Engineering Structures 242 (2021) 112550

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Nonlinear study on the structure-soil-structure interaction of seismic

response among high-rise buildings
Hui Long a, Zicheng Wang b, Chunshun Zhang c, *, Haiyang Zhuang d, Wenzhao Chen a,
Cheng Peng a
a
School of Civil Engineering, University of South China, Hengyang 421001, China
b
School of Civil Engineering, Central South University, Changsha 410075, China
c
Department of Civil Engineering, Monash University, VIC 3800, Australia
d
Institute of Geotechnical Engineering, Nanjing Tech University, Nanjing 210009, China

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

Keywords: Based on the 1–3 identical high-rise buildings with pile-raft foundation, a series of building-site two-dimensional
Dynamic structure-soil-structure interaction finite element analysis models with different building spacings is established. Then, considering the nonlinear
Nonlinear characteristics of soil and structure, the static-dynamic coupled numerical simulations are conducted to study the
Seismic response
structure-soil-structure interaction (SSSI) under different ground motion. The results show that: 1) Adjacent high-
Static-dynamic coupled numerical simulation
High-rise building
rise buildings have the most unfavorable impact on the peak acceleration of the 1/4–1/2 height floors of the
middle building, and increase the normalized response acceleration spectra value of acceleration near the second
natural period of the structure; 2) Adjacent high-rise buildings have the most significant impact on the inter-story
drift of the top, bottom and basement of the middle building; 3) Adjacent high-rise buildings redistribute the total
horizontal shear force of the pile top of the middle building among the pile foundations, so that the peak value of
the pile top shear force of each pile changes significantly, which is extremely unfavorable to the seismic safety of
the structure. It should be taken seriously in the seismic design of structures. In addition, the structure spacing
that results in favorable or unfavorable effects is discussed, and the reasons for the impact are analyzed.

1. Introduction that the mass and spacing of the foundation have a significant influence
on SSSI. Liang et al. [4,5] also divide the site into soil and bedrock and
The seismic safety of densely constructed urban areas is critical. To embed the soil layer’s semi-circular arc-shaped foundation. Considering
accurately assess its seismic performance and carry out structural the impact of the seismic wave incident angle, the dynamic SSSI of the
seismic design, it is necessary to study the dynamic structure-soil- two buildings is studied. The results show that the structure’s peak
structure interaction (SSSI) of multiple buildings under ground mo­ seismic responses significantly reduced when the distance between
tion. The studies of dynamic SSSI between structures mainly use structures is very close.
boundary element method [1–5], finite element–boundary element The boundary element method cannot build models of complex
coupling method [6–9], finite element method [9–21] or physical geometric shapes, and it is also difficult to simulate the nonlinear me­
experiment method [22–29]. There are also some studies using the chanical properties of materials. In contrast, finite element methods can
discrete lumped parameter model method [30–33]. Karabalis and easily solve the above problems. However, the finite element method
Huang [1,2] regard the site as a homogeneous isotropic linear elastic requires a more considerable amount of calculation than the boundary
half-space, assuming that two rigid foundations are located on the element method. To combine the advantages of the two methods, a finite
ground, using the boundary element method to study the dynamic SSSI element–boundary element coupling method is derived. The coupling
of adjacent structures. It shows that adjacent structures will increase the method uses the finite element method to simulate the superstructure,
natural frequency of the structure-soil-structure system when the dis­ foundation, near-field soil, and use the boundary element to simulate the
tance between structures is very close. Afterwards, Karabalis et al. [3] far-field soil. Lin et al. [6] use the finite element–boundary element
layer the site into soil and bedrock and do similar studies. They believe coupling method to analyze adjacent foundations’ dynamic interaction.

* Corresponding author.
E-mail address: [email protected] (C. Zhang).

https://doi.org/10.1016/j.engstruct.2021.112550
Received 3 November 2020; Received in revised form 23 April 2021; Accepted 14 May 2021
Available online 29 May 2021
0141-0296/© 2021 Elsevier Ltd. All rights reserved.
H. Long et al. Engineering Structures 242 (2021) 112550

Finite element simulation is used for the foundation and near-field soil, shear structure building and the frame structure building are adjacent,
while the far-field soil is assumed to be a semi-infinite elastic body. The the seismic response of the frame-shear structure building will decrease.
influencing factors have been studied, such as the distance between The SSSI effect can be ignored when the building distance is greater than
foundations, the arrangement direction of adjacent foundations, and 30 m (1 time the building width). Bolisetti et al. [21] used linear analysis
whether or not the foundation is embedded on the dynamic SSSI. Wang code SASSI and nonlinear finite-element analysis code LS-DYNA to study
et al. [7] analyze the dynamic SSSI between two buildings and find that the dynamic SSSI of low-to medium-rise buildings, and compared the
the SSSI effect is more evident near the structure’s natural frequency. numerical results with the centrifuge experiment results. The results
The SSSI effect decreases with the increase of the structure spacing and show that SSSI does not cause any significant change in structural
excitation loading depth. Lehmann et al. [8] proposed a numerical seismic response, but does influence a deep foundation’s peak response.
model based on the symmetric Galerkin boundary element method, Generally speaking, the existing finite element analysis mainly focuses
using the finite element to simulate the high-rise structure, and using the on the influence of the structure spacing, ground motion characteristics,
boundary element method to simulate the soil as a three-dimensional structure arrangement mode, foundation form on the SSSI. However, in
viscoelastic domain. In the work of Álamo et al. [9], the dynamic SSSI these analyses, the nonlinear characteristics of the structure and the soil
effect of the seismic responses between three adjacent structures has are rarely considered simultaneously. The influence of the soil-structure
been studied. The obtained results show that the type of wave and its system’s initial static stress on the dynamic response of the model is
angle of incidence greatly influence the SSSI effects. Moreover, SSSI rarely considered, and the most straightforward binding contact is used
effect between structures with close natural frequency is more signifi­ for the contact relationship between soil-structure.
cant. The methods mentioned above still have the problems that it is Centrifuge and shaking table tests are also common methods to study
difficult to describe the nonlinearity of the soil accurately and simulate dynamic SSSI effects. Mason et al. [22] conducted centrifuge tests on
the complex multi-layer soil. two frame structures and believed that whether the SSSI effect is
With the development of computer technology, the finite element harmful depends on the ground motion and structural system. The re­
method (FEM) is more widely used to study dynamic SSSI effects, as FEM sults show that, regardless of the geometric layout of the structure, SSSI
is more accurate in describing structure geometry, material nonlinear cause the peak foundation bending moment and foundation shear force
mechanical properties, and soil-structure contact mechanical behavior. to increase, and the SSSI effect is most significant under low-to-medium
Dou et al. [10] performed a two-dimensional explicit finite element intensity ground motion. Knappett et al. [25] combined the centrifuge
analysis on two adjacent 6-story and 21-story buildings. It is believed tests and finite element simulation to study an independent structure’s
that the impact of high-rise buildings on adjacent multi-story buildings seismic performance and two adjacent structures under a series of strong
is very significant, and the impact decreases as the distance between the ground motions. They concluded that the SSSI may reduce or increase
buildings increases. Zhang and Chen [11] used the equivalent lineari­ structural displacement and settlement, which mainly depends on
zation method (ELM) to describe the foundation soil’s nonlinearity. adjacent structures’ characteristics. In all cases where adjacent struc­
They simplified the superstructure to a series of particles, and studied tures exist, a significant increase in permanent rotation (structure tilt) is
the seismic response of two adjacent 25-story high-rise building with observed. The centrifuge tests of Hayden et al. [26] were carried out on
pile box foundations in different deep and weak site and different isolated structures and adjacent structures on the ground with varying
ground motions. Rahgozar et al. [12] also used ELM to describe the soil’s degrees of liquefaction. Adjacent structures tend to lean toward each
nonlinearity and study the seismic responses of two 15-story and 30- other, and the settlement is less than an isolated structure. The physical
story steel-structured buildings with pile foundations. The results constraints imposed by adjacent foundations limit the soil’s lateral
show that adjacent buildings’ SSSI on the flexible soil depends on their movement and reduce the total settlement of adjacent structures. Ngo
spacing, which increases the dynamic displacement of the building and et al. [27] conducted centrifuge tests on two structures with different
reduces the foundation shear force. A similar study by Ghandil et al. [13] masses, and the results show that the total power spectral density of
indicated that if the site is large enough, the SSSI effect should be structures with larger masses decreased, and the response of structures
considered when the distance between buildings is less than half the with smaller masses was increased. The mass ratio and distance between
building’s width. Besides, buildings with a larger mass are more the two structures are important parameters to control the SSSI effect.
affected. Yahyai et al. [14] used a linear elastic model to describe the Some scholars also use shaking table tests to study the dynamic SSSI
structure and soil, and study the dynamic SSSI between two adjacent 32- between adjacent structures, such as Aldaikh et al. [28] and Chan­
story high-rise buildings on three soil types: soft clay, sandy gravel, and drakala [29]. Another method is to use a discrete lumped parameter
dense sandy gravel. The response law of natural period, base shear force model to study the interaction between the foundations and the site’s
and floor displacement was investigated. Sushma et al. [15] studied the half-space. It is common to assume that the adjacent structures, or the
seismic response of 2–3 pile foundation buildings or rigid foundation structure and the soil are connected by a frequency-independent spring
buildings fixed to the ground, and believed that the middle building [30–33].
caused more displacement because of absorbing seismic waves. Yue In general, the existing numerical studies have insufficiently
et al. [16] analyzed the dynamic SSSI of the adjacent nuclear power considered the nonlinear characteristics of the soil-structure system.
plant and machine room integrated model with different foundations Most of them use equivalent linear methods or elastic materials to
under different seismic waves. Wang and Lou [17,18] studied the dy­ simulate the soil-structure system’s mechanical characteristics. More­
namic SSSI between the subway station and adjacent building, and over, in the analysis, the influence of the stress–strain field of the soil-
analyze the factors that influence the SSSI, such as direction and fre­ structure system under the initial static state is rarely considered. The
quency spectrum of the ground motion, shear wave speed and damping stress–strain state of the soil-structure system under the initial static
ratio of the soil layer, foundation form and depth, height of the super­ state has an essential influence on its seismic response. In order to
structure. The results show that the structure’s arrangement direction simulate the seismic response characteristics of the site-adjacent build­
and vibration direction have the most significant influence on the dy­ ing system more realistically, and study the dynamic SSSI effect between
namic SSSI. Utilizing the Davidenkov model to describe soil nonline­ adjacent structures, this paper will use the finite element software
arity, Li et al. [19] analyzed the seismic response of two elastic shear ABAQUS to perform nonlinear static-dynamic coupling numerical
wall structures. The results show that the seismic response of the analysis on the seismic response of a single, two or three identical high-
structure will increase when the SSSI effect is considered. After that, Gan rise structures in the site. In the analysis, the nonlinearity of the struc­
et al. [20] studied the seismic response of three adjacent high-rise ture and the soil is considered simultaneously, and the influence of the
buildings, considering the distance between the structures, the struc­ static force on the soil-structure system is considered. Under the
ture type, and the structure’s height. It is believed that when the frame- nonlinear condition, the study of the dynamic SSSI effect is more

2
H. Long et al. Engineering Structures 242 (2021) 112550

realistic when adjacent structures are on one or both sides of the

structure. Considering the computational cost and efficiency, the 2D Roof
models are selected to conduct a comparative analysis study considering
these multiple influencing factors. The characteristics of the influence on
the middle structure’s seismic response provide a reference for the
seismic design under similar engineering conditions.

2. Model introduction

2.1. Structure

The typical frame structure with 1 floor underground and 12 floors

above ground is taken as the research object. This modelled structure is
simplified from the actual engineering structure, as in China, the use of a
frame structure of about 10 floors in office buildings is representative.
The basement floor height of the structure is 4.2 m, the standard floor Raft foundation
Basement
height above ground is 3 m, and the total height above ground is 36 m.
The structure has 3 horizontal spans, the side span column center dis­ Pile foundation
tance is 6 m, the middle span column center distance is 2.4 m, the total
width is 15 m. The structure has 6 longitudinal spans, the column center
distance is 6 m, and the total length is 36.6 m. The structural frame
column section size is 0.6 m × 0.6 m, the beam section size is 0.25 m × Fig. 2. Structural section.
0.6 m, and the floor and roof slab thickness is 0.1 m. Except the mid-span
beams, all beams have 0.2 m thick infill walls (including longitudinal
the Chinese Code for Seismic Design of Buildings (GB50011-2010). The
beams) with a density of 1.8 t/m3. The structure adopts a pile-raft
equivalent shear wave velocity within a depth of 20 m from the site’s
foundation with one pile (size of 0.6 m × 0.6 m and length of 7.9 m)
ground surface is 198.47 m/s, and the soil thickness is 30 m. Below it is
under each column. Also, the raft thickness is 0.4 m. For convenience,
the bedrock with shear wave velocity greater than 500 m/s, and the
the columns are numbered C1-C4 from left to right, and the pile foun­
bottom of the soil layer is the ground motion input surface. It is assumed
dations are numbered Z1-Z4 from left to right. The structural plan and
that the site is a horizontally layered site, with no change in the soil
structural section are shown in Fig. 1 - Fig. 2. The modal analysis of the
layers in front and behind the lateral boundary. Since no artificial
structure in ABAQUS shows that the first 3 natural frequencies are f1 =
viscoelastic boundary is used, in order to reduce the influence of the side
0.97636 Hz, f2 = 3.0615 Hz, and f3 = 5.5867 Hz, respectively. Note that
boundary reflection wave in the dynamic analysis, the site scope is
in the modal analysis, it is assumed that the basement floor (Raft fou­
enlarged to not less than 5 times the site thickness (150 m) [18] from the
dation) is the anchor end of the superstructure, constraining the
outer edge of the structure to the side boundary of the site under all
displacement and rotation of the basement floor and pile foundation.
working conditions. Based on the above reason, the total width of the
The mode figure is shown in Fig. 3. A widely-employed concrete damage
site is enlarged to 510 m. Zhuang’s [35,36] viscoplastic memory-nested
constitutive model [34] is used to describe the nonlinear characteristics
surface constitutive model is utilized to describe the soil’s nonlinear
of structural materials.. The concrete strength grade is C30, and its
characteristics. Based on the generalized plasticity theory of soil, the
mechanical parameters are shown in Table 1.
constitutive model employs the hardening modulus field combined with
isotropic and kinematic hardening. Also, the viscoplastic dynamic
2.2. Soil constitutive incremental model is established in the form of total stress.
The model determines the change rule of the yield surface by memo­
An engineering site along Nanjing Metro Line 1 is selected as the rizing the reverse loading surface, the failure surface and the initial
construction site. The site category belongs to the Class II site defined in

Fig. 1. Structural plan (mm).

3
H. Long et al. Engineering Structures 242 (2021) 112550

Mode 1: Freq= 0.97636Hz Mode 2: Freq= 3.0615Hz Mode 3: Freq= 5.5867Hz

Fig. 3. First 3 modes of structure.

Table 1 Failure
Parameters of the plastic-damage model of concrete. surface F

Parameters Value Parameters Value

σ1 Initial loading
yield surface f
Density 2500 kg/m3 Dilation angle 32.4◦ n
Elastic modulus 3.0 × Poisson’s ratio 0.18
104MPa
Limited compressive yield 24.5 MPa Initial tensile yield 2.4 Reverse loading
stress stress MPa surface fr

loading yield surface inscribed to the reverse loading surface at any O σ3

time, so the model parameters are easy to determine. The distribution of
the model memory nesting surface in the π plane and the stress space is
σ2
shown in Figs. 4 and 5. The specific establishment process is shown in
references [18,35]. The analysis assumes that the soil is saturated and Fig. 5. Memorial surfaces of viscoplastic memory nested surface constitutive
will not drain in the earthquake. The calculation parameters of the soil model in stress space.
layer are shown in Table 2.
analysis models are established with the distance to width ratio D/W =
0.25, 0.5, 1, 1.5, 2, 3, 4, 5, 6. The SSSI analysis model is shown in Fig. 6.
2.3. SSSI analysis model
2.4. Soil-structure dynamic contact relationship and element
The SSSI analysis models are divided into three categories: single
building, 2 buildings, and 3 buildings on the site. The three buildings are The dynamic contact relationship between soil and structure is
named A, B, and C from left to right. There is building B in a single shown in Fig. 7, which shows: 1) On the contact surface, the normal
building model. There are building A and building B in 2 buildings compressive stress is transferred to each other through contact con­
model. The distance between the buildings is defined as D, and the straints. When the structure is separated from the soil, the contact
building width is defined as W(W = 15 m). The multiple comparative constraint is eliminated, and the contact surface becomes the mutual
free surface of the two materials. 2) The tangent is the Coulomb friction

σ1 Initial loading
contact, assuming that the friction coefficient is 0.4. When the value of
the tangential contact shear stress is greater than the critical value,
yield surface f f slippage occurs along the contact surface.
Because the beam element cannot accurately describe the above-
Reverse floading
r mentioned contact characteristics between soil and structure, the
surface fr
θij structural element adopts a plane strain full integral solid element with a
αij1 Failure size of 0.2 m × 0.2 m, and the soil element adopts a plane strain
αij2 S ij1 surface F reduction integral solid element with a preferred size of 1 m × 1 m to
Sij2 reduce the amount of model calculation. The finite element meshing of
O the model is shown in Fig. 8, where the parameters of different soil
layers are shown in Table 2.
σ2 Sij3
σ3

Fig. 4. Memorial surfaces of viscoplastic memory nested surface constitutive

model in π plane.

4
H. Long et al. Engineering Structures 242 (2021) 112550

Table 2
Parameters of the viscoplastic model of soil.
Soil layer Thickness/ Density Shear wave velocity Internal friction Dynamic Static Poisson’s Static elastic Reference shear
name m ρ/(t/m3) Vs/(m/s) angle φ/(◦ ) Poisson’s ratio μ ratio μ0 modulus E0/MPa strain γ0

Clay 2.5 1.88 141 13.5 0.49 0.43 5 0.00032

Silty clay 6.5 1.92 185 20 0.49 0.4 10 0.00035
Silt 8 2.01 220 22 0.49 0.38 35 0.00038
Fine sand 7.5 2.03 260 27.5 0.49 0.35 60 0.0004
Silty fine 5.5 2.02 237 26 0.49 0.36 45 0.0004
sand

Building A Building B Building C

36
150

12.5
Y
30

D W=15 D
O X Soil
Bedrock 510

Fig. 6. SSSI analysis model (m).

3. Finite element analysis method

3.1. Two-dimensional equivalent method

The condition of simplifying the three-dimensional problem to the

Soil Structure Soil Structure
plane strain problem (XOY plane) is that the structure is long enough in
the vertical plane direction (Z direction), and the cross-section on the
plane does not change. Therefore, the frame columns are transformed
into equivalent walls distributed along the Z direction, and the frame
beams are transformed into equivalent plates distributed along the Z
Any pressure possible
Sliding occurs when the direction. Besides, to ensure the same mechanical properties after
when in contact.
tangential stress is greater than equivalence, the mass and stiffness of frame columns and beams need to
No pressure when no
the friction be evenly distributed to the equivalent walls and plates. It makes the
contact(Instantaneous separation)
mass ratio and stiffness ratio between soil-structure and structural
Fig. 7. Dynamic contact relationship between soil and structure. members consistent with the actual situation. For example, the section
size of the column is 0.6 m × 0.6 m, and the section size of the equivalent
continuous wall is 0.6 m × 6 m (Longitudinal column center distance is
6 m). According to the principle of equal mass and stiffness, the con­
crete’s density and modulus are reduced to 1/10 of the original after
two-dimensional equivalence. Similarly, the density and modulus of the

Plane strain full

integral solid eelement
0.2m × 0.2
0.2m
Plane strain
strai reduction
integral solid
so element
1m × 1m

Input horizontal ground motion(Normal incidence shear wave)

Fig. 8. Finite element meshing of model.

5
H. Long et al. Engineering Structures 242 (2021) 112550

concrete are reduced to 1/24 of the original. Also, assuming that the 5. Results
uniformly distributed variable load on the floor (roof) is 2kN/m3, the
floor’s self-weight, the uniformly distributed variable load on the floor, For the convenience of description, take One, T, S to represent the
and the infill wall’s self-weight are applied to the beams and columns by number of buildings in the model as 1, 2, 3, and use the distance-to-
increasing the density of beams and columns. width ratio D/W as the suffix to name various working conditions. For
example, the working calculation condition is T-0.5 when the distance-
3.2. Static-dynamic coupling numerical analysis to-width ratio of two buildings is 0.5.

The boundary conditions of the model are shown in Fig. 9. In the 5.1. Acceleration response
research, static-dynamic coupling analysis will be performed on the
model. Since static analysis and dynamic analysis have different re­ 5.1.1. Response of peak acceleration
quirements for boundary conditions, the actual analysis is carried out in Define the influence coefficient βa of acceleration as follows:
three steps, and the boundary conditions are changed during the anal­
ysis process. βa =
amax,2 (amax,3 )
(1)
Step 1: Static analysis. The static analysis of the model is performed amax,1
to obtain the initial stress–strain field of the entire model under the static In the formula, amax,1 indicates the horizontal peak acceleration
state and complete the ground stress balance to keep the site strain less response of each floor of the building when there is only building B in
than 10-5. At this time, the model’s boundary condition is the normal the model, amax,2 and amax,3 respectively indicate the peak horizontal
constraint of the lateral boundary and the normal constraints of the acceleration response of each floor of building B when there are 2
bottom boundary. buildings and 3 buildings in the model.
Step 2: Boundary condition conversion. Since the dynamic analysis Tables 3–5 show the horizontal peak acceleration response of the
input is horizontal ground motion, the normal constraint of the lateral raft, floor 1, floor 4, floor 5, floor 7, floor 10, and roof of building B
boundary of the model needs to be cancelled. In order to keep the model without adjacent structure under different ground motions, and the in­
stress–strain field unchanged, after canceling the normal constraint of fluence coefficient βa of each floor of building B when there are adjacent
the lateral boundary, the constraint reaction force of the lateral structures. Fig. 11 shows the comparison of βa on floor 4, floor 5, floor 6,
boundary node obtained in the first step is applied to the constraint node floor 7 and roof under different adjacent structures.
in the form of a concentrated force to complete the conversion of the Based on the given structure and site conditions, under the influence
boundary condition. of adjacent buildings, the peak acceleration response characteristics of
Step 3: Dynamic analysis. Apply a horizontal ground motion to the each floor of the building B in the middle are as follows:
bottom of the model for dynamic analysis. Under all working conditions,
the distance between the outer edge of the outer building and the lateral 1. In general, the adjacent structures impact the peak acceleration of
boundary of the site shall not be less than 5 times the soil thickness [18] building B. Under LP waves’ action, the most unfavorable effect is an
to reduce the artificial boundary’s influence on the dynamic response of increase of 10.1% (S-3 working condition, floor 5), and the most
the structure. favorable effect is a decrease of 8.8% (S-025 working condition, floor
1). Under the action of NJ wave, the most unfavorable effect is an
4. Input ground motion increase of 13.5% (S-3 working condition, floor 5), and the most
favorable effect is a decrease of 12.8% (S-025 working condition,
The input ground motion is a horizontal ground motion whose vi­ floor 6, amax,1 =1.403 m/s2, amax,3 =1.224 m/s2). Under the WL wave
bration direction is parallel to the direction of the building’s cross- action, the most unfavorable effect is an increase of 14.3% (S-3
section (SV wave). According to the code for seismic design of build­ working condition, floor 5), and the most favorable effect is a
ings (GB 50011–2010, China), no less than two actual seismic records decrease of 13.8% (T- 025 working condition, floor 1).
and one artificial acceleration time history curves can be selected as the 2. The structure spacing has a significant impact on the SSSI. When the
input ground motion for time history analysis. Referring to the code, far- spacing is very close (D/W = 0.25), in most cases, adjacent structures
field seismic wave Loma Prieta wave (starting now referred to as LP tend to be beneficial to the peak acceleration response of each floor
wave), near-field seismic wave Wolong wave (starting now referred to as of building B. As the spacing increases, in most cases, adjacent
WL wave) and Nanjing artificial seismic wave (starting now referred to structures’ influence gradually tends to be unfavorable when 1 ≤ D/
as NJ wave) are selected as input ground motions. The peak accelera­ W ≤ 4, and it is the most unfavorable when D/W = 3. The influence
tions of these seismic waves are all modulated to 1 m/s2. The ground of adjacent structures on the peak acceleration of most building B
motion duration is 30 s, and the acceleration time history curve and floors is within 5% when D/W ≥ 5, which can be ignored.
acceleration response spectrum are shown in Fig. 10. 3. In most cases, adjacent structures have a beneficial effect on the first
floor of building B’s peak acceleration response. The adjacent
structure has the most adverse impact on the peak acceleration

Site Site Site

Constraint
Constraint reaction Acceleration

Step 1: Static analysis Step 2: Boundary condition conversion Step 3: Dynamic analysis

Fig. 9. Transformation of model boundary conditions.

6
H. Long et al. Engineering Structures 242 (2021) 112550

0.12 0.12

0.08 0.08

Acceleration(g)

Acceleration(g)
0.04 0.04

0.00 0.00

-0.04 -0.04

-0.08 -0.08

-0.12 -0.12
0 5 10 15 20 25 30 0 5 10 15 20 25 30
Time(s) Time(s)

(a) (b)
0.12
0.4

Response Acceleration(g)
NJ wave
0.08 LP wave
0.3 WL wave
Acceleration(g)

0.04

0.00 0.2

-0.04
0.1
-0.08

-0.12 0.0
0 5 10 15 20 25 30 0 1 2 3 4
Time(s) Period(s)
(c) (d)
Fig. 10. (a) Acceleration time-history of Nanjing artificial earthquake wave; (b) Acceleration time-history of Loma Prieta earthquake wave; (c) Acceleration time-
history of Wolong earthquake wave; (d) Response acceleration of three earthquake waves.

Table 3
Peak acceleration response and βa of each layer of Building B under LP wave.
Position amax,1 /(m/s− 2) βa

One T-0.25 T-0.5 T-1 T-1.5 T-2 T-3 T-4 T-5 T-6

Roof 2.408 0.982 1.002 1.037 1.016 1.022 1.027 1.032 1.027 1.015
F10 1.882 0.961 0.969 0.997 0.994 0.979 0.992 0.996 0.984 0.983
F7 1.777 0.980 1.000 0.998 1.027 1.027 1.017 1.008 0.992 1.006
F5 2.122 0.984 1.006 1.016 1.048 1.033 1.038 1.021 1.008 1.014
F4 2.145 0.989 1.006 1.028 1.056 1.037 1.049 1.029 1.017 1.018
F1 2.116 0.959 0.946 0.963 0.994 0.985 0.983 1.018 1.003 0.998
Raft 2.115 0.951 0.941 0.960 0.970 0.987 0.982 1.012 0.994 0.995
Position S-0.25 S-0.5 S-1 S-15 S-2 S-3 S-4 S-5 S-6
Roof 0.979 1.028 1.065 1.058 1.061 1.085 1.049 1.041 1.033
F10 0.951 0.960 0.982 0.980 0.989 0.990 0.974 0.979 0.981
F7 0.929 0.975 1.007 1.043 1.071 1.095 1.057 1.010 1.003
F5 0.957 1.005 1.050 1.071 1.083 1.101 1.089 1.048 1.028
F4 0.969 1.014 1.053 1.064 1.074 1.100 1.099 1.069 1.043
F1 0.912 0.921 0.947 0.962 0.970 0.985 0.997 1.029 1.022
Raft 0.944 0.945 0.957 0.982 0.994 0.988 1.006 1.015 1.021

response of the floor 4-floor 7 located at 1/4–1/2 of the height of 5.1.2. Normalized response acceleration spectra
building B when 1 ≤ D/W ≤ 4. Also, the peak acceleration response Considering the damping ratio is 0.05, the normalized response ac­
of other floors of adjacent structures is related to the input ground celeration spectra of building B under different working conditions are
motion, and the same floor may have beneficial or unfavorable ef­ studied to explore adjacent structures’ influence on their seismic
fects under different ground motions. response spectrum characteristics.
4. As shown in Fig. 11, the calculation results of building B with adja­ The normalized response acceleration spectra for structures floor 1,
cent structures on one side or both sides are compared. As 1 ≤ D/W floor 5, floor 7, and building B roof are shown in Fig. 12 when the
≤ 4, compared with adjacent structures on one side, the peak ac­ number and distance of adjacent structures are different. Since the SSSI
celeration responses are more unfavorable to the floor 4-floor 7 and has little impact on the peak acceleration response of building B when
roof when there are adjacent structures on both sides. Whether the D/W ≥ 5, the figures only show the calculation results when D/W =
impact on other floors is beneficial is related to the input ground 0.25, 1, 3, 4.
motion. The results show that the overall trend of building B’s acceleration
spectrum has not changed significantly with the changes in the number
and spacing of adjacent structures. However, some local influences are

7
H. Long et al. Engineering Structures 242 (2021) 112550

Table 4
Peak acceleration response and βa of each layer of Building B under NJ wave.
Position amax,1 /(m/s− 2) βa

One T-0.25 T-0.5 T-1 T-1.5 T-2 T-3 T-4 T-5 T-6

Roof 2.583 0.941 0.948 0.982 0.989 0.987 0.995 0.991 0.991 0.989
F10 1.753 1.010 1.006 0.994 1.004 1.003 1.000 0.923 1.004 1.000
F7 1.333 0.930 0.975 1.023 1.044 1.058 1.048 1.021 0.998 1.031
F5 1.372 0.989 1.004 1.028 1.050 1.063 1.075 1.035 1.033 1.028
F4 1.480 0.938 0.946 0.980 0.982 1.015 1.023 1.076 1.044 1.045
F1 1.871 0.920 0.943 0.963 0.994 1.018 0.957 0.983 1.029 1.009
Raft 1.663 0.932 0.948 0.961 1.015 1.015 0.964 0.976 1.002 1.001
Position S-0.25 S-0.5 S-1 S-15 S-2 S-3 S-4 S-5 S-6
Roof 0.963 0.963 1.020 1.026 1.008 0.995 0.990 1.016 0.997
F10 1.017 1.014 1.009 0.996 1.000 0.993 1.005 1.025 1.025
F7 0.923 0.961 1.029 1.070 1.103 1.060 0.992 1.009 1.056
F5 0.930 0.986 1.047 1.099 1.127 1.135 1.071 1.055 1.051
F4 0.963 0.966 1.006 1.026 1.041 1.081 1.055 1.061 1.055
F1 0.925 0.927 0.991 0.978 1.000 0.982 0.966 0.973 1.030
Raft 0.931 0.927 0.996 1.000 1.007 0.982 1.037 0.996 1.025

Table 5
Peak acceleration response and βa of each layer of Building B under WL wave.
Position amax,1 /(m/s− 2) βa

One T-0.25 T-0.5 T-1 T-1.5 T-2 T-3 T-4 T-5 T-6

Roof 2.303 0.967 0.977 0.998 1.011 1.026 1.037 0.979 1.050 1.039
F10 1.222 1.011 1.007 1.001 1.012 1.010 1.018 0.919 1.043 1.036
F7 0.987 0.956 0.962 1.033 1.056 1.070 1.065 1.079 1.031 1.050
F5 1.549 0.963 0.970 1.029 1.040 1.044 1.077 1.057 1.041 1.051
F4 1.697 0.956 0.961 1.010 1.045 1.053 1.081 1.041 1.038 1.043
F1 2.031 0.862 0.867 0.916 0.947 0.950 0.930 0.905 0.961 0.986
Raft 1.569 0.992 0.996 1.025 1.070 1.065 1.051 1.030 1.054 1.078
Position S-0.25 S-0.5 S-1 S-15 S-2 S-3 S-4 S-5 S-6
Roof 1.021 1.005 1.012 1.052 1.062 1.093 1.102 1.065 1.044
F10 1.093 1.046 1.021 1.041 1.038 1.063 1.067 1.057 1.057
F7 0.925 0.944 1.030 1.092 1.143 1.113 1.061 1.042 1.039
F5 0.957 0.957 1.036 1.069 1.124 1.085 1.077 1.044 1.045
F4 0.963 0.970 1.029 1.073 1.112 1.115 1.072 1.045 1.042
F1 0.879 0.883 1.008 0.977 0.996 1.000 0.967 0.941 0.932
Raft 1.013 1.036 1.066 1.104 1.113 1.132 1.096 1.068 1.041

still produced as follows:

Δumax,2 (Δumax,3 )
βu = (2)
Δumax,1
1. On the whole, whether the influence of adjacent structures is bene­
ficial or not, the influence on the normalized response acceleration In the formula, Δumax,1 indicates the inter-story drift response of each
spectra with adjacent structures on both sides is slightly more sig­ floor of the building when there is only building B in the model. Δumax,2
nificant than that on one side. and Δumax,3 , respectively, indicate the inter-story drift response of each
2. When D/W = 0.25, adjacent structures’ influence on the normalized floor of building B when there are 2 buildings and 3 buildings in the
response acceleration spectra of building B is not consistent. It can be model.
considered that the SSSI between the structures is too complicated The inter-story drift ratio response of each floor of the building B
when the structure spacing is very close, so there is no clear rule for under different spacings and the maximum/minimum βu of the above-
the influence of vibration spectrum characteristics. ground or underground parts of the structure are shown in Fig. 13.
3. When 1 ≤ D/W ≤ 4, regardless of the structural floor’s location and The influence of adjacent structures on the inter-story drift ratio of
the input ground motion, the adjacent structure causes the normal­ the above-ground and underground parts of building B is as follows:
ized response acceleration spectra of building B to increase in the
local area centered on period = 0.3 s-0.35 s. Besides, the impact 1. When adjacent structures on both sides, the influence on the inter-
range of different floors under different input ground motions is not story drift ratio is more significant than when adjacent structures
entirely the same. The specific increase range under each working on one side. Furthermore, the influence of SSSI on the inter-story
condition is shown in Fig. 12. Considering that the second natural drift ratio of some floors cannot be ignored when D/W = 5–6. For
period of the structure is 0.327 s (f2 = 3.0615 Hz), the SSSI increases the above-ground part of the structure, both the maximum or mini­
the seismic response near the second natural period of building B mum values appear on floor 12 or floor 1-floor 2. The SSSI has the
when 1 ≤ D/W ≤ 4, which may increase the effect of second-order most significant impact on the inter-story drift ratio response at the
mode one seismic response of similar structures. The impact should top and bottom of the structure.
be taken seriously. 2. Inter-story drift ratio response under far-field LP wave. For the
above-ground part of building B, almost all floors’ inter-story drift
ratio response has been reduced. It can be considered that the
5.2. Response of inter-story drift
adjacent structure has a favorable influence on the inter-story drift
response of the above-ground part of building B, and it is most
Define the influence coefficient βu of inter-story drift as follows:

8
H. Long et al. Engineering Structures 242 (2021) 112550

1.15 1.15
LP wave NJ wave

1.1 1.1

1.05 1.05
a

a
1 1

0.95 0.95
1 1.5 2 3 4 1 1.5 2 3 4
D/W D/W
T-F4 T-F5 T-F6 T-F7 T-Roof T-F4 T-F5 T-F6 T-F7 T-Roof
S-F4 S-F5 S-F6 S-F7 S-Roof S-F4 S-F5 S-F6 S-F7 S-Roof

1.15
WL wave

1.1

1.05
a

0.95
1 1.5 2 3 4
D/W
T-F4 T-F5 T-F6 T-F7 T-Roof
S-F4 S-F5 S-F6 S-F7 S-Roof

Fig. 11. Comparison of βa for different adjacent building numbers.

favorable to the top of the building. Furthermore, the adjacent lateral deformation around the basement1of building B will increase. As
structure increases the inter-story drift ratio response of basement 1, a result, the inter-story drift ratio of the basement1of building B in­
and the effect is significant only when D/W = 0.25 (βu max = 1.282). creases. Especially when the structure spacing is minimal, the soil be­
When D/W ≥ 1, adjacent structures’ influence is small and does not tween the structures absorbs little deformation. This is why the adjacent
change much with the spacing. structures significantly increase the inter-story drift ratio response of the
3. Inter-story drift ratio response under near-field WL wave. For the basement1of building B when the structure spacing is very close.
above-ground part of building B, regardless of the number of adja­ Besides, the structure’s inter-story drift ratio response is small under
cent buildings, the adjacent structure decreases the inter-story drift the action of WL waves. The lateral soil deformation in contact with
ratio responses of floor 4-floor 12 when 0.25 ≤ D/W ≤ 2, and the basement 1of the adjacent structure may be greater or less than the soil’s
increase inter-story drift ratio responses when 3 ≤ D/W ≤ 6. Besides, deformation at the same location without adjacent structures. As a
the inter-story drift ratio responses of the floor 1-floor 2 at the bot­ result, the inter-story drift ratio of basement 1 of building B increases or
tom of building B are increased by the adjacent structure. There is no decreases.
apparent influence law of inter-story drift ratio responses on base­ The foudation’s relative horizontal displacements of 3 buildings and
ment 1. a single building under NJ wave action were compared in Table 6. The
4. Inter-story drift ratio response under NJ wave. For the above-ground table shows the maximum relative horizontal displacement between the
part of building B, in most cases, the adjacent structures increase the ground surface and the soil at a depth of 4.35 m. Among them, S-A, S-C
inter-story drift ratio responses of each floor when there are on both represent the maximum relative horizontal displacement of the soil in
sides. Besides, the adjacent structure increases the inter-story drift contact with the underground part on the right side of building A or the
ratio responses of basement 1. left side of building C, and O-A, O-C represent the maximum relative
horizontal displacement of the soil at the same location corresponding to
The inter-story drift ratio response is larger under LP wave and NJ the former when there is only building B. When 0.25 ≤ D/W ≤ 5, S-A >
wave than under WL wave. Furthermore, the adjacent structure in­ O-A, S-C > O-C, which is consistent with the results of the increase in
creases the inter-story drift ratio response of basement 1 of building B inter-story drift ratio of basement 1 of building B under the influence of
under LP wave and NJ wave. The inter-story drift of the underground adjacent structures. When D/W = 6, SA < OA, and SC > OC, the inter-
part of the structure is greatly affected by the surrounding soil defor­ story drift ratio response of basement 1 of building B also increases.
mation. When the inter-story drift of adjacent structure basement 1 is This increase shows that the response results from the interaction of
significant, the lateral soil deformation in contact with it will be greater adjacent buildings A and C.
than that at the same position without adjacent structure. If the soil Also, the influence of adjacent structures on inter-story drift ratio
between the structures cannot fully absorb the deformation, the soil’s response of the above-ground part of building B depends more on the

9
H. Long et al. Engineering Structures 242 (2021) 112550

4.3 4.3
3.8 3.8

Normalized response acceleration

Normalized response acceleration

5 5
3.3 3.3
4 2.8 4 2.8
2.3 2.3
3 3
1.8 1.8
2 1.3 0.2s-0.5s 2 1.3 0.2s-0.5s
0.8 0.8
1 0.2 0.4 0.6 0.8 1 1.2 1 0.2 0.4 0.6 0.8 1 1.2
LP wave LP wave
00 1 2 3 4 00 1 2 3 4
Period(s) Period(s)
One T-0.25 T-1 T-3 T-4 One S-0.25 S-1 S-3 S-4

5 5.5
4.5 5
Normalized response acceleration

Normalized response acceleration

5 5
4 4.5
4 3.5 4
4
3.5
3
3 3 3
2.5 2.5
`

2 2
2 0.2s-0.5s 2 0.2s-0.5s
1.5 1.5
1 1 1 1
0.1 0.3 0.5 0.7 0.9 0.1 0.3 0.5 0.7 0.9
NJ wave NJ wave
00 1 2 3 4 00 1 2 3 4
Period(s) Period(s)
One T-0.25 T-1 T-3 T-4 One S-0.25 S-1 S-3 S-4

5 5
Normalized response acceleration
Normalized response acceleration

5 4.5 5 4.5
4 4
4 3.5 4 3.5
3 3
3 2.5 3 2.5
2 2
2 1.5 2 1.5
0.2s-0.55s 0.2s-0.55s
1 1
1 0.5 1 0.5
0.1 0.3 0.5 0.7 0.9 0.1 0.3 0.5 0.7 0.9
WL wave WL wave
00 1 2 3 4 00 1 2 3 4
Period(s) Period(s)
One T-0.25 T-1 T-3 T-4 One S-0.25 S-1 S-3 S-4

(a)
4.3 4.3
3.8 3.8
Normalized response acceleration

Normalized response acceleration

5 3.3 5 3.3
2.8 2.8
4 4
2.3 2.3
3 1.8 3 1.8
1.3 0.3s-0.45s 1.3 0.3s-0.45s
2 2
0.8 0.8
0.2 0.4 0.6 0.8 1 1.2 0.2 0.4 0.6 0.8 1 1.2
1 1
LP wave LP wave
00 1 2 3 4 00 1 2 3 4
Period(s) Period(s)
One T-0.25 T-1 T-3 T-4 One S-0.25 S-1 S-3 S-4

Fig. 12. Normalized response acceleration spectra. (a) Floor 1; (b) Floor 5; (c) Floor 7; (d) Roof.

input ground motion and the vibration characteristics of the structure- 5.3. Shear force response of pile foundation
soil-structure system, and the influence may be favorable or
unfavorable. Define the influence coefficient βv of shear force as follows:

10
H. Long et al. Engineering Structures 242 (2021) 112550

4.5 4.5
4 4

Normalized response acceleration

Normalized response acceleration

5 5
3.5 3.5
4 3 4 3
2.5 2.5
3 3
2 2
2 1.5 0.25s-0.3s 2 1.5 0.25s-0.3s
1 1
1 0.1 0.3 0.5 0.7 0.9 1.1 1.3 0.1 0.3 0.5 0.7 0.9 1.1 1.3
1
NJ wave NJ wave
00 1 2 3 4 00 1 2 3 4
Period(s) Period(s)
One T-0.25 T-1 T-3 T-4 One S-0.25 S-1 S-3 S-4

4.4 4.4

Normalized response acceleration

Normalized response acceleration

5 3.9 5 3.9
3.4 3.4
4 2.9 4 2.9
2.4 2.4
3 3 1.9
1.9
0.25s-0.45s 0.25s-0.45s
2 1.4 2 1.4
0.9 0.9
1 0.4 1 0.4
0.2 0.4 0.6 0.8 1 0.2 0.4 0.6 0.8 1
WL wave WL wave
00 1 2 3 4 00 1 2 3 4
Period(s) Period(s)
One T-0.25 T-1 T-3 T-4 One S-0.25 S-1 S-3 S-4

(b)
3.5 3.5
3 3
Normalized response acceleration

Normalized response acceleration

5 5
2.5 2.5
4 4
2 2
3 3
1.5 1.5
2 0.25s-0.4s 2 0.25s-0.4s
1 1
0.2 0.4 0.6 0.8 1 1.2 0.2 0.4 0.6 0.8 1 1.2
1 1
LP wave LP wave
00 1 2 3 4 00 1 2 3 4
Period(s) Period(s)
One T-0.25 T-1 T-3 T-4 One S-0.25 S-1 S-3 S-4

4.5 4.5
4 4
Normalized response acceleration

Normalized response acceleration

5 5
3.5 3.5
4 3 4 3
2.5 2.5
3 3
2 2
2 1.5 0.2s-0.4s 2 1.5 0.2s-0.45s
1 1
1 0.1 0.3 0.5 0.7 0.9 1.1 1.3 1 0.1 0.3 0.5 0.7 0.9 1.1 1.3
NJ wave NJ wave
00 1 2 3 4 00 1 2 3 4
Period(s) Period(s)
One T-0.25 T-1 T-3 T-4 One S-0.25 S-1 S-3 S-4

Fig. 12. (continued).

Vmax,2 (Vmax,3 ) In the formula, Vmax,1 indicates the peak shear force response on each
βv = (3)
Vmax,1 pile top or the peak total shear force on all piles top when there is only
building B in the model, Vmax,2 and Vmax,3 respectively indicate the peak

11
H. Long et al. Engineering Structures 242 (2021) 112550

4.5 4.5

Normalized response acceleration

4

Normalized response acceleration

4
5 3.5 5 3.5
3 3
4 4
2.5 2.5
3 2 3 2
1.5 1.5
2 0.25s-0.35s 0.25s-0.45s
1 2 1
0.5 0.5
1 0.2 0.4 0.6 0.8 1 1 0.2 0.4 0.6 0.8 1
WL wave WL wave
00 1 2 3 4 00 1 2 3 4
Period(s) Period(s)
One T-0.25 T-1 T-3 T-4 One S-0.25 S-1 S-3 S-4
(c)

4.4 4.4

Normalized response acceleration

3.9
Normalized response acceleration

3.9
5 5
3.4 3.4
4 4 2.9
2.9
3 2.4 3 2.4
1.9 0.25s-0.4s 1.9 0.25s-0.4s
2 2
1.4 1.4
0.2 0.4 0.6 0.8 1 1.2 1 0.2 0.4 0.6 0.8 1 1.2
1
LP wave LP wave
0 0
0 1 2 3 4 0 1 2 3 4
Period Period
One T-0.25 T-1 T-3 T-4 One S-0.25 S-1 S-3 S-4

4.3 4.2
Normalized response acceleration

Normalized response acceleration

3.8 3.7
5 5
3.3 3.2
4 4
2.8 2.7
3 3
2.3 2.2 0.25s-0.5s
2 0.25s-0.5s 2
1.8 1.7
0.2 0.4 0.6 0.8 1 1.2 0.2 0.4 0.6 0.8 1 1.2
1 1
NJ wave NJ wave
0 0
0 1 2 3 4 0 1 2 3 4
Period Period
One T-0.25 T-1 T-3 T-4 One S-0.25 S-1 S-3 S-4

4.3 4.3
Normalized response acceleration
Normalized response acceleration

5 3.8 5 3.8
3.3 3.3
4 4
2.8 2.8
3 2.3 3 2.3
1.8 1.8
2 2
1.3 0.35s-0.45s 1.3 0.35s-0.45s
1 0.8 1 0.8
WL wave 0.2 0.4 0.6 0.8 1 WL wave 0.2 0.4 0.6 0.8 1
0 0
0 1 2 3 4 0 1 2 3 4
Period Period
One T-0.25 T-1 T-3 T-4 One S-0.25 S-1 S-3 S-4

(d)

Fig. 12. (continued).

shear force response on each pile top or the peak total shear force on all shear force response on all piles top when there are adjacent structures.
piles top when there are 2 buildings and 3 buildings in the model. The influence of adjacent structures on the peak shear force response
Tables 7–9 show the peak shear force response on each pile top or the of the pile foundation of building B is as follows:
peak total shear force on all piles top of building B without adjacent
structure, and the influence coefficient βv of each pile top or peak total

12
H. Long et al. Engineering Structures 242 (2021) 112550

12 12
Above =1.031, T-0.25, F1 Above =0.997, S-1, F1
u,max
11 u,max
11
ground =0.908, T-0.25, F12 ground =0.815, S-0.25, F12
u,min u,min
10 10
9 9
ONE ONE
8 8
T-0.25 S-0.25
7 T-1 7 S-1
Floor

Floor
6 T-2 6 S-2
5 T-3 5 S-3
T-5 S-5
4 4
T-6 S-6
3 3
=1.282, T-0.25 =1.190, S-0.25
2 B1
u,max
2 B1
u,max

=1.012, T-5 =1.005, S-5

1 u,min
1 u,min

B1 B1
0.8 1.2 1.6 2 2.4 2.8 3.2 3.6 0.8 1.2 1.6 2 2.4 2.8 3.2 3.6
Inter-story drift ratio (×10-3) Inter-story drift ratio (×10-3)

(a)
12 Above =1.088, T-0.25, F1
u,max
12 Above =1.209, S-3, F12
u,max

11 ground =0.938, T-6, F1 11 ground =0.874, S-0.25, F12

u,min u,min

10 10 ONE
9 ONE 9 S-0.25
T-0.25 S-1
8 8
T-1 S-2
7 T-2 7
Floor
Floor

S-3
6 T-3 6 S-5
5 T-5 5 S-6
T-6
4 4
3 =1.285, T-0.25
u,max
3 =1.384, S-0.25
u,max
B1 B1
2 =1.019, T-6
u,min
2 =1.001, S-5
u,min

1 1
B1 B1
0.4 0.8 1.2 1.6 2 2.4 0.4 0.8 1.2 1.6 2 2.4
Inter-story drift ratio (×10-3) Inter-story drift ratio(×10-3)

(b)

12 Above =1.087, T-2, F2

u,max
12 Above =1.115, S-5, F1
u,max

11 ground =0.882, T-1, F12 11 ground =0.787, S-0.25, F12

u,min u,min

10 10
ONE ONE
9 9
T-0.25 S-0.25
8 T-1 8 S-1
7 T-2 7 S-2
Floor
Floor

T-3 S-3
6 6
T-5 S-5
5 5
T-6 S-6
4 4
3 3
2 =1.000, T-5
u,max 2 =1.135, S-5
u,max
B1 B1
1 u,min=0.882, T-3 1 u,min=0.864, S-2

B1 B1
0.4 0.6 0.8 1 1.2 1.4 1.6 0.4 0.6 0.8 1 1.2 1.4 1.6
Inter-story drift ratio (×10-3) Inter-story drift ratio (×10-3)

(c)

Fig. 13. Interlayer displacement angle under different ground motion. (a) LP wave; (b) NJ wave; (c) WL wave.

13
H. Long et al. Engineering Structures 242 (2021) 112550

Table 6 the shear force response of the pile foundation. The reason is that
Maximum relative horizontal displacement between the ground and the soil at a when the structure spacing is small, the foundations of adjacent
depth of 4.35 m(mm). structures and the soil between the foundations are likely to form a
D/W S-A O-A S-C O-C D/W S-A O-A S-C O-C sound interaction system. In this way, the soil’s stiffness between the
0.25 3.14 1.4 4.50 2.52 3 2.53 2.25 2.28 2.20
foundations is significantly increased, which leads to a reduction in
1 3.12 1.20 3.75 1.49 5 3.04 2.41 3.54 2.55 the peak total shear force on the pile top of building B. Furthermore,
2 2.96 1.93 3.68 2.19 6 3.06 3.33 3.67 1.74 the peak total shear force is reduced more when adjacent high-rise
structures are on both sides. Similar results were also given in the
research of Rahgozar et al. [12]. They studied the seismic response of
1. When D/W = 0.25, the peak total shear force response on the pile top
1 or 2 high-rise building with pile foundation and found that the
is reduced under all ground motion. Compared with adjacent struc­
adjacent structure will reduce each other’s foundation’s shear force
tures on one side, the peak shear force responses are smaller when
response. On the other hand, when the structure spacing is large, the
there are adjacent structures on both sides. When D/W ≥ 2, in most
foundations of adjacent structures and the soil between the founda­
cases, the peak total shear force response on pile top increases.
tions cannot form a sound interaction system, and the reflection and
Moreover, the influence range of adjacent structures is extensive, and
superposition of seismic waves between adjacent structures will in­
the peak total shear force response on pile top can increase by 9.1%
crease the peak total shear force of the pile top of building B.
(WL wave, S-6) when D/W = 6. It can be considered that when 2 ≤
D/W ≤ 6, adjacent structures tend to have an unfavorable effect on

Table 7
Peak shear force response and βv on each pile top under LP wave.
Pile foundation Vmax,1 /(kN) βv

One T-0.25 T-1 T-2 T-3 T-4 T-5 T-6

Z1 637.8 1.053 1.181 1.097 0.994 1.150 0.943 0.882

Z2 628.7 1.001 1.293 1.385 1.365 1.333 1.303 1.418
Z3 808.5 0.741 0.791 0.863 0.807 0.846 0.852 0.828
Z4 749.2 0.997 0.944 0.943 0.911 0.974 0.927 0.986
Total 1695.4 0.905 0.945 1.077 1.073 1.091 1.126 1.012
S-0.25 S-1 S-2 S-3 S-4 S-5 S-6
Z1 0.835 0.870 0.881 0.883 0.612 0.725 0.947
Z2 1.171 1.550 1.599 1.558 1.613 1.578 1.483
Z3 0.882 0.836 0.882 0.898 0.980 0.873 0.852
Z4 0.843 0.711 0.663 0.656 0.664 0.601 0.881
Total 0.800 0.993 1.014 1.045 1.075 1.116 1.084

Table 8
Peak shear force response and βv on each pile top under NJ wave.
Pile foundation Vmax,1 /(kN) βv

One T-0.25 T-1 T-2 T-3 T-4 T-5 T-6

Z1 597.7 1.143 1.048 1.076 0.909 1.007 1.005 0.988

Z2 542.2 1.041 1.153 0.968 1.094 1.139 1.067 1.051
Z3 751.5 0.941 0.991 1.008 1.011 1.099 1.035 1.111
Z4 737.9 1.110 1.055 1.057 0.897 0.997 1.073 0.895
Total 1572.9 0.932 1.063 1.054 1.051 1.112 1.016 1.027
S-0.25 S-1 S-2 S-3 S-4 S-5 S-6
Z1 1.076 1.043 0.881 0.798 0.586 0.700 0.977
Z2 1.386 1.552 1.544 1.396 1.501 1.595 1.595
Z3 0.974 1.142 1.119 1.145 1.310 1.024 1.118
Z4 0.871 0.802 0.800 0.724 0.697 0.684 0.945
Total 0.713 1.077 1.045 0.961 1.009 1.038 1.027

Table 9
Peak shear force response and βv on each pile top under WL wave.
Pile foundation Vmax,1 /(kN) βv

One T-0.25 T-1 T-2 T-3 T-4 T-5 T-6

Z1 635.9 1.161 0.982 0.988 0.949 1.215 0.927 0.862

Z2 607.5 1.150 1.230 0.984 1.116 1.307 1.138 1.366
Z3 752.4 0.911 0.963 0.987 0.944 1.127 0.994 1.057
Z4 769.0 0.932 0.873 0.930 0.968 0.925 0.888 0.914
Total 1313.1 0.963 1.025 0.992 1.104 1.192 1.115 1.077
S-0.25 S-1 S-2 S-3 S-4 S-5 S-6
Z1 0.966 0.814 0.794 0.678 0.427 0.730 0.977
Z2 1.325 1.362 1.448 1.280 1.333 1.509 1.397
Z3 0.939 1.084 1.078 1.175 1.222 0.968 1.022
Z4 0.875 0.748 0.630 0.711 0.652 0.599 0.875
Total 0.821 1.096 1.055 1.116 1.154 1.193 1.091

14
H. Long et al. Engineering Structures 242 (2021) 112550

2. Adjacent structures significantly influence the peak shear force 4. Generally speaking, adjacent structures on both sides of the middle
response of each pile, especially when there are adjacent structures structure have a more significant impact on the inter-story drift than
on both sides. For example, in the S-5 work condition under NJ wave, when there is an adjacent structure on one side. For the above-
the peak shear forces of the pile tops of Z2 and Z3 increased by 59.5% ground part of the structure, adjacent structures increase or
and 2.4%. In contrast, the peak shear forces of Z1 and Z4 decreased decrease each floor’s inter-story drift response. This effect is related
by 30% and 31.6%. The corresponding peak total shear force on the to the input ground motion and the vibration characteristics of the
pile top only increased by 3.8%. It shows that the adjacent structure’s structure-soil-structure system. Moreover, the most influential loca­
influence leads to the redistribution of the horizontal shear force tions are the top and bottom of the above-ground part, which have a
among the pile foundations. The peak shear force response of the maximum increase of 20.9%. For the basement, the inter-story drift
middle pile increases while the side piles’ peak shear force decreases. is mainly affected by the surrounding soil deformation. If the adja­
Even when the peak total shear force is almost constant, the peak cent structure increases the soil deformation around the middle
shear force on each pile top will change dramatically. In general, in structure, the inter-story drift of the middle structure’s basement will
pile foundation design, the total horizontal shear force is usually also increase; otherwise, the inter-story drift may decrease. As a
evenly distributed to each pile foundation. The redistribution of each result, the maximum increase is 38.4%.
pile’s horizontal shear force by the adjacent structure may lead to 5. The influence of the SSSI effect on the peak total shear force response
insufficient horizontal bearing capacity of some pile foundations. It on the pile top of the middle structure also shows a trend of first
will be highly detrimental to the seismic safety of the structure. decreasing and then increasing with the increase of the structure
spacing. More importantly, adjacent structures’ influence leads to
6. Summary and conclusions the redistribution of horizontal shear forces among the pile founda­
tions. Even when the peak total shear force is almost constant, the
Under actual engineering conditions, the dynamic SSSI effect of peak shear force on each pile top will change dramatically, and the
adjacent structures in earthquakes is highly complex. Based on the soil- maximum increase in the peak shear force is 61.3%. It is highly
structure system model that is too simplified or does not consider ma­ unfavorable to the structure’s seismic safety and should be taken
terial nonlinearity, the numerical analysis of the SSSI effect of adjacent seriously in the seismic design.
structures is carried out without considering the static state’s influence.
The results may deviate from the actual situation. To more accurately In short, the influence of the SSSI effect on the seismic response
study the dynamic SSSI effects between adjacent structures, the finite between adjacent structures cannot be ignored. The SSSI effect of
element method is used to conduct a two-dimensional static-dynamic adjacent structures should be fully considered in the seismic design of
coupled numerical analysis of the seismic response of a single, two or structures. The study of SSSI based on nonlinear conditions is closer to
three identical high-rise structures in the site. The finite element method engineering reality, but it also increases the problem’s complexity. Some
is used to perform a static-dynamic coupling numerical analysis of the conclusions obtained are similar to the previous research results, but
seismic response of single, 2 or 3 identical high-rise structures in the site. there are many differences. Nevertheless, the study based on the two-
In the analysis, soil and structural materials’ nonlinearity is considered, dimensional plane problem is only applicable to the situation where
and a contact method that can produce slippage and instantaneous the structures are aligned horizontally in a straight line in practical
separation between soil and structure is adopted. The purpose is to more engineering. Therefore, in the follow-up, based on the results of the
realistically simulate the influence of the SSSI effect on the structure’s current 2D study, 3D models will be established to study the SSSI effect
seismic response when there are the same adjacent structures on one or between structures further.
both sides. Then the main discussion and conclusions are as follows:
CRediT authorship contribution statement
1. The influence of the frequency spectrum characteristics of the input
ground motion, the structure spacing and the number of adjacent Hui Long: Conceptualization, Methodology, Software, Writing -
structures on the SSSI effect between the structures cannot be original draft. Zicheng Wang: Software, Data curation, Writing - review
ignored. The actual seismic response of the structure is the result of & editing. Chunshun Zhang: Supervision, Resources, Formal analysis,
the combined effect of the above factors. Due to the complexity of Writing - review & editing. Haiyang Zhuang: Methodology. Wenzhao
nonlinear problems, not all SSSI effects can be found to have Chen: Investigation. Cheng Peng: Investigation.
consistent laws on the seismic response of structures, but some laws
and trends can still be found.
2. As the structure spacing increases, the impact of the SSSI effect on the Declaration of Competing Interest
peak acceleration response of the middle structure (Building B) de­
creases first. It then increases, and the impact on different floors is The authors declare that they have no known competing financial
different. The SSSI effect has the most unfavorable impact on the interests or personal relationships that could have appeared to influence
peak acceleration response at the 1/4–1/2 height of the middle the work reported in this paper.
structure, with a maximum increase of 14.3%. Also, having adjacent
structures on both sides of the same location is more disadvanta­ Acknowledgement
geous than having adjacent structures on one side. When D/W ≥ 5,
the influence of the SSSI effect on the middle structure’s response This study was supported by the National Natural Science Founda­
under peak acceleration can be ignored. tion of China NSFC (Grant No.51708273), Natural Science Foundation
3. The influence of the SSSI effect on the standard acceleration response of Hunan Province (Grant No.2017JJ3263) and State Scholarship Fund
spectrum of the middle structure is mainly reflected in the change of of China Scholarship Council (File No.201908430064), which are
the local spectrum value. When there are adjacent structures on both gratefully acknowledged.
sides, its influence on the spectrum value is more significant than
adjacent structures on one side. Besides, the normalized acceleration References
response spectrum value of the structure near the second natural
period of the structure increases when 1 ≤ D/W ≤ 4, which may [1] Karabalis DL, Huang CFD. 3-D foundation-soil-foundation interaction. WIT Trans
Modelling Simulation 1970;8.
increase the effect of the second-order mode on the seismic response [2] Huang CFD. Dynamic soil-foundation and foundation-soil-foundation interaction in
of similar high-rise structures 3-D. Columbia, South Carolina: University of South Carolina; 1993.

15
H. Long et al. Engineering Structures 242 (2021) 112550

[3] Karabalis DL, Mohammadi M. 3-D dynamic foundation-soil-foundation interaction [20] Gan J, Li P, Liu Q. Study on dynamic structure-soil-structure interaction of three
on layered soil. Soil Dyn Earthquake Eng 1998;17(3):139–52. https://doi.org/ adjacent tall buildings subjected to seismic loading. Sustainability 2019;12(1):
10.1016/S0267-7261(97)00047-X. 1–19. https://doi.org/10.3390/su12010336.
[4] Liang J, Han B, Todorovska MI, Trifunacb MD. 2D dynamic structure-soil-structure [21] Bolisetti C, Whittaker AS. Numerical investigations of structure-soil-structure
interaction for twin buildings in layered half-space I: Incident SH-waves. Soil Dyn interaction in buildings. Eng Struct 2020;215:110709. https://doi.org/10.1016/j.
Earthquake Eng 2017;102:172–94. https://doi.org/10.1016/j. engstruct.2020.110709.
soildyn.2017.08.017. [22] Mason HB, Trombetta NW, Chen Z, Brayd JD, Hutchinsonb TC, Kuttere BL. Seismic
[5] Liang J, Han B, Todorovska MI, Trifunacb MD. 2D dynamic structure-soil-structure soil-foundation-structure interaction observed in geotechnical centrifuge
interaction for twin buildings in layered half-space II: Incident SV-waves. Soil Dyn experiments. Soil Dyn Earthquake Eng 2013;48:162–74. https://doi.org/10.1016/
Earthquake Eng 2018;113:356–90. https://doi.org/10.1016/j. j.soildyn.2013.01.014.
soildyn.2018.05.023. [23] Trombetta NW, Mason HB, Hutchinson TC, Zupan JD, Bray JD, Kutter BL.
[6] Lin HT, Roesset JM, Tassoulas JL. Dynamic interaction between adjacent Nonlinear soil-foundation-structure and structure-soil-structure interaction:
foundations. Earthquake Eng Struct Dyn 1987;15(3):323–43. https://doi.org/ centrifuge test observations. J Geotech Geoenviron Eng 2014;140(5):04013057.
10.1002/eqe.4290150304. [24] Trombetta NW, Benjamin Mason H, Hutchinson TC, Zupan JD, Bray JD, Kutter BL.
[7] Wang S, Schmid G. Dynamic structure-soil-structure interaction by FEM and BEM. Nonlinear soil-foundation-structure and structure-soil-structure interaction:
Comput Mech 1992;9(5):347–57. https://doi.org/10.1007/BF00370014. engineering demands. J Struct Eng 2015;141(7):04014177.
[8] Lehmann L, Antes H. Dynamic structure-soil-structure interaction applying the [25] Knappett JA, Madden P, Caucis K. Seismic structure-soil-structure interaction
symmetric Galerkin boundary method (SGBEM). Mech Res Com 2001;28(3): between pairs of adjacent building structures. Géotechnique 2015;65(5):429–41.
297–304. https://doi.org/10.1016/S0093-6413(01)00177-X. https://doi.org/10.1680/geot.SIP.14.P.059.
[9] Álamo Guillermo M, Padrón Luis A, Aznárez Juan J, Maeso Orlando. Structure-soil- [26] Hayden CP, Zupan JD, Bray JD, Allmond JD, Kutter BL. Centrifuge tests of adjacent
structure interaction effects on the dynamic response of piled structures under mat-supported buildings affected by liquefaction. J Geotech Geoenviron Eng 2015;
obliquely incident seismic shear waves. Soil Dyn Earthquake Eng 2015;78:142–53. 141(3):04014118.
https://doi.org/10.1016/j.soildyn.2015.07.013. [27] Ngo VL, Ewusi-Wilson R, Ike E. Investigation of the structure-soil-structure
[10] Dou LJ, Yang BP. Dynamic interaction of tall building with neighboring multi- interaction between two structures in centrifuge test. In: Geotechnics for
storied building. Earthquake Eng Eng Vibration 2000;20(3):15–21. Sustainable Infrastructure Development. Singapore: Springer; 2020. p. 1151–7.
[11] Zhang JL, Chen GX. Numerical simulation of the earthquake response for adjacent https://doi.org/10.1007/978-981-15-2184-3_150.
double high-rise building with pile-box foundation on deep soft sites. World [28] Aldaikh H, Alexander NA, Ibraim E, Knappett J. Shake table testing of the dynamic
Earthquake Eng 2003;19(4):99–105. interaction between two and three adjacent buildings (SSSI). Soil Dyn Earthquake
[12] Rahgozar MA. Accounting for soil nonlinearity in three-dimensional seismic Eng 2016;89:219–32. https://doi.org/10.1016/j.soildyn.2016.08.012.
structure-soilstructure-interaction analyses of adjacent tall buildings structures. Int [29] Chandrakala B. Dynamic analysis of soil pile structure interaction in soft clay using
J Civil Eng 2015;13(3):213–25. shake table test. J Eng Sci 2019; 6(2).
[13] Ghandil M, Behnamfar F, Vafaeian M. Dynamic responses of structure-soil- [30] Aldaikh H, Alexander NA, Ibraim E, Oddbjornssona O. Two dimensional numerical
structure systems with an extension of the equivalent linear soil modeling. Soil Dyn and experimental models for the study of structure-soil-structure interaction
Earthquake Eng 2016;80:149–62. https://doi.org/10.1016/j.soildyn.2015.10.014. involving three buildings. Comput Struct 2015;150:79–91. https://doi.org/
[14] Yahyai M, Mirtaheri M, Mahoutian M, Daryan AS, Assareh MA. Soil structure 10.1016/j.compstruc.2015.01.003.
interaction between two adjacent buildings under earthquake load. Am J Eng Appl [31] Vicencio F, Alexander NA. Dynamic interaction between adjacent buildings
Sci 2008;1(2):121–5. through nonlinear soil during earthquakes. Soil Dyn Earthquake Eng 2018;108:
[15] Sushma P, Kumar RP. Dynamic soil structure interaction analysis of pile supported 130–41. https://doi.org/10.1016/j.soildyn.2017.11.031.
high rise structures. International Conferences on Recent Advances in Geotechnical [32] Vicencio F, Alexander NA. Higher mode seismic structure-soil-structure interaction
Earthquake Engineering and Soil Dynamics. California, USA, 2010; 12. between adjacent building during earthquakes. Eng Struct 2018;174:322–37.
[16] Yue D, Ghiocel D M, Fuyama H, Ogata T, Stark G. Structure-soil-structure https://doi.org/10.1016/j.engstruct.2018.07.049.
interaction effects for two heavy npp buildings with large-size embedded [33] Vicencio F, Alexander NA. Dynamic structure-soil-structure interaction in
foundations. SMiRT22 Proceedings, California, USA, 2013; Division V. unsymmetrical plan buildings due to seismic excitation. Soil Dyn Earthquake Eng
[17] Wang HF, Lou ML, Chen X, Zhai YM. Structure-soil-structure interaction between 2019;127:105817. https://doi.org/10.1016/j.soildyn.2019.105817.
underground structure and ground structure. Soil Dyn Earthquake Eng 2013;4: [34] Lee J, Fenves GL. Plastic-damage model for cyclic loading of concrete structures.
31–8. https://doi.org/10.1016/j.soildyn.2013.07.015. J Eng Mech 1998;124(8):892–900.
[18] Wang HF, Lou ML, Zhang RL. Influence of presence of adjacent surface structure on [35] Zhuang H, Chen G, Zhu D. Dynamic visco-plastic memorial nested yield surface
seismic response of underground structure. Soil Dyn Earthquake Eng 2017;100: model of soils and its verification. Chin J Geotech Eng 2006;28(10):1267–72.
131–43. https://doi.org/10.1016/j.soildyn.2017.05.031. [36] Zhuang HY, Chen GX. Improvement of dynamic viscoplastic memorial nested yield
[19] Li P, Liu S, Lu Z. Studies on pounding response considering structure-soil-structure surface model of soil. Rock Soil Mech 2009;30(1):118–22.
interaction under seismic loads. Sustainability 2017;9(12):2219. https://doi.org/
10.3390/su9122219.

16