Hindawi

Advances in Materials Science and Engineering

Volume 2019, Article ID 4592847, 13 pages
https://doi.org/10.1155/2019/4592847

Research Article
Seismic Fragility Analysis of Buildings Based on
Double-Parameter Damage Models considering
Soil-Structure Interaction

Panpan Zhai,1,2 Peng Zhao,2,3 Yang Lu,2,3 Chenying Ye,2,3 and Feng Xiong 2,3

1
Institute of New Energy and Low-carbon Technology, Sichuan University, Chengdu 610065, China
2
MOE Key Laboratory of Deep Underground Science and Engineering, School of Architecture and Environment,
Sichuan University, Chengdu 610065, China
3
College of Architecture & Environment, Sichuan University, Chengdu 610065, China

Correspondence should be addressed to Feng Xiong; fxiong@scu.edu.cn

Received 7 June 2019; Accepted 20 September 2019; Published 24 October 2019

Academic Editor: Paweł Kłosowski

Copyright © 2019 Panpan Zhai et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Most conventional seismic fragility analyses of RC buildings usually ignore or greatly simplify the soil-structure interaction (SSI),
and the maximum interstory drift ratio (MIDR) is often adopted to establish seismic fragility curves. In this work, an eight-story
RC building was designed to study the influence of the SSI on the seismic fragility of RC buildings. Three double-parameter
damage models (DPDMs) were considered for the fragility assessment: the Park–Ang model, the Niu model, and the Lu–Wang
model. Results show that considering SSI induces a higher fragility than that of the fixed model and that employing the DPDMs for
the fragility analysis provides more reasonable results than those evaluated using the MIDR damage index.

1. Introduction It is of great importance to include soil in the numerical

model in the fragility analysis. When an earthquake appears,
Seismic fragility analysis of structures is the basis of seismic the propagation of seismic waves in the soil results in soil
loss estimation and city risk management. This type of movement and is proven that soil-structure interaction (SSI)
analysis is helpful in understanding the response of struc- affects the dynamic responses of the structure. For example,
tures under an earthquake so as to define the structural Myat [3] studied the seismic performance of low-rise shear
damage corresponding to a specific earthquake intensity wall buildings with a strip foundation, and the results
level. Therefore, decision-makers can adopt the fragility showed that both the axial forces and shear forces of the
analysis as a significant reference to reduce the consequences columns increased compared with those with the fixed base.
of earthquakes [1]. Dutta et al. [4] found that the SSI could increase the base
Developing fragility curves is useful for seismic fragility shear of a low-rise RC frame building but decreased the
assessment of structures. In principle, fragility curves can seismic response of high-rise buildings. Pinto and Ciampoli
be established using the following methods [2]: (1) pro- [5] investigated typical piers and assessed the effect of SSI on
fessional expert judgement; (2) quasistatic and design code dynamic response, which indicated that in most cases, SSI
consistent analysis; (3) statistical analysis of damage data led to a higher displacement demands than fixed-base
associated with past earthquakes; and (4) numerical sim- structures.
ulation based on the structural dynamic analysis. Due to Unfortunately, most existing numerical studies of fra-
the scarcity and low reliability of postearthquake data, the gility do not consider the SSI effects and simply fix the
numerical method is the most commonly adopted method bottom of the structure. Only a few works currently include
in current research. the SSI effects in the structural fragility analysis, which shows
2 Advances in Materials Science and Engineering

the changes caused by the SSI effect. For example, the study (DI). Through comparing the damage index with the limit
of Rajeev and Tesfamariam [6] highlighted the effects of the values defined for specific damage states, the damage of
SSI on three-, five-, and nine-story, three-bay moment structures can be assessed. Three typical DPDMs are selected
reinforced concrete frames resting on dense silty sand, and in this investigation, including the Park–Ang model, which
the fragility curves were developed for both fixed-base and is widely accepted in the earthquake engineering field, the
SSI models. The results showed that the SSI effect induced Niu model, of which the relevant parameters are easy to
higher fragility than that of the fixed-base model. Similarly, determine, and the Lu–Wang model. The expressions of the
Pitilakis et al. [7] studied three RC buildings resting on very DPDMs adopted can be seen in Table 1.
soft soil, and the comparative fragility curves were obtained In Table 1, δm is the maximum displacement under an
with the fixed-base and SSI models. A substantial increase in earthquake, δu is the ultimate displacement under a
the fragility was observed for the SSI models with respect to monotonic loading, δy is the yield displacement, Fy is the
the fixed-base structures. The research of Saez et al. [8] also calculated yield strength, Eu is the limit hysteretic energy, Eh
studied the effects of the SSI on the fragility assessment for is the cumulative hysteretic energy, and βo is a nonnegative
an old RC building resting on a mixture of sandy and clayey parameter. In the Niu damage model, μ � 0.1387 and
soil, and a general reduction in the fragility was presented. c � 0.0814. When evaluating the damage that appears in a
As is well known, the SSI has a very complex influence on the specified story level or an entire structure, a weighted
structural responses; therefore, it does not always lead to combining index is required:
beneficial effects. As a result, it is appropriate to study the SSI n
effects on the structural fragility, which is helpful to generate DI � 􏽘 λi Di . (1)
more accurate fragility curves. i�1
Furthermore, all the seismic fragility analyses above
Some researchers [15–17] proposed different methods to
adopt a simplified soil model represented by a series of
calculate the weighting coefficients. In this study, the method
springs. The characteristics of the springs will influence the
proposed by Du and Ou is first adopted to generate the story
structural fragility results. With the development of com-
level damage index, and the method proposed by Ou and He
puting technologies, it is possible to employ a full soil-
is then adopted to define the damage index of the total
structure model in the fragility analysis to obtain satisfactory
structure. The weighting coefficient method adopted in this
results. On the other hand, the current fragility analysis uses
study can be seen in Table 2.
the maximum interstory drift ratio (MIDR) to assess the
To obtain the structural fragility curves, the relationship
damage state of structures. The MIDR is the most commonly
between the structural response and damage states must be
used metric, as it is easily obtained in the performance
clearly defined. Based on the existing methods, five damage
analysis of the structure under an earthquake. However, this
states and four limit states (LS) of slight damage, moderate
response parameter cannot fully reflect the cumulative
damage, severe damage, and complete damage are defined,
damage caused by the low cycle fatigue under the re-
and the corresponding indices of four limit states are shown
ciprocating load of an earthquake [9]. Therefore, using
in Table 3.
MIDR may underestimate the degree of damage, especially
when under a long-lasting earthquake, which will lead to
severe seismic loss. Because the double-parameter damage 3. Numerical Modeling
models (DPDMs), generally using the combination of de- To study the SSI effects on structural fragility using the
formation and energy as the damage index, can reflect both DPDMs, a typical eight-story, three-bay reinforced concrete
the maximum deformation and cumulative damage effects, frame building is designed according to the Chinese code for
it is considered to be more comprehensive when describing Seismic Design of Buildings [19]. Two foundation systems
the structural damage under earthquakes, which may give an are considered in this study: fixed-base and piled raft sup-
accurate prediction of the seismic damage of buildings. ported on soft soil considering the SSI effects. The di-
In this study, to investigate the SSI effects on fragility mensions of the reference structure are shown in Figures 1
assessment and supply a method for the fragility analysis of and 2.
SSI structures, an eight-story RC frame structure was
designed and fully modeled as a test bed to carry out a series
of numerical fragility analyses. At the same time, three 3.1. Superstructure Model. The reference structure adopted
DPDMs of Park–Ang, Niu, and Lu–Wang model are in this study is a high-rise building with 8 stories, with a story
adopted to evaluate the seismic fragility of structures height of 3.0 m. The building has a floor plan of 6.0 m by
compared to using the MIDR. 14.4 m with three bays in the horizontal direction (shown in
Figures 1 and 2). The section size of the beam is 300 mm by
2. Damage Models 600 mm and of the column is 800 mm by 800 mm, and the
thickness of the floor is 120 mm. The structure is designed
The seismic damage model is an important tool to describe using the PKPM software according to the “Code for Seismic
the degree of damage to structures under an earthquake, Design of Buildings” (GB50011-2010) [18] and the “Code for
making it a key issue in fragility analyses. In general, the Concrete Structure Design” (GB50010-2010) [20]. The
damage model is a quantitative mathematical model de- concrete grade of the beam, column, and floor is C30, and
scribing the severity of damage, measured by a damage index the equivalent elastic modulus is 32000 MPa. The C30
Advances in Materials Science and Engineering 3

Table 1: The DPDMs adopted in this study. to about 200 m/s is used, corresponding to ground type II of
Damage models Expression
the Chinese code for seismic design of buildings. A very soft
soil is specifically selected to amplify the effects of SSI. The
Park–Ang [10, 11] DI � (δm /δu ) + βo (Eh /Fy δu )
Niu [12] DI � (δm /δu ) + μ(Eh /Eu )c
damping ratio for energy dissipation during seismic loading
is equal to 5%. The specific soil parameters can be seen in
DI � (1 − βo )(δm − δy )/(δu −
Lu–Wang [13, 14]
δy ) + βo (Eh /Fy (δu − δy ))
Table 4.
The soil is modeled using the Solid185 element. To
simulate the contact performance in the interface between
the soil and the pile, the contact elements are employed in
Table 2: Weighting coefficient method adopted in this study.
this study, where the target surface used the Target170
Weighting coefficient Expression element and the contact surface used the Contact173
Du method [15] λi � Di /􏽐 Di element.
Ou method [16] λi � (N + 1 − i)/􏽐N
j− 1 (N + 1 − j)
Note: Di � weighting coefficient for component i or story i; λi � damage 3.3. Input Ground Motions. A representative series of
index of component i or story i; N � the number of stories.
earthquake ground motions are selected for the dynamic
analysis, including 5 real ground motion recordings (Ta-
concrete and HRB400 rebars (with a reinforcement ratio of
ble 5) selected from the Ground Motion Database of the
0.1%) were used, and the reinforced concrete was treated as a
Pacific Earthquake Engineering Research Center (PEER:
bulk material that has been shown to give reasonably ac-
http://peer.berkeley.edu/) according to the target spectrum
curate results [21–23]. The multilinear isotropic strength-
of the Chinese codes. The earthquake magnitudes are
ening (MISO) constitutive model in ANSYS was used to
5 < Mw < 8, and the epicentral distances are 0 < R < 25 km.
simulate the material behaviour. The MISO model was
Because the foundation soil used in the study, on average,
calibrated against Pan’s [24] uniaxial compressive stress-
belongs to the soil category according to GB50011, with a
strain curves.
Vs,30 lower than 500 m/s, the selected ground motions were
To analyse the structural responses, the ANSYS software
all recoded on soil sites with Vs,30 ranging from 288 to 413 m/
is employed. The structure and the soil are meshed as in
s. Figure 4 shows the spectra of these five ground motions.
Figure 3. In the model, the floor of the structure is simulated
The soil-structure interaction system is first analysed under
using the Shell181 element, and the beams, columns, and
the self-weight loading to achieve a geostatic condition, and
piles are all simulated using the Beam188 element.
then the selected ground acceleration time series are applied
In addition, the weight of superstructure and soil is
to the bottom of the soil layer.
added to the calculation model by means of concentrated
nodal force, in which the live load of the floor and the weight
of the wall in the frame design is considered, to balance the 4. Dynamic Analysis
computational efficiency. The live load of the floor and the
weight of the walls are all loaded on the floor, and the mean To illustrate the effect of different intensities of seismic
value of 9.2 kN/m2. Due to the large scale of the SSI system, ground motions on the structural damage distribution
the finite element model consists of a total 137,216 elements. using the DPDMs of Park–Ang, Niu, and Lu–Wang
compared with that using MIDR for the analysed building
with or without SSI effects, a series of preliminary com-
3.2. Soil Model. When considering the SSI in the fragility parative dynamic analyses are conducted for the SSI and
analysis, it is crucial to determine the range of the soil to fixed-based models based on Park–Ang, Niu, and Lu–
ensure the reliability of the analysis. Because the proper size Wang damage models with respect to the MIDR under
of the soil has a considerable influence on the results of the different intensities of seismic ground motions. The Tai-
numerical analysis, Lu and Jiang [25] studied the soil do- wanSMART1(5) earthquake record is used in this case as
main range using the displacement response analysis the input motion, scaled to reach PGA values of 0.1 g, 0.2 g
method. The results showed that the soil domain should be and 0.4 g.
taken as at least 8L (L is the width of the structure) with more Figures 5–7 show the heightwise distribution of MIDR
than 10L yielding very satisfactory results. and the story DI based on the DPDMs of Park–Ang, Niu,
Therefore, the size of the soil is defined to be 180 m by and Lu–Wang at different seismic intensity levels of 0.1 g,
180 m (more than 10L), and the depth of the soil is 20.0 m. 0.2 g and 0.4 g, respectively. It can be seen that the SSI effects
The Drucker–Prager constitutive model can better reflect the lead to higher MIDR compared with the fixed-base model.
soil characteristic when simulating the soil elastic-plastic The MIDR of the building considering SSI effects is con-
constitutive relation, and the model requires several pa- centrated at the first story, but the MIDR of the fixed-base
rameters consisting of the elastic modulus, the Poisson ratio, building appears at the third story. For the damage distri-
the density, the soil cohesion, and the internal friction angle. bution of the building using the DPDMs, it is shown that the
In addition, the Drucker–Prager model has achieved good maximum damage appears at the second story for the fixed-
results in soil constitutive simulation through many prac- base building and that at the first floor for the building
tical engineering verifications; therefore, it is adopted in this considering SSI effects. The results show that SSI effects will
study [26, 27]. An average shear wave velocity of Vs,30 equal influence the damage distribution of the structure using both
4 Advances in Materials Science and Engineering

Table 3: Damage indices of the limit state.

Limit state Slight damage Moderate damage Extensive damage Complete damage
MIDR [18] 1/450 1/300 1/150 1/50
Park–Ang [10] 0.1 0.25 0.4 1.0
Niu [12] 0.2 0.4 0.65 0.9
Lu–Wang [13] 0.2 0.4 0.6 0.9

6.0m 2.4m 6.0 m

3.0∗8 = 24.0m
6.0 m 2.4m 6.0m
6.0m

Plan Elevation
Figure 1: Dimensions of the fixed-base model.

Free field

Vso, 30

Input motion
Elastic bedrock
180.0 m
Figure 2: Dimensions of the SSI model.

(a) (b)

Figure 3: Three-dimensional finite element meshes of the structure adopted in this study. (a) Fixed-base model and (b) SSI model.
Advances in Materials Science and Engineering 5

Table 4: Parameters of soil.

Internal
Thickness Density Shear wave velocity Passion’s Damping Shear modulus Elastic modulus Cohesion
friction
(m) (kg/m3) (m/s) ratio ratio (MPa) (MPa) (Pa)
angle (°)
5 2000 190 0.4 0.05 20 48 25000 15
5 2100 200 0.4 0.05 20.4 72.58 40000 20
10 2200 210 0.4 0.05 43.12 103.49 50000 25

Table 5: List of earthquake records used for analysis.

Earthquake recording Station Year Mw R (km) PGA (g) Vs,30 (m/s)
Parkfield Cholame-ShandonArray #5 1966 6.19 9.5 0.391 289
Ma.Nicaragua-01 Managua, ESSO 1972 6.24 4.06 0.26 288
Friuli, Italy-02 Forgaria Cornino 1976 5.91 14.75 0.225 413
San Fernando LA-Hollywood Stor FF 1971 6.61 22.27 0.225 316
TaiwanSMART1(5) SMART1 007 1981 5.9 24.93 0.085 314

1.4

1.2

0.8
Sa (g)

0.6

0.4

0.2

0
0 1 2 3 4
Period (s)
Adopted earthquake records
Mean earthquake records
Target spectrum
Figure 4: Response spectra of the selected input seismic ground motions.

8 8

7 7

6 6

5 5
Stories
Stories

4 4

3 3

2 2

1 1
0 0.2 0.4 0.6 0.8 1 0 0.002 0.004 0.006 0.008 0.01
DI MIDR
SSI-Niu Fixed-Niu SSI model
SSI-Park–Ang Fixed-Park–Ang Fixed-base model
SSI-Lu–Wang Fixed-Lu–Wang

(a) (b)

Figure 5: Distribution of damage at PGA of 0.1 g with DPDMs (a) and the MIDR (b).
6 Advances in Materials Science and Engineering

8 8

7 7

6 6

Stories 5 5

Stories
4 4

3 3

2 2

1 1
0 0.2 0.4 0.6 0.8 1 1.2 0 0.002 0.004 0.006 0.008 0.01
DI IDR
SSI-Niu Fixed-Niu SSI model
SSI-Park–Ang Fixed-Park–Ang Fixed-base model
SSI-Lu–Wang Fixed-Lu–Wang
(a) (b)

Figure 6: Distribution of damage at PGA of 0.2 g with DPDMs (a) and the MIDR (b).

8 8

7 7

6 6

5 5
Stories
Stories

4 4

3 3

2 2

1 1
0 0.4 0.8 1.2 1.6 2 0 0.002 0.004 0.006 0.008 0.01 0.012
DI IMR
SSI-Niu Fixed-Niu SSI model
SSI-Park–Ang Fixed-Park–Ang Fixed-base model
SSI-Lu–Wang Fixed-Lu–Wang

(a) (b)

Figure 7: Distribution of damage at PGA of 0.4 g with DPDMs (a) and the MIDR (b).

the DPDMs and the IMDR. In addition, the maximum the DPDMs and the IMDR. For the building considering the
damage of the fixed-base model is concentrated at the SSI effects, it can be seen that when the PGA is 0.4 g, the
second story using the DPDMs, and the MIDR appears at the structure assessed using DPDMs is in complete damage
third story, but the maximum damage is concentrated at the state, while that evaluated using MIDR is in the severe
first floor using the DPDMs and the IMDR for the building damage state, showing a large difference of results using
considering SSI effects. It is shown that the influence of the different DI. In addition, with the increase in PGA, the
damage distribution of the SSI model using the DPDMs difference in damage assessment using DPDMs is more
compared with the IMDR is different from that of the fixed- noticeable than the MIDR.
base model. The corresponding influence of SSI effects and
the difference using the DPDMs and the MIDR is similar 5. Seismic Fragility Analysis
under different intensity levels of seismic ground motions.
Figure 8 gives the global DI of the structure under 5.1. Incremental Dynamic Analysis. The Incremental dy-
different ground motions of the three DPDMs and the namic analysis (IDA) is an effective analysis method that
MIDR. It depicts that SSI effects induce higher DI for both involves performing a series of nonlinear dynamic analyses
Advances in Materials Science and Engineering 7

1.4 0.012

1.2
0.01

1
0.008
0.8

MIDR
0.006
DI

0.6
0.004
0.4

0.2 0.002

0 0
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
PGA (g) PGA (g)
SSI-Niu Fixed-Niu SSI model
SSI-Park–Ang Fixed-Park–Ang Fixed-base model
SSI-Lu–Wang Fixed-Lu–Wang
(a) (b)

Figure 8: Distribution of the global damage index under different intensities of seismic ground motions with DPDMs (a) and the MIDR (b).

under a suite of scaled ground motion records whose in- structural model. Figure 9 illustrates the PGA-MIDR re-
tensities should ideally cover the entire range from elasticity lationships and PGA with double-parameter damage index
to the whole dynamic instability [28]. An advanced tracing for the fixed-base structural model and the structural model
algorithm of Hunt and Fill is used to scale the earthquake considering SSI effects. It is worth nothing that the structural
records. The PGA is adopted as the intensity measure (I) for model considering SSI effects tends to yield higher damage
the IDA analysis, and each earthquake record is scaled for 10 index compared with the fixed-base model, and the DPDMs
different levels (i.e., 0.05 g, 0.15 g, 0.225 g, 0.3 g, 0.4 g, 0.5 g, behave different distribution compared with the MIDR.
0.625 g, 0.75 g, 0.8 g, and 0.85 g). Based on the dynamic Various uncertainties are taken into account through
analyses above, for both the fixed-base models and SSI the log-standard deviation parameter β, which describes
models, the uncertainty of the input seismic motions, and 50 the total dispersion related to each fragility curve. Three
typical structural samples are obtained for fixed-base model primary sources of uncertainty contribute to the total
and the SSI model to conduct the fragility analysis, variability for any given damage state [30], including the
respectively. variability associated with structural modeling, the capacity
of each structural type, and the seismic demand. The log-
standard deviation value in the definition of limit states is
5.2. Fragility Analysis. Seismic fragility refers to the failure assumed to be equal to 0.2 [31], while the corresponding
probability under different seismic intensities, which also value in the capacity is assumed to be 0.2, 0.3, and 0.4 for
describes the seismic performance of the engineering slight/moderate and severe/complete damage states, which
structures in the sense of a probability. Seismic fragility considers that the uncertainty of the structural capacity
expresses the relationships between the damage degree and increases with the deepening of structural damage [32]. The
the seismic intensity macroscopically using fragility curves. third source of uncertainty associated with the seismic
The response results of the IDA are used to derive fragility demand is taken into consideration by calculating the
curves expressed as double-parameter lognormal distribu- dispersion of the logarithms of simulated results with re-
tion functions. The cumulative probability of exceeding an spect to the regression fit. Under the assumption that these
LS conditioned on a measured of the seismic intensity IM is three log-standard deviation components are statistically
as follows [29]: independent, and the total log-standard deviation is esti-
ln(IM) − ln(IM) mated as the root of the sum of the squares of the com-
Pf (LS | IM) � Φ􏼨 􏼩, (2) ponent dispersions. The computed log-standard deviation
β
β values of the fragility curves in terms of PGA-MIDR vary
where Φ is the standard normal cumulative distribution from 0.79 to 0.86 and 0.85 to 0.92 for the model considering
function, IM is the intensity measure of the adopted SSI effects and the fixed-base model, respectively. However
earthquake expressed in terms of PGA (in units of g ), IM the β values for the DPDMs vary from 0.76 to 0.99 and from
and β are the median values (in units of g ) and the log- 0.73 to 1.07 for the model considering SSI effects and the
standard deviations of the building fragilities, and LS is the fixed-base model, respectively.
limit state. The median PGA values corresponding to the Table 6 presents the lognormal distributed fragility pa-
prescribed performance levels are determined based on a rameters of median and log-standard deviation for the fixed-
regression analysis of the nonlinear IDA results for each base building, and Figure 10 gives the comparative fragility
8 Advances in Materials Science and Engineering

0.1 10

LS4
SSI: y = 0.0162x0.6608 SSI: y = 1.9212x0.6615
LS4
0.01 LS3 1
LS3
LS2
θmax

LS2

DI
LS1
LS1
0.001 0.1

Fixed: y = 0.6598x0.8251
Fixed: y = 0.0065x0.8063
0.0001 0.01
0.01 0.1 1 0.01 0.1 1
PGA (g) PGA (g)
(a) (b)
10 10

SSI: y = 1.5443x0.6501 SSI: y = 1.644x0.765

1 LS4 1 LS4
LS3 LS3
LS2 LS2
DI

DI

LS1 LS1

0.1 0.1

Fixed: y = 0.6599x0.6716
Fixed: y = 0.6071x1.1932

0.01 0.01
0.01 0.1 1 0.01 0.1 1
PGA (g) PGA (g)
(c) (d)

Figure 9: Comparative relationships for the SSI models based on (a) MIDR model and three double-parameter damage models: (b)
Park–Ang model, (c) Niu model, and (d) Lu–Wang model.

Table 6: Parameters of the fragility functions in terms of PGA for with the DPDMs of Park–Ang, Niu, and Lu–Wang. It may
the analysed structure without considering SSI effect. be due to that the failure of the structure is mainly de-
termined by the accumulated damage in this state, but the
Median PGA (g) Dispersion β MIDR only takes the deformation into account. Thus, there
Damage model
LS1 LS2 LS3 LS4 LS1 LS2 LS3 LS4 is a big difference to evaluate the structural fragility using the
IMDR 0.29 0.46 1.03 3.69 0.85 0.88 0.88 0.92 DPDMs and the MIDR damage index, and therefore, it is
Park–Ang 0.11 0.31 0.55 1.65 0.87 0.90 0.90 0.94 unreasonable to only adopt the IMDR as the damage index
Niu 0.17 0.48 0.98 1.58 0.73 0.76 0.76 0.81 for fragility analysis.
Lu–Wang 0.4 0.71 1.04 1.36 1.01 1.04 1.04 1.07 Table 7 summarizes the lognormal distributed fragility
parameters (median and log-standard deviation) in terms
of PGA-MIDR for the buildings with or without SSI effect.
curves for different DPDMs of Park–Ang, Niu, and Lu– Figure 11 depicts the comparative fragility curves for the
Wang compared with that of the MIDR damage index. It is SSI and the fixed-base models using the MIDR damage
shown that the fragility distribution of the fixed-base index. From Figure 11, it can be seen that the SSI effects
building is changed when DPDMs are adopted. From the decrease the median values compared with those of the
slight damage state to the severe damage state, a significant fixed-base model, which lead to that the SSI model presents
overall increase of the fixed-base building’s fragility of higher fragility compared with that of the fixed-base
Park–Ang damage model with respect to the MIDR damage models. This observation is noticeable for four limit
index. Especially in the complete damage state, the median damage limit states. The overall fragility of SSI models
values of Park–Ang, Niu, and Lu–Wang are far smaller than increased significantly even by 50%. It is indicated that the
that of the MIDR. It shows the fragility of the fixed-base seismic performance of the structure is overestimated
building is greatly reduced by adopting the MIDR compared without considering SSI effects. But for the slight and the
Advances in Materials Science and Engineering 9

1 1

0.8 0.8

0.6 0.6
Pf Pf
0.4 0.4

0.2 0.2

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
PGA (g) PGA (g)
MIDR Niu MIDR Niu
Park–Ang Lu–Wang Park–Ang Lu–Wang
(a) (b)
1 1

0.8 0.8

0.6 0.6
Pf Pf
0.4 0.4

0.2 0.2

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
PGA (g) PGA (g)
MIDR Niu MIDR Niu
Park–Ang Lu–Wang Park–Ang Lu–Wang
(c) (d)

Figure 10: Fragility curves of the fixed-base building based on DPDMs compared with the MIDR damage index: (a) slight damage, (b)
moderate damage, (c) severe damage, and (d) complete damage.

Table 7: Parameters of the fragility functions in terms of PGA-MIDR for the analysed structure with and without SSI effect.
Median PGA (g) Dispersion β
RC buildings
LS1 LS2 LS3 LS4 LS1 LS2 LS3 LS4
Fixed base 0.29 0.46 1.03 3.69 0.85 0.88 0.88 0.92
SSI model 0.05 0.09 0.26 1.38 0.79 0.82 0.82 0.86

moderate damage states, the effects decrease the gap be- the fragility curves of the SSI models induce a higher
tween the fragility of SSI model and that of the fixed-base fragility. The higher fragility could be because the SSI effect
model with the increase of PGA. Thus, SSI effects may play introduces additional translation and rotation effects and
a crucial role in the expected structural damage, even if it is increases the displacement demands on the structure or
expressed in terms of the MIDR, and SSI effects should not resonance effect appears.
be neglected for assessment purposes. The results here are The obtained results show that the seismic behaviour
similar to those in the study of Karapetrou et al. [33], which of the structure will be affected when the SSI is consid-
shows that considering the SSI effects leads to a significant ered, especially when using MIDR as the damage index,
increase in the fragility compared with that of the fixed which will lead to an unsafe structure. Therefore, the SSI
base. Additionally, in the research of Rajeeva et al. [6], the effects should not be neglected in the structural fragility
reference structures are of three different heights, and all analysis.
10 Advances in Materials Science and Engineering

1 1

0.8 0.8

0.6 0.6
Pf Pf
0.4 0.4

0.2 0.2

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
PGA (g) PGA (g)
Fixed-base Fixed-base
SSI model SSI model
(a) (b)
1 1

0.8 0.8

0.6 0.6
Pf Pf
0.4 0.4

0.2 0.2

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
PGA (g) PGA (g)
Fixed-base Fixed-base
SSI model SSI model
(c) (d)

Figure 11: Comparative fragility curves between the double-parameter and single-parameter damage models: (a) slight damage, (b)
moderate damage, (c) severe damage, and (d) complete damage.

Based on the analysis mentioned above, it can be seen distribution compared with the single-parameter index of
that SSI effect has a great impact on the fragility of the MIDR. From the slight damage state to the severe damage
structure using the MIDR as the damage index. But the state, it can be seen that the median value of MIDR is higher
MIDR only considers the structural damage caused by de- that of PA model and Niu model but is lower than that of
formation, which could underestimate the damage degree, Lu–Wang model. So the fragility of PA and Niu models is
especially under a long-lasting earthquake. However, the higher than that of MIDR, and the fragility of Lu–Wang
DPDMs taking the combination of the deformation and model is the lowest. For the complete damage state, the
energy as the damage index can reflect both the maximum median values of the three DPDMs are all far lower than that
deformation and cumulative damage effects, which is of MIDR, which leads to the failure probability of the
thought to be more comprehensive at describing the structure in this state that is much smaller than that of the
structural damage. double-parameter damage index. This may be attributed to
Table 8 summarizes the lognormal distributed fragility that the proportion of the cumulative damage on the
parameters including the median and log-standard deviation structural damage is increased, and the cumulative damage
in terms of PGA for the building considering SSI effects with gradually dominates the structural damage in this state,
three DPDMs and the MIDR. Figure 12 depicts the com- while the MIDR does not consider this part, so the fragility of
parative fragility curves for three different DPDMs of PA MIDR is far lower than that of DPDMs.
model, Niu model, and Lu–Wang model compared with that Next, regarding the influence of the SSI effects on the
of the MIDR. It is shown that adopting the double-pa- fragility of DPDMs compared with the fixed-base building,
rameter damage index behaves different fragility comparing the median values of Tables 6 and 8, as well as
Advances in Materials Science and Engineering 11

1 1

0.8 0.8

0.6 0.6
Pf Pf
0.4 0.4

0.2 0.2

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
PGA (g) PGA (g)
MIDR Niu MIDR Niu
Park–Ang Lu–Wang Park–Ang Lu–Wang
(a) (b)
1 1

0.8 0.8

0.6 0.6
Pf Pf
0.4 0.4

0.2 0.2

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
PGA (g) PGA (g)
MIDR Niu MIDR Niu
Park–Ang Lu–Wang Park–Ang Lu–Wang
(c) (d)

Figure 12: Fragility curves of the fixed-base and SSI models with MIDR:(a) slight damage, (b) moderate damage, (c) severe damage, and (d)
complete damage.

Table 8: Parameters of the fragility functions in terms of PGA for the analysed structural models with considering SSI by adopting IMDR
and DPDMs.
Median PGA (g) Dispersion β
Damage model
LS1 LS2 LS3 LS4 LS1 LS2 LS3 LS4
MIDR 0.05 0.09 0.26 1.38 0.79 0.82 0.82 0.86
Park–Ang 0.01 0.05 0.1 0.38 0.81 0.84 0.84 0.88
Niu 0.03 0.08 0.2 0.32 0.76 0.79 0.79 0.83
Lu–Wang 0.06 0.16 0.3 0.46 0.93 0.95 0.95 0.99

the fragility curves of Figures 10 and 12, it is seen that SSI to an underestimate of the seismic fragility of the structure
effects also increase the fragility of the buildings compared and will cause the structure to be unsafe, especially in the
with the fixed-base case using DPDMs. Especially for the complete damage state.
severe and complete damage states, this observation is
more noticeable. 6. Conclusions
Therefore, it is more reasonable and comprehensive to
employ the DPDMs to describe and evaluate the seismic There are two objectives presented in this study. One is to
fragility of the SSI model, while adopting the MIDR will lead investigate the SSI effects on the structural fragility
12 Advances in Materials Science and Engineering

compared with those of the fixed-base model. The results 30 years,” ISET Journal of Earthquake Technology, vol. 43,
represented in this study show that consideration of the SSI no. 3, pp. 104–175, 2006.
effects induces a higher fragility than that of the fixed-base [2] M. Shinozuka, M. Q. Feng, J. Lee, and T. Naganuma, “Sta-
model by adopting both the DPDMs and the IMDR, and it tistical analysis of fragility curves,” Journal of Engineering
will lead to the structure being unsafe. Therefore, the SSI Mechanics, vol. 126, no. 12, pp. 1224–1231, 2000.
[3] K. Z. W. Myat, Analysis on Seismic Performance of Strip
effects should not be neglected in the structural fragility
Foundation of a Low Rise Building with Soil-Structure In-
analysis.
teraction Approach, 2008.
The other objective is to supply a method for fragility [4] S. C. Dutta, K. Bhattacharya, and R. Roy, “Response of low-
assessment to study the influence of the DPDMs on the SSI rise buildings under seismic ground excitation incorporating
structural fragility analysis compared with the MIDR. First, a soil-structure interaction,” Soil Dynamics and Earthquake
series of dynamic analyses are conducted based on three Engineering, vol. 24, no. 12, pp. 893–914, 2004.
DPDMs compared with the MIDR for the building with and [5] P. E. Pinto and M. Ciampoli, “Effects of soil-structure in-
without considering SSI effects. It shows that the building teraction on inelastic seismic response of bridge piers,”
using the MIDR will lead to a significant decrease in the Journal of Structural Engineering, vol. 121, no. 5, pp. 806–814,
damage assessment compared with the DPDMs of Park– 1995.
Ang, Niu, and Lu–Wang models. With the increase in PGA, [6] P. Rajeev and S. Tesfamariam, “Seismic fragilities of non-
the difference in damage assessment using the DPDMs is ductile reinforced concrete frames with consideration of soil
more noticeable than that assessed using the MIDR. Next, structure interaction,” Soil Dynamics and Earthquake Engi-
IDA is performed by 50 scaled input ground motions, and neering, vol. 40, no. 3, pp. 78–86, 2012.
the fragility curves based on the DPDMs are obtained [7] K. D. Pitilakis, S. T. Karapetrou, and S. D. Fotopoulou,
“Consideration of aging and SSI effects on seismic vulnera-
compared with that of the MIDR. It is shown that adopting
bility assessment of RC buildings,” Bulletin of Earthquake
the DPDMs of Park–Ang, and Niu increases the fragility
Engineering, vol. 12, no. 4, pp. 1755–1776, 2014.
compared with that of the MIDR from the slight damage [8] E. Sáez, F. Lopez-Caballero, and A. Modaressi-Farahmand-
state to the severe damage state, but the Lu–Wang damage Razavi, “Effect of the inelastic dynamic soil-structure in-
model decreases the fragility of the structure in the corre- teraction on the seismic vulnerability assessment,” Structural
sponding damage state. More significantly, in the complete Safety, vol. 33, no. 1, pp. 51–63, 2011.
damage state, the MIDR greatly increases the fragility of the [9] M. E. Rodriguez and J. C. Aristizabal, “Evaluation of a seismic
structure compared with that of Park–Ang, Niu, and Lu– damage parameter,” Earthquake Engineering & Structural
Wang damage models, which will overestimate the struc- Dynamics, vol. 28, no. 5, pp. 463–477, 2015.
tural seismic capacity to lead to severe seismic losses. [10] Y. J. Park and A. H. S. Ang, “Mechanistic seismic damage
Therefore, it is more reasonable to adopt the DPDMs for model for reinforced concrete,” Journal of Structural Engi-
fragility analysis by considering SSI effects. neering, vol. 111, no. 4, pp. 722–739, 1985.
All these conclusions are based on the results of the [11] Y. J. Park, A. M. Reinhorn, and S. K. Kunnath, “IDARC:
eight-story RC building supported by pile foundation on soft inelastic damage analysis of reinforced concrete frame-shear
soil. Other influential factors (e.g., number of stories, wall structures,” Technical Report, National Center for
Earthquake Engineering Research, Taipei, Taiwan, 1987.
structural types, foundation types, and soil types) will be
[12] D. Niu and L. Ren, “A modified seismic damage model with
considered in future studies.
double variables for reinforced concrete structures,” Earth-
quake Engineering and Engineering Vibration, vol. 16, no. 4,
Data Availability pp. 45–55, 1996, in Chinese.
[13] D. Lu and W. Guangyuan, “Decision making method for
The data used to support the findings of this study are in- optimal design of earthquake resistant structures based on
cluded within the article. damage performance,” Civil Engineering Journal, vol. 34,
no. 1, pp. 44–49, 2001, in Chinese.
Conflicts of Interest [14] D. Lu, “Intelligent optimization design of disaster resistant
structures based on optimal fortification load and perfor-
The authors declare that they have no conflicts of interest. mance,” Doctoral Dissertation, Harbin University of Archi-
tecture, Heilongjiang, China, 1999.
[15] X. Du and J. Ou, “Seismic damage assessment model of
Acknowledgments buildings,” World Earthquake Engineering, vol. 7, no. 3,
pp. 52–58, 1991, in Chinese.
This work was partially developed in the framework of the [16] J. Ou, Z. He, B. Wu, and L. Xu, “Seismic damage control
research project funded by the Sichuan International Science design of reinforced concrete structures,” Journal of Buildings
and Technology Innovation Cooperation/Hong Kong, Structures, vol. 21, no. 1, pp. 63–67, 2000, in Chinese.
Macao and Taiwan Science and Technology Innovation [17] J. M. Bracci, A. M. Reinhorn, J. B. Mander et al., “De-
Cooperation (2019YFH0120). terministic model for seismic damage evaluation of reinforced
concrete structures,” Technical Report, State University, New
References York, NY, USA, 1989.
[18] Y. Li, B. Liu, and Q. Shi, “Performance levels and estimation
[1] G. M. Calvi, R. Pinho, G. Magenes et al., “Development of indices of structures,” World Earthquake Engineering, vol. 19,
seismic vulnerability assessment methodologies over the past no. 2, pp. 132–137, 2003, in Chinese.
Advances in Materials Science and Engineering 13

[19] National Standard of the People’s Republic of China

(NSPRC), Code for Seismic Design of Buildings (GB50011-
2010), Ministry of Construction of Peoples Republic of China,
Beijing, China, 2010.
[20] National Standard of the People’s Republic of China
(NSPRC), Code for Concrete Structure Design (GB 50010-
2010), Ministry of Construction of Peoples Republic of China,
Beijing, China, 2010.
[21] D., Kong, Study on Equivalent Method of Reinforced Concrete
Material Properties, Shenyang Jianzhu University, Shenyang,
China, 2003.
[22] Y. H. Zhao and G. J. Weng, “Effective elastic moduli of
Ribbon-reinforced Composites,” Journal of Applied Me-
chanics, vol. 57, no. 1, pp. 158–167, 1990.
[23] Y. H. Zhao, G. P. Tandon, and G. J. Weng, “Elastic moduli for
a class of porous materials,” Acta Mechanica, vol. 76, no. 1-2,
pp. 105–131, 1989.
[24] Z. Pan, Research and Application of Reinforced Concrete
Equivalent Material, Harbin Institute of Technology, Hei-
longjiang, China, 2008.
[25] Z. Lu and B. Jiang, “Study on range of mesh about finite
element for back analysis of displacement,” China Civil En-
gineering Journal, vol. 32, no. 1, pp. 26–30, 1999, in Chinese.
[26] D. C. Drucker, W. Prager, and H. J. Greenberg, “Extended
limit design theorems for continuous media,” Quarterly of
Applied Mathematics, vol. 9, no. 4, pp. 381–389, 1952.
[27] D. C. Drucker and W. Prager, “Soil mechanics and plastic
analysis or limit design,” Quarterly of Applied Mathematics,
vol. 10, no. 2, pp. 157–165, 2013.
[28] D. Vamvatsikos and C. A. Cornell, “Incremental dynamic
analysis,” Earthquake Engineering and Structural Dynamics,
vol. 31, no. 3, pp. 491–514, 2002.
[29] Y. K. Wen, B. R. Ellingwood, and J. M. Bracci, “Vulnerability
function framework for consequence-based engineering,”
Technical Report No.DS-4, Mid-America Earthquake Center
(MAE), University of Illinois at Urbana-Champaign,
Champaign, IL, USA, 2004.
[30] National Institute of Building Science, Direct Physical
Damage-General Building Stock, HAZUS-MH Technical
Manual, Chapter 5, Federal Emergency Management Agency,
Washington, DC, USA, 2004.
[31] B. R. Ellingwood, O. C. Celik, and K. Kinali, “Fragility as-
sessment of building structural systems in Mid-America,”
Earthquake Engineering and Structural Dynamics, vol. 36,
no. 13, pp. 1935–1952, 2007.
[32] X. Yu, D. Lu, and B. Li, “Estimating uncertainty in limit state
capacities for reinforced concrete frame structures through
pushover analysis,” Earthquakes and Structures, vol. 10, no. 1,
pp. 141–161, 2016.
[33] S. T. Karapetrou, S. D. Fotopoulou, and K. D. Pitilakis,
“Seismic vulnerability assessment of high-rise non-ductile RC
buildings considering soil-structure interaction effects,” Soil
Dynamics and Earthquake Engineering, vol. 73, pp. 42–57,
2015.
 Nanomaterial
Nanomaterials
Journal of

The Scientific Journal of International Journal of

World Journal
Hindawi Publishing Corporation
Applied Chemistry
Hindawi Hindawi
Scientifica
Hindawi
Polymer Science
Hindawi
http://www.hindawi.com
www.hindawi.com Volume 2018
2013 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Advances in Advances in
Chemistry
Hindawi
Physical Chemistry
Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Submit your manuscripts at

www.hindawi.com

International Journal of Advances in

Analytical Chemistry Condensed Matter Physics
Hindawi Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Engineering
Journal of
Journal of International Journal of Advances in Journal of
Chemistry
Hindawi
Biomaterials
Hindawi
High Energy Physics
Hindawi Hindawi
Nanotechnology
Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

International Journal of Journal of BioMed Advances in Advances in

Corrosion
Hindawi
Materials
Hindawi
Research International
Hindawi
Tribology
Hindawi
Materials Science and Engineering
Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018