Fig. 8). The motion derived
Fig. 10. The comparison of base shear for models 1 & 2 shows a
from the surface (Fig. 8a) and from inside the excavation (Fig. 8b) have
considerable decrease for the 2nd model, i.e. the fixed base model
been used for the fixed-base models, models 1 & 2, corresponding to
induced from the excavation accelerograms. The disagreement of the
fixed-base surface motion and fixed base excavation motion, respec­
tively. This method is typically employed in common structural engi­
neering analysis. Studying the fixed base models with the mentioned Table 6
Summary of the models studied in this research.
input motions and the interactive models give a perspective on the ef­
fects of soil-structure interaction on the surface and inside an Model Fixed-base/ Input Seismic Piled-raft Inside
No. Interactive Motion Foundation Excavation
excavation.
Three interactive models studied in this research are as following: Model 1 Fixed-base Surface motion – –
Model 2 Fixed-base Excavation – –
motion
(i) 20-story building with piled-raft foundation on the surface Model 3 Interactive Earthquake Yes No
(ii) 20-story building without piled-raft foundation inside an Model 4 Interactive Earthquake No Yes
excavation Model 5 Interactive Earthquake Yes Yes
(iii) 20-story building with piled-raft foundation inside an excavation

Fig. 8. The input motion derived from free field models.

6
H. Bolouri Bazaz et al.
Fig. 9. The five models studied in the present research.
Fig. 10. The comparison of base shear in six studied models.
base shear is generally caused by the amplification effects of the top soft
soil layer which is present in the accelerograms of the 1st model.
Therefore the decrease of base shear in model no. 2 is quite evident.
When considering soil-structure interaction, a reduction of base shear is
commonly observed in medium to stiff subsoils (e.g. Refs. [49,50])
whereas it usually increases in soft soils (e.g. Refs. [51–53]). Although
the 3rd model is located over a soft subsoil, the base shear has signifi­
cantly decreased comparing to the non-interactive model (1st model)
which is quite unexpected. This could be the result of employing
piled-raft foundation to the building. On the other hand, the comparison
of models no. 4 & 5 shows that the use of piled-raft foundation inside the
excavation not only decreases the base shear, but also causes a small
increase in it.
3.2. Peak accelerations undergone in the structure
The vertical and horizontal acceleration of the left-side nodes of the
structure have been recorded during every seismic motion. PVA and
PHA correspond to the maximum values obtained for the vertical and
horizontal accelerations of each node respectively. The mean values of
PVA and PHA are illustrated in Fig. 11. For a better differentiation be­
tween the responses of the structure, the average values (see Fig. 12) can
be used as a basis for comparison.
In the fixed-base models (models 1&2), since vertical accelerations of
7
Soil Dynamics and Earthquake Engineering 145 (2021) 106712
Fig. 11. The peak accelerations observed in the superstructure.
the earthquakes were not applied at the base of the structure, the PVAs
developed in this model are noticeably lower than expected. which is
mainly due to foundation rocking in other models. Therefore, it is not
appropriate to compare the PVA of these models with the others.
The peak horizontal acceleration (PHA) of models no. 1 & 2 show an
average of 3% decrease for the 2nd model where the accelerograms
inside the excavation have been employed. The PHA in the 3rd model
has decreased about 40% regarding the first model which clearly shows
the impact of the piled-raft foundation on the seismic response of the
superstructure. Although piled-raft foundation has a decreasing impact
in model 3, it has an opposite effect in model 5 regarding model 4 and
the effect of piled-raft foundation, in this case for PHA, is an averagely
7% increase in the seismic response of the superstructure. Additionally,
it can be seen that the use of piled-raft foundation in the excavation has
increased the PVA of the superstructure (34% increase in PVA in model
no. 5 in respect to model no. 4). Furthermore, it is quite clear that the
PVA in model no. 3 is not less than the other models (model no. 4 & 5)
which implies that the use of piled-raft foundation in soft soils does not
H. Bolouri Bazaz et al.
Fig. 12. The average of peak accelerations of the superstructure.
Fig. 13. Drift & displacements of the superstructure.
have a decreasing effect for PVA whereas it considerably decreases PHA
in the response of the superstructure.
3.3. Drift and displacements
Displacements of structures during earthquakes are important
because of the disruptive impacts of adjacent buildings on each other.
Fig. 14. Average values of drift & displacements of the superstructure.
8
Soil Dynamics and Earthquake Engineering 145 (2021) 106712
Besides structural inter-story drifts throughout earthquakes are sub­
stantial to study for the case of sustainability and integrity of the
structure. Hereon the drift and displacements of the left column of the
superstructure have been observed to give a perspective on the seismic
structural response. Fig. 13 shows the mean values of drift and dis­
placements and Fig. 14 shows their average values for a more conve­
nient comparison. It should be noted that the displacements of the
structure have been recorded relative to the connection of the structure
and foundation.
The displacements and drifts recorded in models 1 & 2 (fixed base
models without considering SSI) show a 37% decrease in these param­
eters in model no. 2 where the accelerograms from the excavation have
been employed. A noticeable change in the results of this session can be
seen in the results of model no. 3. It can be seen that the horizontal
displacement and drifts of this model is between models no. 1 & 2 while
for the previous parameters, the results were less than them. In other
words the drifts and displacements of this model (model no. 3) have
increased by foundation rocking. The rotation of the foundation in this
model has been recorded and the final rotation of the foundation is
averagely 0.16% among the 7 earthquake records. Hence, the in­
compatibility of drifts and horizontal displacements recorded for the
structure of this model is originated by foundation rocking (rotation of
foundation).
For models 4 & 5 there is not a considerable difference (only 1%)
between the responses of the superstructure which confirm that the
employment of piled-raft foundation inside the excavation and in non-
soft soils does not have a significant impact on the response of the
structure.
3.4. Moments
The moments developed in a structural frame during an earthquake
can give an insight on the performance of the structure. The investiga­
tion of moments were carried out by analyzing the moments of the
beams of the left bay and the left column of the superstructure. The
results show that the moment of the beams have mostly reached their
maximum limit (i.e. plastic moment [Mp]) during the earthquakes. The
H. Bolouri Bazaz et al. Soil Dynamics and Earthquake Engineering 145 (2021) 106712

investigated by comparing different Soil-Structure-Pile interactive sys­

tems. The systems were defined to investigate the effect of employing a
piled-raft foundation on the seismic response of a superstructure resting
on clayey and soft soils. For this study, 5 different models were utilized
to study the mentioned issues. Although studying each subject requires a
large amount of research and the conclusions of this study cannot be a
basis for similar situations, but the results can give an insight on the
effects of piled-raft foundations on tall structures. By comparing the
seismic response of the aforementioned models, the following can be
concluded:

(1) The effect of piled-raft foundations on the seismic response of

Fig. 15. Moments of the superstructure.
mean moments of the left end of beams of the left side of the building is
illustrated in Fig. 15. It can be clearly seen that in models 3 to 5 the
moments have reached their plastic moment while in the fixed-base
models (models 1 & 2) the moments are less and the moments of
model number 2 are predominantly less than model number 1. A similar
behavior can be seen in the moments of columns. As shown in Fig. 16 it
can be seen that the average moments of model number 2 is less than
model number 1 which is, as discussed previously, the result of the
earthquakes induced in both models. The moments of beams and column
of model no. 3 does not accommodate the previous sections where it is
more than the other models. This behavior could be the consequence of
foundation rocking which has implied an extra moment to the structural
members where in the models studied in this research this model only
has foundation rocking. For the interactive models inside the excavation
(models 4 & 5) there is no significant difference between the results and
it can be seen that piled-raft foundation in excavation do not have a
remarkable impact on the seismic response of the studied
superstructure.
4. Concluding remarks
In this study the effects of piled-raft foundations have been
structure is primarily dependent on the situation of the structure.
For the case where the superstructure is located on the surface,
piled-raft foundations have a significant decreasing effect on the
seismic response of the structure while on the other hand when it
is located inside an excavation the effects are not noteworthy. The
effortlessness of piled-raft foundations for the situation where the
structure is located inside an excavation can be the result of the
soil underneath the structure where compared to the building on
the surface, is quite denser or the effect of the surcharge caused
by the soil above the bottom of the foundation. It should be noted
that for the models embedded in the ground (models no. 4 & 5)
the structure and retaining walls are not connected to eliminate
the effect of period shortening of the structure due to excavation
which the structure in excavation tends to act as a shorter
structure. In other words the underground stories do not act as a
rigid box and the period of the superstructure is not affected by
locating the building in an excavation. Nevertheless, the impact
of utilizing piles for a better seismic response is only limited to
soft soils. Due to insufficient data regarding this issue, it is rec­
ommended to perform SSI analysis in these situations to better
understand the extent of structural response reduction, since the
effects are undeniable. In this study, considering the 40%
reduction in the base shear for the superstructure located on the
surface, can have a substantial economic benefit in the design of
the superstructure.
(2) Locating any structure inside excavation mainly gives the
impression of better structural response, but in the case of this
study, the 20-story building does not react preferable inside the
excavation rather than the surface. It should be noted that this
event seems to be mainly caused by the utilization of piled-raft
foundation.
(3) The effects of foundation rocking for tall structures are quite
significant. In the case of this study, the drifts, displacements and
moments of the model with foundation rocking were influenced.
Although this model had a piled-raft foundation, but it did not
totally prevent the disadvantageous effects of foundation rocking.
The main aim of this study was to give a general understanding on

Fig. 16. Average Moments of the superstructure.

9
H. Bolouri Bazaz et al.
the behavior of a superstructure with piled-raft foundation in different
cases which shows a deficiency in the general belief of the response of
the structure. It should be noted that many seismic building codes
currently refer to SSI studies when encountering soft soils and some have
misleading recommendations. Hence, the results of this study can be a
fundamental point for further studies to overcome the current shortage
of information in this area and to predict the behavior of superstructures
in soft soil situations more accurately. Additionally, a lack of knowledge
in performance-based seismic design of structures can be observed
especially when the inelastic response of superstructures is to be
considered (e.g. Refs. [9,54]). Further scope of the study can be utilizing
inelastic and complex behavior for the superstructure and piled-raft
foundation as well as the surrounding and interactive soil.
Author statement
HamidReza Bolouri Bazaz: Methodology, Software, Validation,
Formal analysis, Investigation, Resources, Writing – original draft,
Visualization, Ali Akhtarpour: Conceptualization, Methodology, Project
administration, Supervision, Abbas Karamodin: Resources, Supervision.
Declaration of competing interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
References
[1] Peizhen L, Hongmei R, Yueqing C, Xilin L, Heping S. Shaking table testing of hard
layered soil-pile-structure interaction system. Front Architect Civ Eng China 2007;
34(3):307–13.
[2] Zhang L, Liu H. Seismic response of clay-pile-raft-superstructure systems subjected
to far-field ground motions. Soil Dynam Earthq Eng 2017;101(1):209–24.
[3] Roy J, Kumar A, Choudhury D. Natural frequencies of piled raft foundation
including superstructure effect. Soil Dynam Earthq Eng 2018;112(1):69–75.
[4] Poulos HG, Davis EH. Pile foundation analysis and design. The University of
Sydney; 1980.
[5] Luo C, Yang X, Zhan C, Jin X, Ding Z. Nonlinear 3D finite element analysis of
soil–pile–structure interaction system subjected to horizontal earthquake
excitation. Soil Dynam Earthq Eng 2016;84(1):145–56.
[6] Hokmabadi A, Fatahi B, Samali B. Assessment of soil-pile-structure interaction
influencing seismic response of mid-rise buildings sitting on floating pile
foundafoundations. Comput Geotech 2014;55:172–86.
[7] Boulanger RW, Curras CJ, Kutter BL, Wilson DW, Abghari A. Seismic soil-pile-
structure interaction experiments and analyses. J Geotech Geoenviron Eng 1999:
750–9.
[8] Takewaki I, Kishida A. Efficient analysis of pile-group effect on seismic stiffness
and strength design of buildings. Soil Dynam Earthq Eng 2005;25:355–67.
[9] Dutta SC, Saha R, Haldar S. Inelastic seismic behavior of soil–pile raft–structure
system under bi-directional ground motion. Soil Dynam Earthq Eng 2014;67:
133–57.
[10] Chau K, Shen C, Guo X. Nonlinear seismic soil-pile-structure interactions: shaking
table tests and FEM analyses. Soil Dynam Earthq Eng 2009;29(2):300–10.
[11] Dungca J, Kuwano J, Takahashi A, Saruwatari T, Izawa J, Suzuki H, Tokimatsu K.
Shaking table tests on the lateral response of a pile buried in liquefied sand. Soil
Dynam Earthq Eng 2006;26(2–4):287–95.
[12] Durante M, Di Sarno L, Mylonakis G, Taylor C, Simonelli A. Soil-pile-structure
interaction: experimental outcomes from shaking table tests. Earthq Eng Struct
Dynam 2015;45(7):1041–61.
[13] Lombardi D, Bhattacharya S. Modal analysis of pile-supported structures during
seismic liquefaction. Earthq Eng Struct Dynam 2014;43:119–38.
[14] Banerjee S. Centrifuge and numerical modeling of soft clay-pile-raft foundations.
Singapore: National University of Singapore; 2009. Ph.D. Thesis.
[15] Banerjee S, Goh S, Lee F. Earthquake-induced bending moment in fixed-head piles
in soft clay. Geotechnique 2014;64(6):431–46.
[16] Meymand P. Shaking table scale model test of nonlinear soil-pile-superstructure
interaction in soft clay. Berkeley: University of California; 1998. Ph.D. Thesis.
[17] Zhang L, Goh S, Liu H. Seismic response of pile-raft-clay system subjected to a long-
duration earthquake: centrifuge test and finite element analysis. Soil Dynam Earthq
Eng 2017;92:488–502.
[18] Zhang L, Goh S, Yi J. A centrifuge study of the seismic response of pile–raft systems
embedded in soft clay. Geotechnique 2017;67(6):479–90.
[19] Saha R, Haldar S, Dutta SC. Influence of dynamic soil-pile raft-structure
interaction: an experimental approach. Earthq Eng Eng Vib 2015;14(4):625–45.
10
Soil Dynamics and Earthquake Engineering 145 (2021) 106712
[20] Ayothiraman R, Matsagar V, Kanaujia V. Influence of superstructure flexibility on
seismic response pile foundation in sand. In: Proceedings of the 15th world
conference on earthquake engineering; 2012. Lisbon.
[21] De Sanctis L, Maiorano R, Aversa S. A method for assessing kinematic bending
moments at the pile head. Earthq Eng Struct Dynam 2010;39(10):1133–54.
[22] Maheshwari B, Truman K, El Naggar M, Gould P. Three-dimensional nonlinear
analysis for seismic soil-pile-structure interaction. Soil Dynam Earthq Eng 2004;24
(4):343–56.
[23] Mohanty P, Dutta S, Bhattacharya S. Proposed mechanism for mid-span failure of
pile supported river bridges during seismic liquefaction. Soil Dynam Earthq Eng
2017;102:41–5.
[24] Itasca Consulting Group. Fast Lagrangian analysis of continua. 2016. Minneapolis,
Minnesota.
[25] Bolouri Bazaz H, Akhtarpour A, Kharaghani S. A study on the efficiency of a capped
hardening elasto-plastic model for soft clays. Soils Found 2019;59(1):122–35.
[26] Chaudhry A. Effects of applied stress path on the stress–strain behaviour and
strength characteristics of soft nogn ngoo hao clay. 1975.
[27] Ratananikom W, Likitlersuang S, Yimsiri S. An investigation of anisotropic elastic
parameters of Bangkok Clay from vertical and horizontal cut specimens.
Geomechanics Geoengin: Int J 2012;8(1):15–27.
[28] Likitlersuang S, Kyaw K. Shear wave velocity correlations for the Bangkok subsoil.
Obras y Proyectos 2010;7:24–30.
[29] Surarak C. Geotechnical aspects of the Bangkok MRT blue line project. Queensland:
Griffith University; 2010.
[30] Likitlersuang S, Surarak C, Wanatowski D, Oh E, Balasubramaniam A. Finite
element analysis of a deep excavation: a case study from the. Bangkok MRT 2013;
53(5):756–73. October.
[31] Phienwej N. Ground movement in station excavations of Bangkok first MRT. In:
Proceedings of the 6th international symposium on tunnelling for urban
development. Shanghai, China: IS-Shanghai 2008); 2008.
[32] Suwansawat S, Chaiwonglek C, Horny U. Design aspects of NATM and cut and
cover underground stations for the Bangkok MRTA Blue Line Extension. In:
Proceedings of the 7th international symposium on tunnelling for urban
development (IS-pattaya 2007). Thailand: Pattaya City; 2007.
[33] Atkinson J. Non-linear soil stiffness in routine design. Geotechnique 2000;50(5):
487–508.
[34] Likitlersuang S, Teachavorasinskun S, Surarak C, Oh E, Balasubramaniam B. Small
strain stiffness and stiffness degradation curve of Bangkok Clays. Soils Found 2013;
53(4):498–509. August.
[35] Bhattacharya S, Orense RP, Lombardi D. Seismic design of foundations: concepts
and applications. London: ICE Publishing; 2019.
[36] Teachavorasinskun S, Thongchim P, Lukkunaprasit P. Shear modulus and damping
of soft Bangkok Clays. Can Geotech J 2002;39:1201–8.
[37] Akhtarpour A, Khodaii A. Nonlinear numerical evaluation of dynamic behavior of
an asphaltic concrete core rockfill dam (A case study). JSEE 2009;11(3):143–53.
[38] Ohtori Y, Christenson R, Spencer B, Dyke S. Benchmark control problems for
seismically excited nonlinear buildings. J Eng Mech 4 2004;130(4):366–85.
[39] Katsanos E, Sextos A, Manolis G. Selection of earthquake ground motion records: a
state-of-the-art review from a structural engineering perspective. Soil Dynam
Earthq Eng 2010;30(4):157–69.
[40] Yeganeh N, Fatahi B. Effects of choice of soil constitutive model on seismic
performance of moment-resisting frames experiencing foundation rocking
subjected to near-feld earthquakes. Soil Dynam Earthq Eng 2019;121:442–59.
[41] McCormac J, Csernak S. Structural steel design. Pearson Education; 2012.
[42] Yeganeh N, Bolouri Bazaz J, Akhtarpour A. Seismic analysis of the soil–structure
interaction for a high rise building adjacent to deep excavation. Soil Dynam Earthq
Eng 2015;79(A):149–70. December.
[43] Naeini A, Choobbasti A, Saadati M. Seismic behaviour of pile in three-layered soil
(case study: babol City Center Project). Arabian Journal of Geosciences 2013;6
(11):4487–97.
[44] Chen C, Martin G. Soil–structure interaction for landslide stabilizing piles. Comput
Geotech 2002;29(5):363–86.
[45] Kramer SL. Geotechnical earthquake engineering. New Jersey: Prentice Hall; 1996.
[46] Foutch DA, Yun S-Y. Modeling of steel moment frames for seismic loads. J Constr
Steel Res 2002;58:529–64.
[47] Tabatabaiefar HR, Samali B, Fatahi B. Effects of dynamic soil-structure interaction
on inelastic behaviour of mid-rise moment resisting buildings on soft soils. In:
Australian earthquake engineering society conference; 2010. Perth.
[48] Fhwa-Nhi-14-007. Geotechnical engineering circular No. 7 soil nail walls -
reference manual. Washington: Federal Highway Administration; 2015.
[49] García JA. Soil structure interaction in the analysis and seismic design of reinforced
concrete frame buildings. In: The 14th world conference on earthquake
engineering; 2008. Beijing, China.
[50] Khosravikia F, Mahsuli M, Ghannad MA. Probabilistic evaluation of 2015 NEHRP
soil-structure interaction provisions. J Eng Mech 2017;143(9).
[51] Saad G, Saddik F, Najjar S. Impact of soil structure interaction on the seismic
design of reinforced concrete buildings with underground stories. Lisboa: 15
WCCE; 2012.
[52] Chandran J N, Rajan A, Syed S. Seismic analysis of building with underground
stories considering soil structure interaction. International Journal of Emerging
Technology and Advanced Engineering 2014;4(11):112–7.
[53] Singh V, Mala K. Effect on seismic response of building with underground storey
considering soil structure interaction. Int J Eng Res Technol 2016;5(6):163–7.
[54] Saha R, Dutta SC, Haldar S, Kumar S. Effect of soil-pile raft-structure interaction on
elastic and inelastic seismic behaviour. Structure 2020;26:378–95.