Vol. 19, No.

3 EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION July, 2020

Earthq Eng & Eng Vib (2020) 19: 561-571 DOI: https://doi.org/10.1007/s11803-020-0580-z

Seismic resilience of retrofitted RC buildings

Ghazanfar Ali Anwar† and You Dong‡
Department of Civil and Environmental Engineering, The Hong Kong Polytechnic University, Hong Kong, China

Abstract: Existing buildings can be at a greater seismic risk due to non-conformance to current design codes and may
require structural retrofitting to improve building performance. The performance of buildings is measured in terms of
immediate consequences due to direct damage, but the continuing impacts related to recovery are not considered in seismic
retrofit assessment. This paper introduces a framework of retrofit selection based on the seismic resilience of deficient buildings
retrofitted with the conventional mitigation approaches. The assembly-based methodology is considered for the seismic
resilience assessment by compiling a nonlinear numerical model and a building performance model. The collapse fragility is
developed from the capacity curve, and the resulting social, economic, and environmental consequences are determined. The
seismic resilience of a building is assessed by developing a downtime assessment methodology incorporating sequence of
repairs, impeding factors, and utility availability. Five functionality states are developed for the building functionality given
investigated time interval, and a functionality curve for each retrofit is determined. It is concluded that seismic resilience can
be used as a performance indicator to assess the continuing impacts of a hazard for the retrofit selection.

Keywords: resilience; reinforced-concrete; performance-based; seismic retrofit; non-ductile

1 Introduction during the design and construction of these buildings

(Gautam and Chaulagain, 2016). The seismic loss
Seismic hazard is a low probability high consequence can be significant for the deficient existing buildings,
event, which until 1961 was considered in the seismic particularly in low-to-medium seismicity regions where
provisions of Uniform Building Code (UBC) as a the seismic codes have not been adopted. This highlights
single lateral load on a building equivalent to 7.5% the importance of improving the performance of existing
of the buildings weight (10% in the case of poor soil deficient structures to reduce seismic consequences and
conditions). It was not until 1960 when seismic design increase resilience.
provisions were made mandatory for the communities Repair, rehabilitation, and retrofitting is used to
in USA adopting UBC codes. Two major earthquakes, improve the performance of existing buildings. Recently,
Alaska 1964, and San Fernando 1971, revealed poor it has become an important construction activity,
structural performance of buildings which resulted in considering that the amount of money spent globally on
developing new seismic design procedures (Beavers, repair and rehabilitation of existing structures is higher
2002). Improving seismic design codes is an ongoing than new constructions (Ma et al., 2017). The retrofitting
process, and it is not until recently that the seismic techniques include adding lateral force-resisting systems
design codes are being implemented in low-to- or upgrading the existing elements for structural
moderate seismicity regions (Mwafy and Elkholy, performance improvement (Zheng et al., 2019). The
2017). The existing building infrastructure around the upgrading of existing elements can be implemented by
world is therefore at risk of poor performance in an either reducing the demands on a lateral force-resisting or
earthquake event due to inadequate structural detailing improving the capacity, achieved by modifying strength,
and inefficient seismic design provisions implemented stiffness, ductility properties or through any of these
combinations (Thermou and Elnashai, 2006). Ductility
depends on the detailing of structural components;
Correspondence to: You Dong, Department of Civil and
therefore, its retrofitting would require improving beam-
Environmental Engineering, The Hong Kong Polytechnic
University, Hung Hom, Kowloon, Hong Kong, China
column joints and rebar reinforcements, which can be
E-mail: [email protected] disruptive and expensive. Hence this type of retrofitting
†
PhD; ‡Assistant Professor is rarely used in the low-to-medium seismicity region
Supported by: Chinese National Engineering Research Centre (Calvi, 2013). A more desirable approach for ductility
(CNERC) for Steel Construction (Hong Kong Branch) at the related retrofitting is to reduce the demands on the
Hong Kong Polytechnic University under Project No. P0013864; structure by modifying or replacing lateral force resisting
Programme Code: BBV9) members. This study is related to improving the strength
Received April 6, 2020; Accepted June 5, 2020 and stiffness of existing lateral force resisting members
562 EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION Vol. 19

by using Reinforced Concrete Jacketing (RCJ), Steel structural models, high-rise buildings, and in cases where
Jacketing (SJ), and Fiber-Reinforced-Polymer wrapping buildings have to be analyzed several times. According
around columns, which is a commonly utilized approach to the authors’ best knowledge, seismic resilience
(Billah and Alam, 2014). assessment of deficient reinforced concrete buildings
Performance-based assessment is used for the retrofitted with conventional mitigation approaches
seismic upgrading of existing buildings. Performance has not been studied, especially through assembly-
is expressed in terms of discrete performance levels based quantification of downtime. Additionally, the risk
defined as immediate occupancy, life safety, and assessment indicators only consider the robustness of a
collapse prevention. The performance levels are structure, while the resilience indicator also considers
correlated with social, economic and downtime losses, the recovery of a building. In this paper, retrofit selection
but these correlations are observation based or empirical is investigated based on the seismic resilience indicator.
in nature, and are site specific (Whitman et al., 1997). In this paper a performance-based seismic resilience
This approach to risk reduction requires threshold limit assessment framework is presented applied to a deficient
state values which cannot be precisely determined reinforced concrete building. The increase in seismic
for various type of buildings, since they depend on resilience is investigated by applying three conventional
several factors, such as structural configurations, design structural mitigation approaches commonly utilized for
criteria, importance factors, level of detailing, among the structural retrofit of a building. The methodology
others (Qian and Dong, 2020). The recovery time of a considers a component-level approach, which requires
building, which is a key input in the seismic resilience assembling a fragility and consequence functions in
assessment, is also related to building performance building performance model. The proposed assembly-
levels, which are mostly presented in crude terms (e.g., based component-level approach considers the collapse
the most widely used HAZUS risk-assessment platform fragility, determined from pushover analysis, hence
assumes the building to achieve full functionality bypassing computationally expensive time history
within one year, irrespective of the amount of damage analyses. Social, economic, and environmental
and hazard scenario). Numerous studies have adopted consequences are assessed in terms of casualties,
a performance-based seismic assessment approach for monetary loss, and equivalent carbon emissions. The
risk and resilience evaluation (Dong and Frangopol, seismic resilience for retrofit alternatives are assessed
2015; Zheng et al., 2018; Li et al., 2020a; Kilanitis by developing a downtime assessment methodology
and Sextos, 2019; Giouvanidis and Dong, 2020; Li incorporating sequence of repairs, impeding factors,
et al., 2020b), also linking to seismic sustainability and utility availability. The key contribution of this
(Rodriguez-Nikl, 2015; Bocchini et al., 2013; Dong et paper includes the development of collapse fragilities
al., 2014). A component-level approach incorporating for conventional retrofit alternatives utilizing
seismic loss, sustainability and resilience has also been computationally efficient pushover analysis, the
investigated by many researchers (Dong and Frangopol, development of a framework for social, economic and
2016; Hashemi et al., 2019; Anwar et al., 2020; Asadi et environmental consequence assessment for considered
al., 2019). Tirca et al. (2016) investigated improvement retrofit alternatives, and retrofit selection based on
in seismic resilience through local modifications of the seismic resilience taking into account the robustness
components of office buildings. Incremental Dynamic and recovery of a building. The performance-based
Analysis (IDA) was used to develop damage fragilities, seismic resilience assessment methodology is presented
and functionality curves developed by Cimellaro et al. in section 2, and an illustrative example is presented in
(2010) were used to evaluate seismic resilience. Guo et section 3. Final section presents the conclusion of the paper.
al. (2017) studied seismic resilience of a frame building
retrofitted with self-centering walls with friction
devices. The performance of a building was compared 2 Methodology
through engineering demand parameters (EDPs), but a
quantification framework for seismic resilience was not The framework begins by selecting a building and
considered. Similar studies can be found in the literature retrofit methods used for investigating and enhancing
for the seismic resilience improvements considering seismic resilience. The first step is to develop the
seismic retrofit (Pekcan et al., 2014; Vona et al., 2018; nonlinear models for the reference un-retrofitted and
Khanmohammadi et al., 2018; Rousakis, 2018; Anelli the retrofitted buildings. The nonlinear model should be
et al., 2019), but none utilizes a performance-based able to effectively capture the steel yielding, concrete
quantification framework of resilience assessment. crushing, strength, and stiffness degradation. The
Molina Hutt et al. (2016) proposes a seismic loss capacity curve representing base shear given lateral
and downtime assessment approach for increasing displacement can then be developed from the nonlinear
seismic resilience for tall buildings by utilizing IDA, static analysis procedure and is used to estimate the
which employs series of time history analyses with deficiencies in the lateral force resisting system. The
increasing intensity measure levels, and which can be capacity curves are developed by applying a series of
computationally expensive, particularly for complex lateral loads with increasing magnitude and recording
No. 3 Ghazanfar Ali Anwar et al.: Seismic resilience of retrofitted RC buildings 563

the lateral displacements. Increasing the lateral loads (2006) investigated a series of single-degree-of-freedom
in each iteration will eventually cause elements to start systems with a wide range of time periods through
to yield, and as a result of each yielding of structural incremental dynamic analysis. The resulting hysteresis
members, the redistribution of loads will take place (Su loops were converted to backbone curves ranging from
et al., 2019). The model is revised in each iteration by simple bilinear to quadrilinear, comprising an elastic,
adjusting the member yielding, strength and stiffness hardening, softening, and a residual plateau segment
degradation, and the process continues till the yield that ends at a zero-strength. The relationship between
pattern and strength and stiffness degradation for the the characteristic segments of IDA curves were linked
whole structure is identified. The maximum base shear to the backbone curves of many systems, suggesting
and the lateral displacements are identified and compared that nonlinear static analysis procedure (i.e., pushover
with the design loads and a strength factor is determined. analysis) can be used to estimate nonlinear dynamic
If the strength factor is greater than one or within the response. In this paper, the pushover analysis is used
desirable limits of the codes, the structure is considered to estimate nonlinear IDA results by utilizing the static
safe; otherwise structural retrofitting is required. The pushover to incremental dynamic analysis (SPO2IDA)
capacity curves are developed for the retrofitted models tool. FEMA (2012) recommends that this tool can be
using the same procedure (i.e., pushover analysis). If used to develop collapse fragilities for low-rise buildings
the strength factors are not desirable, then the retrofit dominated by the fundamental mode of vibration. This
techniques are revised and the process is repeated to method can bypass the computationally expensive part
achieve the desirable preliminary performance. The of the methodology and can rapidly generate the collapse
methodology is presented in a flowchart as shown in fragility. Following are the steps to develop collapse
Fig. 1. fragility using the SPO2IDA tool.
1. Develop a suitable nonlinear mathematical
2.1 Developing collapse fragilities from pushover model of a structure for the pushover analysis.
2. Perform a nonlinear static analysis procedure
If the retrofit techniques satisfy the preliminary to develop capacity curve in the principle building
strength and stiffness requirements, the next step direction.
is to develop the collapse fragilities and building 3. Approximate the capacity curve into
performance model. Vamvatsikos and Allin Cornell quadrilinear curve by identifying four control points
each indicating the endpoint and the start point of the
four defined segments.
4. Execute the SPO2IDA tool and input the control
points and relevant information (e.g., building weight,
building height, fundamental time period etc.), and
extract the median collapse capacity.
5. Construct the collapse fragility using lognormal
cumulative distribution function with a dispersion of 0.6.

2.2 Consequence assessment

The collapse fragility analysis provides information

on the probability of collapse given an intensity measure.
It is more interesting for decision makers to obtain
information that is more meaningful (e.g., economic
loss in terms of dollars, casualties in terms of numbers,
equivalent carbon emissions etc.). In the consequence
assessment, collapse fragility and the probability of
damage to components of a building are converted to
social, economic, and environmental consequences
(Dong et al., 2016; Wang et al., 2018). For that purpose,
a building performance model is assembled, comprising
fragility functions and consequence functions for
damageable structural and non-structural components.
Fragility functions determine the probability of
exceeding given damage states for each damageable
component. Consequence functions use the probabilities
of components being in different damage states, and
Fig. 1 The methodology for assessing seismic resilience using determine the social, economic, or environmental
nonlinear static procedure consequences. The following steps can determine
564 EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION Vol. 19

economic and environmental consequences given a time as shown in Eq. (3)

hazard scenario.
1. Define a hazard scenario against which TR
Q ( t ) dt
consequences are to be determined
2. Evaluate EDPs from the developed nonlinear
R= ∫0
TR
(3)

mathematical model.
3. Determine probability of exceeding different where R is the resilience metric, Q ( t ) is the
damage states for all the damageable components. functionality, and TR is the investigated time interval
4. Utilize the probability of exceeding different after an earthquake.
damage states for all the damageable components, and In this paper five functionality states are developed
the collapse fragility to determine consequences using depending upon the structural and non-structural damage
total probability theorem. and utility availability. The mapping of the functional
The social consequences (i.e., injuries, fatalities) can states and the recovery to full functionality is presented in
be determined as a flowchart shown in Fig. 2. Five functionality states are
represented mathematically by a designated weighting
Sm|IM = ∅CTrand f ( p | Trand ) pT pR pC|IM (1) factor. Full-functionality (FF) is assigning a weighting
factor of 1, and the Restricted-entry (RE) is given a
weighting factor of 0.2. The remaining functionality
where Sm|IM is the social metric of seismic sustainability;
states are assigned weighting factors between 0.2 and 1
∅C is the casualty function, which depends on the type
with an increment of 0.2. After an earthquake event, the
of construction and can be determined using historical process starts with the inspection of a building, which
casualties from past earthquakes; Trand is the randomly is performed by a professional building inspector. In
generated time of the day and day of the week for a this paper the structural and non-structural damage is
particular realization; f ( p | Trand ) is the time-dependent computed using fragility functions to quantify the extent
population model; pT is the total population of a building; of damage to the building. Depending on the extent of
pC|IM is the probability of collapse of a building given damage and the information about the utility availability,
IM; and pR is the population at risk depending upon the the initial functional state of a building can be
failure mode of a building. determined. For example, if the building has experienced
The economic and environmental consequences can moderate to extensive structural damage, the building
be determined as is tagged in the Restricted-entry (RE) functional state
∞ (i.e., the occupants are not allowed to enter the premises
before the necessary repairs). The logical sequence of
CLT|IM ∑∫C
DS 0
LR|DS pDS|EDP f EDP|IM dEDP ⋅
repair is designed in the next step to bring the building
functionality to pre-hazard state. Before the building
(1 − pC|IM ) + CL C|C
. pC|IM (2) repairs, additional delays, called impeding delays, will
occur due to financing, engineering reviews, permitting,
where CL is the total consequence given IM; CLC|C contractor mobilization and sometimes long lead times.
T|IM
The building is tagged as Restricted-entry (RE) if it
is the consequence given probability of collapse; pC|IM
suffers moderate to extensive structural damage, and if
CLR|DS is the random value of a consequence loss function the building only suffers non-structural damage, then the
of a component for a given damage state; pDS|EDP is the building is in Restricted-use (RU) functional state. The
probability of damage state given EDP; and f EDP|IM is building will be recovered to full functionality after the
the probability density function of EDP given IM. impeding delays, necessary non-structural repairs, and
availability of all the utilities. If minor or no damage
2.3 Seismic resilience assessment is observed, then, depending upon the availability of
utilities, a building is assigned as one of the remaining
Seismic resilience is the ability of a structure to three functional states. If no utility is available, the
absorb damage without suffering collapse and to recover building is in Re-occupancy (RO) functional state (i.e.,
from the earthquake hazard efficiently. The building with the building space can be occupied for shelter purposes
greater seismic resilience would have less damage in an but cannot be utilized for its intended purpose). If only
immediate aftermath of an earthquake and would recover critical utilities are available (i.e., electricity and water),
faster. Functionality of a building after an earthquake then the building is Baseline-functional (BF), and the
and its recovery can be used as a performance indicator building will achieve Full-functionality (FF) after the
for assessing seismic resilience. The functionality curve availability of all the utilities.
provides the functionality state given the investigated The repair time required for the repair of each
time interval and its recovery to full functionality after a damaged structural and non-structural component can
hazard event. Seismic resilience can be mathematically be determined from Eq. (2). The downtime for each
evaluated by integrating the functionality curve over functional state can be determined by considering the
No. 3 Ghazanfar Ali Anwar et al.: Seismic resilience of retrofitted RC buildings 565

repair schedule (i.e., sequence of repairs determined into number of fibers to effectively monitor the stresses
from the repair times of all the damageable components), and strains of different sections of elements. A uniaxial
impeding delays (i.e., financing, engineering review and constant confinement concrete model is utilized for the
permitting, contractor mobilization and long lead times) reinforced concrete jacketing, a bilinear steel model
and the utility availability. The impeding delays and the with constant strain hardening is utilized for the steel
utility availability are considered in this paper through jacketing, and a uniaxial trilinear fiber-reinforced plastic
the lognormal distribution functioned developed by model is used for the fiber-reinforced polymer overlays.
Almufti and Willford (2013) in the REDi Rating System
(Resilience-based Earthquake Design Initiative for the 3.1 Developing collapse fragilities from pushover
Next Generation of Building). The functionality curve
can be developed after determining the downtime for Pushover analysis is performed on a model by
each functional state, and the seismic resilience can be applying an increasing inverted triangular lateral loads
evaluated. pattern, representing deformation of a building under
fundamental mode, and evaluating the maximum lateral
displacements. Information on the structure’s strength,
3 Illustrative example stiffness and ductility can be extracted from the resulting
capacity curve of a building, and a strength factor can be
The non-ductile reinforced concrete building selected determined to evaluate the performance of a building and
for this illustrative purpose is a two-story intermediate retrofit methods (Elkady and Lignos, 2015). Figure 4
moment-resisting frame building with a total height of shows the capacity curve of the reference building and
8.5 m. The residential building is designed according the considered retrofit techniques. The capacity curve
to the building codes implemented at the time of its gives important information about member yielding,
design and construction, which largely ignored seismic stiffness, and ultimate strength of a building structure.
provisions, and in which only wind loads are considered The ultimate strengths are compared with the design
in the design of a building against the lateral loads. The strength, and the strength factor is determined, which is
concrete strength of 20 Mpa and a mild steel with yield
strength of 240 MPa is used for the design, resulting
in large cross-sections, increased weight, and stiffness.
Three retrofitting techniques, namely, Reinforced
Concrete Jacketing (RCJ), Steel Jacketing (SJ), and Fiber
Reinforced Polymer (FRPs) overlays are considered for
improving performance of a non-ductile building. The
considered seismic retrofit techniques require modifying
the existing lateral force-resisting components (i.e.,
column in the considered example). The enhancement
of the cross sections follow FEMA-547 (2006) and
ASCE-41-13 (2013) recommendations, which explicitly
highlight the detailing, construction practices and the
seismic evaluation of existing buildings. The layout and Fig. 2 Functionality states and recovery considering structural
the design details of a building are shown in Fig. 3. and non-structural damage, impeding delays, sequence
Ten fiber-based nonlinear models are developed, one of repairs and utility availability
for the reference un-retrofitted structure and nine models
for the retrofitted structures (i.e., three retrofit models
for each retrofit technique). The numerical models are
developed in an open source nonlinear analysis platform
ZUES-NL (Jeong and Elnashai, 2005). The built-in
nonlinear material models are used to represent concrete
and steel behavior. The nonlinear concrete material
model with crushing strain of 0.02 and a confinement
factor of 1.05 is used depending upon the reinforcement
details. A bilinear elasto-plastic model with kinematic
strain hardening is used for the steel material modeling.
The material and geometric nonlinearities, P-delta effects
and large displacements are considered. The models for
the reinforced concrete jacketing, steel jacketing, and
FRP overlays are represented by modeling sections of the
columns into reinforcing steel, confined, and unconfined
concrete regions. The element cross sections are divided Fig. 3 Building plan and structural details
566 EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION Vol. 19

the ratio of the design strength to ultimate strength. If factors for the SJ retrofit with steel jacket thicknesses of
the strength factor is more than one, then the building 3 mm, 5 mm, and 10 mm are 2.40, 2.99 and 3.98. It is
is satisfactory; otherwise retrofit techniques are used to interesting to note that steel jacketing has greater impact
improve the strength factor. The non-ductile reinforced in increasing the lateral capacity of a building, while
concrete building is designed only to resist gravity and FRPs provide comparatively the least improvement in the
wind loads, since before 1991 the region was classified ultimate lateral capacity. Nonetheless, all the considered
as zone ‘0’, and the lateral seismic loads were not retrofit techniques provide satisfactory strength factors
considered during the design process. According to the (i.e., greater than one).
revised zone classification of the region, static lateral The capacity curves are converted into idealized
force procedure provides a required design strength of curves, a bilinear approximation is provided in Fig. 4(a),
655 kN, and the ultimate strength determined from the and more details on idealization from the capacity curve
capacity curve is 605 kN. Since the ultimate strength can be obtained from Elnashai and Di Sarno (2008).
is less than the required design strength of a building, The four segments of the idealized curve will give four
the reference building is not conforming to the design control points, which are used as an input in the SPO2IDA
requirement of the current code of practice. RCJ retrofit tool, and the median and dispersion values for the
with retrofit thickness of 50 mm, 75 mm, and 100 mm gives collapse fragilities are evaluated. Lognormal cumulative
the strength factors of 2.69, 3.13, and 3.57. Similarly, distribution function is then used to develop collapse
the strength factors for the FRP retrofit for one, two and fragilities for each model. Figure 5 shows the collapse
three layers are 1.54, 2.01, and 2.12, and the strength fragilities developed by using pushover analysis. It is
Base shear (kN)

(a) (b)
Base shear (kN)

Roof displacement (mm) Roof displacement (mm)

(c) (d)
Fig. 4 Capacity curve for (a) reference structure, (b) reinforced concrete jacketing, (c) FRP overlays, and (d) steel jacketing
P(Collapse|IM)

Spectral acceleration Sa(T) Spectral acceleration Sa(T)

(a) (b)
P(Collapse|IM)

Spectral acceleration Sa(T)

(c)
Fig. 5 Collapse fragilities for (a) FRP overlays, (b) reinforced concrete jacketing, and (c) steel jacketing
No. 3 Ghazanfar Ali Anwar et al.: Seismic resilience of retrofitted RC buildings 567

noted that SJ retrofit reduces the probability of collapse acceleration-sensitive components. The components
significantly, RCJ retrofit also significantly reduces the partitions, finishes, and glazing are sensitive to lateral
probability of collapse, while for the FRP retrofit, the story drifts, and ceiling and sprinklers are sensitive
reduction in the probability of collapse is not significant. to floor accelerations. The social consequences are
Nonetheless, probability of collapse is reduced for all determined by constructing a population model, and
the considered retrofit techniques. defining casualty function and the population at risk.
The time dependent population model represents the
3.2 Consequence assessment percentage of people present during the time of the
day, and day of the week for a given realization. The
Consequence assessment starts with selecting a casualty function for the reinforced concrete residential
hazard scenario and assembling a building performance construction indicates that 90% will suffer fatalities in the
model. The hazard scenario with a design PGA of 0.16 g event of collapse, and 10% will suffer a major injury in
is selected for the case of this illustrative example. In the case of reinforced concrete frame structure (FEMA,
order to investigate the variation of social, economic 2012). Figure 6 shows the social losses in terms of total
and environmental consequences with varying intensity number of expected fatalities given four scenarios. The
measure, four hazard scenarios are considered for the social losses for the reference un-retrofitted building
consequence assessment (i.e., half the design hazard has the highest number of expected fatalities. Applying
scenario, twice the design hazard scenario, and four retrofit reduces the social losses, with SJ and RCJ being
times the design hazard scenario). Three retrofit the most effective in reducing the social consequences.
techniques (i.e., FRPs with 1 layer, RCJ with 75 mm of In order to evaluate the economic and environmental
jacket thickness, and SJ with 3 mm of jacket thickness) losses, structural analyses of a nonlinear building models
are considered for the consequence and resilience are performed and engineering demand parameters (i.e.,
assessment. story drifts and accelerations) are extracted for each
The building performance model consists of fragility story, correlated with damage through fragility functions
functions and consequence functions. Fragility functions and consequences through consequence functions.
relate the structural analysis results to the damage, The total economic and environmental consequences
and consequence functions translate the damages into
social, economic, and environmental consequences. The
fragility and consequence functions used in this example
are extracted from (FEMA, 2012; Hashemi et al., 2019;
Fatalities

Mitrani-Reiser, 2007), and are shown in Table 1. The

fragility and consequence functions for various types
of retrofitted structural components is not yet available
in the literature. Therefore, in this illustrative example,
conventional fragility and consequence functions are Intensity measure (PGA)
utilized for the retrofitted buildings. Fig. 6 Social consequence in terms of expected fatalities
The components are divided into drift-sensitive and given IM

Table 1 Fragility functions and consequence functions of damageable components

Fragility Consequence functions
Components Quantity per Damage functions
floor state Economic (USD) Environmental (kgCO2) Repair time (Day)
Median CoV. Median CoV. Median CoV. Median CoV.
Structural 20 units DS1 1.75 0.40 6270 0.39 1.794 0.4 18.9 0.46
columns DS2 2.25 0.40 9540 0.32 1.794 0.4 28.7 0.40
DS3 3.22 0.40 11580 0.30 19.73 0.4 35.3 0.39
Partition 6 m2 × 22 DS1 0.39 0.17 115 0.20 12.72 0.4 0.136 0.30
DS2 0.85 0.23 679 0.10 25.52 0.4 0.797 0.30
Finish 6 m2 × 44 DS1 0.39 0.17 115 0.20 1.336 0.4 0.135 0.51
DS2 0.85 0.23 321 0.10 2.686 0.4 0.376 0.61
Glazing 2.8 m2 × 5.654 DS1 4.00 0.36 564 0.17 96.30 0.4 0.582 0.29
DS2 4.60 0.36 564 0.17 183.2 0.4 0.582 0.40
Ceiling 232 m2 × 0.22 DS1 0.35 0.40 4541 0.40 1.023 0.4 5.699 0.63
DS2 0.55 0.40 37612 0.50 5.846 0.4 47.05 0.40
DS3 0.80 0.40 70769 0.55 19.73 0.4 88.40 0.40
Sprinklers 4 m × 8.8 DS1 0.32 1.40 1154 0.37 58.07 0.4 1.227 0.80
568 EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION Vol. 19

kgCO2 ×100000
USD ×100000

Intensity measure (PGA) Intensity measure (PGA)

(a) (b)
Fig. 7 Consequences (a) economic in terms of monetary loss, and (b) environmental in terms of kgCO2 emissions

cumulative distribution functions. The utility disruption

curves represent the restoration of utilities to the building
resileince
Expected resileince

and are determined from previous earthquake data and

simulation studies (Almufti and Willford, 2013). The
Expected

utility disruptions depend on the amount of local damage

to the distribution system and are considered through
repair rate (RR), which is computed based on the peak
Investigated time TR (days) ground velocity at a building site. The related lognormal
distribution function is selected for repair rates greater or
Fig. 8 Expected resilience of a reference un-retrofitted building less than 0.2 repairs/km, as shown in Table 2.
under given scenarios In a pre-hazard state, the building is performing
its intended purpose and is in full functional state
determined from Eq. (2) are shown in Fig. 7. The (i.e., all the utilities are available and no structural or
economic and environmental consequences increase non-structural damage hinders the normal intended
with increasing IM levels. The un-retrofitted structure functions). After an earthquake event, the building can
has the highest consequences, reduced using retrofit be in any of the functional states as presented in
techniques. Comparatively, the percent reduction in the Fig. 2, depending upon the structural and non-structural
social, economic, and environmental consequences is damage and utility availability. The functionality state
highest for the 0.16 g and 0.32 g hazard scenario, and recovery times can be evaluated, and a functionality
lowest for the 0.08 g and 0.64 g hazard scenario. In the recovery curve can be generated, which gives the
given illustrative example, SJ and RCJ are more effective propagation of functional states to full functionality
in reducing the consequences for the design and twice given the investigated time interval. The functionality
the design seismic hazard scenario. curve can be utilized to develop resilience using Eq. (3).
The resilience of a reference building determined for the
3.3 Seismic resilience assessment given four scenarios is shown in Fig. 8. It is observed that
for a hazard scenario with maximum PGA of 0.08 g, the
The first step in evaluating the seismic resilience building showed better resilience, but for the rest of the
is to extract the repair times for all the damageable hazard scenarios it showed poor resilience. In the hazard
components of a building. Table 2 shows the repair scenario of 0.64, the building has negligible expected
time functions given damage state, utilized to determine resilience even at 500 days of investigated time interval,
repair times for all the components for a given story. showing that the building has collapsed and cannot be
The next step is to develop a logical repair sequence for repaired.
the downtime of a building. The building repair starts Applying the retrofit reduces the damage, hence
with repairing the structural components serially (i.e., improving the functionality curves and seismic resilience.
structural components of the first story are repaired Figure 9 shows the functionality curves and the resulting
first, before moving to the higher stories). Not all non- seismic resilience of the reference building along
structural components can be repaired simultaneously with the retrofit techniques applied. The reference un-
(e.g., to repair ceilings the sprinklers need to be repaired retrofitted building at a PGA of 0.32 g takes an expected
first, and in order to do finishes, partitions needed to be 272.5 days to achieve full functionality, which is reduced
repaired). In the example considered, partition, glazing to 260.5, 107, and 85.5 days after applying FRP, RCJ,
and sprinklers are simultaneously repaired in parallel, and SJ retrofits. The improvement in seismic resilience
followed by finish and ceilings. Additional delays due in the case of FRP retrofit techniques is negligible, while
to impeding factors (i.e., delays due to inspection, significant improvement is observed for the RCJ, and SJ
engineering mobilization, financing, contractor retrofit techniques. Since, seismic resilience is a function
mobilization, and permitting), and utilities (i.e., water, of collapse fragility, EDPs, fragility functions, and the
gas, and electricity) are considered using lognormal consequence functions. It is observed that applying RCJ
No. 3 Ghazanfar Ali Anwar et al.: Seismic resilience of retrofitted RC buildings 569

Table 2 Impeding factors for delay and utility disruption curves

Impeding Factors and utility

Mitigation measures Damage conditions Median CoV.
system
Inspection BORP Equivalent - 1 day 0.54
Engineering mobilization Engineer on contract Minor 2 weeks 0.32
Extensive 4 weeks 0.54
Financing Pre-arranged credit - 1 week 0.54
Contractor mobilization GC on contract Minor 3 weeks 0.66
Extensive 7 weeks 0.35
Permitting GC on contract Minor 1 week 0.86
Extensive 8 weeks 0.32
Electricity system - - 3 days 1.0
Water system RR <= 0.2 repairs/km - 4 days 0.5
RR > 0.2 repairs/km 21 days 1.0
Natural gas system RR <= 0.2 repairs/km - 10 days 0.5
RR > 0.2 repairs/km 42 days 0.6
RR = Repair rate, BORP = Building resumption program, GC = General contractor
Functionality state

Expected resileince

Investigated time TR (days) Investigated time TR (days)

(a) (b)
Fig. 9 Seismic hazard scenario of 0.32g showing (a) functionality curves, and (b) seismic resilience

and SJ can effectively reduce the collapse fragility and required to improve the performance of a building.
the demands on EDPs as compared to FRPs. As a result, 2. Three retrofit techniques, namely, RCJ, SJ
the seismic resilience for RCJ and SJ is larger compared and FRPs, were used for improving the performance
with the FRP retrofit alternative. of a deficient building. Capacity curves for the retrofit
buildings showed improved strength factors, hence
4 Conclusions improving the overall seismic performance of a building.
The SJ retrofit technique significantly improved the
This paper presents a performance-based performance of a building, followed by the RCJ retrofit.
methodology for evaluating seismic resilience under The FRPs also improved the performance above the
conventional structural retrofit techniques. The social, acceptable code performance, but comparatively the
economic, and environmental consequences are performance improvement was not significant.
evaluated and compared for a reference un-retrofitted, 3. The social, economic, and environmental
and a retrofitted building. It is concluded that applying consequences for the reference and retrofit buildings were
retrofit techniques reduces the probability of collapse, assessed in term of casualties, monetary loss in USD, and
social, economic, and environmental consequences. The equivalent carbon emissions. The consequences were
repair times of a building’s components are also reduced, reduced significantly by applying SJ jacketing, followed
hence improving the seismic resilience. by the RCJ. In the case of FRP retrofit, the reduction in
The following conclusions can be drawn. consequences were not significant.
1. Pushover analysis provides important 4. The seismic resilience assessment considers
information on a structure’s strength, stiffness and component-level repair time of a building considering
ductility, which can be used for preliminary evaluation sequence of repairs, utility repair times, and impeding
of a building and the suitability of the considered retrofit delays for the downtime assessment of a building.
technique. The strength factor determined from the Five discrete functionality states were considered for
capacity curve for the reference un-retrofitted building developing the functionality repair curve to evaluate
was 0.92, indicating non-conformance with the current seismic resilience. Among the considered retrofit
building codes, and hence, structural modifications are alternatives, SJ and RCJ showed better seismic resilience,
570 EARTHQUAKE ENGINEERING AND ENGINEERING VIBRATION Vol. 19

while FRPs and the un-retrofitted building showed poor Resilience Assessment of Bridges under Mainshock and
seismic resilience. Aftershocks Incorporating Uncertainties,” Engineering
Structures, 83: 198‒208.
Acknowledgement Dong Y and Frangopol DM (2016), “Performance-Based
Seismic Assessment of Conventional and Base-Isolated
The first author acknowledge funding provided by Steel Buildings Including Environmental Impact
the Research Grants Council (RGC) of Hong Kong and Resilience,” Earthquake Engineering Structural
under the Hong Kong PhD Fellowship Scheme (HKPFS) Dynamics, 45(5): 739‒756.
of 2018. This study has been supported by the Chinese Dong Y, Frangopol DM and Saydam D (2014),
National Engineering Research Centre (CNERC) for “Pre-Earthquake Multi-Objective Probabilistic
Steel Construction (Hong Kong Branch) at the Hong Retrofit Optimization of Bridge Networks Based on
Kong Polytechnic University (Project No. P0013864; Sustainability,” Journal of Bridge Engineering, 19(6):
Programme Code: BBV9). 04014018.
Dong Y, Frangopol DM and Sabatino S (2016), “A
References Decision Support System for Mission-Based Ship
Routing Considering Multiple Performance Criteria,”
Almufti I and Willford M (2013), REDi™ Rating Reliability Engineering & System Safety, 150: 190‒201.
System: Resilience Based Earthquake Design Initiative Elkady A and Lignos DG (2015), “Effect of Gravity
for the Next Generation of Buildings, Version 1.0, Framing on the Overstrength and Collapse Capacity of
October, Arup. Steel Frame Buildings with Perimeter Special Moment
Anelli A, Santa-Cruz S, Vona M, et al. (2019), “A Frames,” Earthquake Engineering and Structural
Proactive and Resilient Seismic Risk Mitigation Dynamics, 44(8): 1289‒1307.
Strategy for Existing School Buildings,” Structure and Elnashai AS and Di Sarno L (2008), Fundamentals of
Infrastructure Engineering, 15(2): 137‒151. earthquake engineering, Wiley Online Library.
Anwar GA, Dong Y and Zhai C (2020), “Performance- FEMA-547 (2006), Techniques for the Seismic
Based Probabilistic Framework for Seismic Risk, Rehabilitation of Existing Buildings, Building Seismic
Resilience, and Sustainability Assessment of Reinforced Safety Council for the Federal Emergency Management
Concrete Structures,” Advances in Structural Agency.
Engineering, 23(7): 1454‒1472.
FEMA (2012), Seismic Performance Assessment of
Asadi E, Salman AM and Li Y (2019), “Multi- Buildings: Vol. 1–Methodology, Report Number FEMA
Criteria Decision-Making for Seismic Resilience and P-58-1, Applied Technology Council for the Federal
Sustainability Assessment of Diagrid Buildings,” Emergency Management Agency, Washington, D.C.
Engineering Structures, 191: 229‒246.
Gautam D and Chaulagain H (2016), “Structural
ASCE-41-13 (2013), Seismic Evaluation and Retrofit of Performance and Associated Lessons to be Learned from
Existing Buildings, American Society of Civil Engineers, World Earthquakes in Nepal After 25 April 2015 (MW
Reston, Virginia. 7.8) Gorkha Earthquake,” Engineering Failure Analysis,
Beavers JE (2002), “A Review of Seismic Hazard 68: 222‒243.
Description in US Design Codes and Procedures,” Giouvanidis AI and Dong Y (2020), “Seismic Loss
Progress in Structural Engineering and Materials, 4(1): and Resilience Assessment of Single-Column Rocking
46‒63. Bridges,” Bulletin of Earthquake Engineering, DOI:
Billah AM and Alam MS (2014), “Seismic Performance 10.1007/s10518-020-00865-5.
Evaluation of Multi-Column Bridge Bents Retrofitted Guo T, Xu Z, Song L, et al. (2017), “Seismic Resilience
with Different Alternatives Using Incremental Dynamic Upgrade of RC Frame Building Using Self-Centering
Analysis,” Engineering Structures, 62: 105‒117. Concrete Walls with Distributed Friction Devices,”
Bocchini P, Frangopol DM, Ummenhofer T, et al. (2013), Journal of Structural Engineering, 143(12): 04017160.
“Resilience and Sustainability of Civil Infrastructure: Hashemi MJ, Al-Attraqchi AY, Kalfat R, et al. (2019),
Toward a Unified Approach,” Journal of Infrastructure “Linking Seismic Resilience into Sustainability
Systems, 20(2). Assessment of Limited-Ductility RC Buildings,”
Calvi G (2013), “Choices and Criteria for Seismic Engineering Structures, 188: 121‒136.
Strengthening,” Journal of Earthquake Engineering, Jeong S-H and Elnashai AS (2005), “Analytical
17(6): 769‒802. Assessment of an Irregular RC Frame for Full-Scale
Cimellaro GP, Reinhorn AM and Bruneau M (2010), 3D Pseudo-Dynamic Testing part I: Analytical Model
“Framework for Analytical Quantification of Disaster Verification,” Journal of Earthquake Engineering,
Resilience,” Engineering Structures, 32(11): 3639‒3649. 9(01): 95‒128.
Dong Y and Frangopol DM (2015), “Risk and Khanmohammadi S, Farahmand H and Kashani H
No. 3 Ghazanfar Ali Anwar et al.: Seismic resilience of retrofitted RC buildings 571

(2018), “A System Dynamics Approach to the Seismic Systems, 32(1-2): 157‒169.

Resilience Enhancement of Hospitals,” International Rousakis TC (2018), “Inherent Seismic Resilience of
Journal of Disaster Risk Reduction, 31: 220‒233. RC Columns Externally Confined with Nonbonded
Kilanitis I and Sextos A (2019), “Integrated Seismic Composite Ropes,” Composites Part B: Engineering,
Risk and Resilience Assessment of Roadway Networks 135: 142‒148.
in Earthquake Prone Areas,” Bulletin of Earthquake Su L, Wan H-P, Dong Y, et al. (2019), “Seismic
Engineering, 17(1): 181‒210. Fragility Assessment of Large-Scale Pile-Supported
Li Y, Dong Y, Frangopol DM, et al. (2020a), “Long- Wharf Structures Considering Soil-Pile Interaction,”
Term Resilience and Loss Assessment of Highway Engineering Structures, 186: 270‒281.
Bridges Under Multiple Natural Hazards,” Structure Thermou G and Elnashai AS (2006), “Seismic
and Infrastructure Engineering, 1‒16, DOI: Retrofit Schemes for RC Structures and Local‐Global
10.1080/15732479.2019.1699936. Consequences,” Progress in Structural Engineering and
Li Y, Dong Y and Qian J (2020b), “Higher-Order Analysis Materials, 8(1): 1‒15.
of Probabilistic Long-Term Loss under Nonstationary Tirca L, Serban O, Lin L, et al. (2016), “Improving the
Hazards,” Reliability Engineering & System Safety, Seismic Resilience of Existing Braced-Frame Office
107092. Buildings,” Journal of Structural Engineering, 142(8):
Ma C-K, Apandi NM, Sofrie CSY, et al. (2017), “Repair C4015003.
and Rehabilitation of Concrete Structures Using Vamvatsikos D and Allin Cornell C (2006), “Direct
Confinement: A Review,” Construction and Building Estimation of the Seismic Demand and Capacity
Materials, 133: 502‒515. of Oscillators with Multi-Linear Static Pushovers
Mitrani-Reiser J (2007), An Ounce of Prevention: Through IDA,” Earthquake Engineering and Structural
Probabilistic Loss Estimation for Performance- Dynamics, 35(9): 1097‒1117.
Based Earthquake Engineering, California Institute of Vona M, Mastroberti M, Mitidieri L, et al. (2018), “New
Technology. Resilience Model of Communities Based on Numerical
Molina Hutt C, Almufti I, Willford M, et al. (2016), Evaluation and Observed Post Seismic Reconstruction
“Seismic Loss and Downtime Assessment of Existing Process, International Journal of Disaster Risk
Tall Steel-Framed Buildings and Strategies for Increased Reduction, 28: 602‒609.
Resilience,” Journal of Structural Engineering, 142(8): Wang Z, Jin W, Dong Y and Frangopol DM (2018),
C4015005. “Hierarchical Life-Cycle Design of Reinforced Concrete
Mwafy A and Elkholy S (2017), “Performance Structures Incorporating Durability, Economic Efficiency
Assessment and Prioritization of Mitigation Approaches and Green Objectives,” Engineering Structures, 157:
for Pre-Seismic Code Structures,” Advances in Structural 119‒131.
Engineering, 20(6): 917‒939. Whitman RV, Anagnos T, Kircher CA, et al. (1997),
Pekcan G, Itani AM and Linke C (2014), “Enhancing “Development of a National Earthquake Loss Estimation
Seismic Resilience Using Truss Girder Frame Systems Methodology,” Earthquake Spectra, 13(4): 643‒661.
with Supplemental Devices,” Journal of Constructional Zheng Y, Dong Y, Chen B, et al. (2019), “Seismic
Steel Research, 94: 23‒32. Damage Mitigation of Bridges with Self-Adaptive
Qian J and Dong Y (2020), “Multi‐Criteria Decision SMA-Cable-Based Bearings,” Smart Structures and
Making for Seismic Intensity Measure Selection Systems, 24(1): 127‒139.
Considering Uncertainty,” Earthquake Engineering and Zheng Y, Dong Y and Li Y (2018), “Resilience and Life-
Structural Dynamics. Cycle Performance of Smart Bridges with Shape Memory
Rodriguez-Nikl T (2015), “Linking Disaster Resilience Alloy (SMA)-Cable-Based Bearings,” Construction and
and Sustainability,” Civil Engineering and Environmental Building Materials, 158: 389‒400.