Displacement Based Simplified Seismic Loss Assessment of Post 70s RC Buildings - 2020

Journal of Earthquake Engineering
ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/ueqe20
Displacement-Based Simplified Seismic Loss

Assessment of Post-70s RC Buildings
Luca Landi , Daniel Saborio-Romano , David P. Welch & Timothy J. Sullivan
To cite this article: Luca Landi , Daniel Saborio-Romano , David P. Welch & Timothy J. Sullivan
(2020) Displacement-Based Simplified Seismic Loss Assessment of Post-70s RC Buildings,
Journal of Earthquake Engineering, 24:sup1, 114-145, DOI: 10.1080/13632469.2020.1735577
To link to this article: https://doi.org/10.1080/13632469.2020.1735577
Published online: 08 Jun 2020.
Submit your article to this journal
Article views: 25
View related articles
View Crossmark data
Citing articles: 1 View citing articles
Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=ueqe20
JOURNAL OF EARTHQUAKE ENGINEERING
2020, VOL. 24, NO. S1, 114–145
https://doi.org/10.1080/13632469.2020.1735577
Displacement-Based Simplified Seismic Loss Assessment of

Post-70s RC Buildings
a
Luca Landi , Daniel Saborio-Romanob, David P. Welchc, and Timothy J. Sullivan d
a
Department DICAM, University of Bologna, Bologna, Italy; bUME School, IUSS Pavia, Pavia, Italy; cROSE School,
IUSS Pavia, Pavia, Italy; dDepartment of Civil and Natural Resources Engineering, University of Canterbury,
Christchurch, New Zealand
ABSTRACT ARTICLE HISTORY

To help facilitate performance-based earthquake engineering (PBEE) Received 2 August 2019
in Italy, this study examines an existing reinforced concrete frame Accepted 22 February 2020
building damaged during the 2009 L’Aquila earthquake and com- KEYWORDS
pares direct monetary loss estimates computed via a simplified Displacement-based
Displacement-Based Assessment (DDBA) method with those assessment; simplified direct
obtained using the refined PEER-PBEE approach. The considered loss estimation;
case study is representative of a category of buildings that has performance-based
undergone some level of seismic design, typical of post-70s construc- earthquake engineering;
tion in Italy, where previous buildings were largely designed for reinforced concrete frames;
gravity loads only. The research finds that the expected annual loss reparability
predicted using the two approaches are comparable and thus DDBD
could aid loss-assessment and hence reparability evaluations.
Different retrofit options are examined and it is concluded that the
DDBA approach could provide useful indications of the impact that
retrofit can have on expected annual losses.
1. Introduction
The main purpose of seismic design has been traditionally the life safety. However, several
strong earthquakes such as the 1994 Northridge earthquake, showed that the limitation of loss
of life was accompanied by high widespread of damage. This laid the groundwork for
improvement of the seismic design with the aim of a more comprehensive control of the
seismic response, arriving at performance-based seismic design methods (SEAOC 1995). In the
past decade, further improvements of the Performance-Based Earthquake Engineering (PBEE)
procedures have allowed to include the uncertainties in seismic engineering and to define
performance measures in terms of expected losses, which is more useful for decision makers.
Much research in seismic design has also aimed to mitigate inherent issues with current
force-based design procedures (Priestley 1993, 2003). In particular, new design approaches
have been proposed for designing structures in order to achieve a specified deformation
state under a design level of earthquake shaking. These approaches, characterized by the
assumption of structural displacement being the fundamental design parameter, are
known as displacement-based design procedures and can be particularly effective in
combination with the PBEE framework (Priestley 2000; Sullivan, Welch, and Calvi
2014). The approach proposed by Priestley (Priestley 2003; Priestley, Calvi, and
CONTACT Luca Landi l.landi@unibo.it Department DICAM, University of Bologna, Viale Risorgimento 2, 40136
Bologna, Italy
© 2020 Taylor & Francis Group, LLC
JOURNAL OF EARTHQUAKE ENGINEERING 115
Kowalsky 2007) is a Direct Displacement-Based Design (DDBD) procedure based on the

characterization of a substitute structure and on the use of highly damped displacement
spectra. The substitute structure is an equivalent single-degree-of-freedom (SDOF) system
with secant stiffness to the limit state displacement. The determination of the SDOF
system properties depends on the expected plastic mechanism. The effects of hysteretic
response are considered using an equivalent viscous damping approach. The DDBD
procedure has then been studied and applied with reference to different structural
typologies (Priestley, Calvi, and Kowalsky 2007; Sullivan, Priestley, and Calvi 2012). The
same approach can be followed when carrying out the seismic assessment of existing
buildings, leading to Direct Displacement-Based Assessment (DDBA) methods. However,
in the assessment procedure, the definition of the expected plastic mechanism is
a challenging issue strongly related to the characteristics of existing buildings, which
often have not been designed to meet modern design criteria.
The focus of this article is on the displacement-based seismic assessment of buildings
that are typical of post-70s construction in Italy. This definition is broadly taken to mean
any RC frame building that has undergone some level of seismic design. Buildings
constructed from around 2003 onwards are likely to have improved detailing and capacity
design provisions than those from the latter part of the twentieth century, owing to code
changes that took place in 2003 (OPCM 2003) and 2008 (NTC 2008). The code of 2008
has then been updated in 2018 (NTC 2018).
The main purpose of this study is to apply the concepts of PBEE and DDBA to a case
study of an existing RC frame building in L’Aquila, Italy. This building, designed con-
sidering seismic actions but not according to more recent code provisions, may be seen as
representative of many Italian RC buildings designed prior to the 2003/2008 code changes.
Whilst variations in structural characteristics would differ between other post-70s build-
ings, the behaviour of the selected case-study building is expected to be representative of
post-70s Italian buildings in general. Moreover, it was damaged during the 2009 L’Aquila
earthquake and subsequently repaired and retrofitted. The choice of repairing the
damaged existing structure rather than replacing it with a new building was mainly linked
to the estimate of the repair and retrofit cost (retrofit for a given percentage of the design
action for new buildings), and was based on the consideration that this cost was estimated
to be lower than the cost of demolition and reconstruction. This highlights with more
evidence the importance of the estimation of the expected losses and the necessity of
simplified tools for this purpose. The specific objectives of the present research were in
particular: (a) the nonlinear modelling and analysis of the building in order to reproduce
the observed damage and the estimate through the DDBA procedure of the response
obtained in the nonlinear analysis; (b) the comparison between the refined PBEE approach
and the simplified displacement-based methods to estimate the expected annual loss; (c)
the study of the effects of different retrofit solutions in terms of expected annual loss.
2. Displacement-Based Loss Assessment (DBLA) Framework and Specific

Considerations for RC Buildings
PBEE has been defined as a framework under which a desired structural system perfor-
mance is sought for a specified seismic intensity level. Nowadays, the most advanced PBEE
seismic assessment procedure appears to be the methodology developed by the Pacific
116 L. LANDI ET AL.
Earthquake Engineering Research (PEER) Center (Porter 2003) and outlined with the
FEMA P-58 guidelines (FEMA 2012), which leads to system performance measures, such
as economic losses, downtime, and casualties, these being parameters of interest in
decision makers.
The PEER-PBEE procedure consists of four successive stages of analysis: hazard
analysis (site definition), structural analysis, damage analysis, and loss (decision) analysis.
In all the stages, the related uncertainties are explicitly considered in a probabilistic
manner. Economic losses can be expressed in terms of a single intensity level, earthquake
scenario (e.g. a given magnitude and distance case) or expressed in annualized terms (e.g.
Expected Annual Loss) (FEMA 2012); the latter of which is very attractive for decision
makers since the entire hazard curve is considered (within reasonable bounds) and
annualized values can be converted to present values to allow for cash flow analysis (i.e.
cost-benefit analysis).
The PEER-PBEE methodology usually requires the use of time-consuming and complex
analysis procedures, this being the main reason for the development of simplified meth-
ods. The simplified loss estimation using Direct Displacement-Based Assessment (DDBA)
concepts (Welch, Sullivan, and Calvi 2014) is intended to achieve the aims of refined
PEER-PBEE methods (i.e. assist decision making), estimating the structural response
without performing a series of Non-Linear Time History (NLTH) analyses, while provid-
ing an alternative means of defining capacity curve models for the simplified method of
FEMA P-58 (FEMA 2012). For more details about the Displacement-Based loss assess-
ment methodology see (O’Reilly et al. 2019).
The basis of the DDBA loss model is the approximation of the mean annual frequency
of exceedance (MAFE) versus Mean Damage Factor (MDF) (i.e. the expected economic
loss) as a multilinear curve defined by specific limit states. Basically, this is a tri-linear
curve, defined by four key limit states: zero loss, operational, damage control, and near
collapse. The zero loss limit state represents the beginning of the loss accumulation with
a high MAFE, attributing an MDF value of 0.0. The operational and damage control limit
states characterize the transition between significant MAFE and low loss to low MAFE and
higher loss, where these points require an estimate of the corresponding MDF. The near
collapse limit state has a low MAFE with a substantial loss accumulation, assuming an
MDF value of 1.0 that represents the total replacement of the building; note that rare high-
intensity levels, despite significant losses, do not contribute significantly to EAL.
Specific considerations for RC buildings and case study analyses regard the following
issues. Firstly, for an existing RC building, the application of the DDBA requires a proper
prediction of the plastic mechanism and of the displacement profile, which is strongly
affected by the case study characteristics. In particular, several RC buildings present
vertical or in-plan irregularities as well as non-conforming detailing with respect to
modern codes. These considerations were the starting point of a previous research effort
aimed at proposing a new DDBA procedure (Sullivan et al. 2018), which was then applied
in the present research to the considered case study. Secondly, different aspects may affect
the estimate of damage and expected annual loss of an existing RC building, as the type of
failure expected for the structural members as well as a proper modelling of non-structural
elements such as masonry infill walls. In particular, specific models for the displacement-
based characterization of the masonry infill walls, as well as consistent fragility functions
and cost functions are required for a proper assessment of the existing RC building
assumed in the present case study. Damage fragility and cost functions for masonry infills
and windows proposed in previous research (Sassun 2014; Sassun et al. 2016) were used in
lieu of absent existing information applicable to the current case study building.
3. The Case Study: An Existing RC Frame Building in L’Aquila

3.1. Building Examined
The case study is an existing six-storey building, consisting of multiple RC frames with
unreinforced masonry partitions and infill walls located in L’Aquila, in the central region
of Italy. The building is regular in plan with seven bays in the longitudinal direction (X
direction) and three principal bays in the transversal direction (Y direction), with a total
area per storey of 340.5 m2, as shown in Fig. 1.
The first storey consists of a basement for wine cellars and storage rooms, separated
from the perimeter retaining walls by a gap of 80 cm. The second storey consists of offices,
shops and garages and the last four storeys are residential. The columns are continuous
along the height while the masonry infill walls and partitions vary along the height due to
differences in architectural configurations. The beam and column sections are gradually
reduced from the lower to the upper storeys according to conventional construction
criteria. However, the storey heights increase from the lower to the upper storeys, causing
an irregularity in the storey stiffness along the height, with stiffer storeys in the lower
levels and more flexible storeys in the upper levels. In detail, the following percentage
reductions in storey stiffness can be estimated for each storey (considering the storey
below and above and presuming shear-type behaviour): at the first (ground) level 51% in
the X direction and 66% in the Y direction, at the second level 17% in both directions, at
the third level 55% in the X direction and 30% in the Y direction, at the fourth level 22%
in the X direction and 50% in the Y direction, and at the fifth level 11% in both directions.
These percentage reductions can be compared with the limit of 30% provided by the
Italian code for buildings regular in elevation (NTC 2018).
Figure 1. Plan view of the building (left) and transverse section (right) in the Y direction, respectively,
taken from Saborío-Romano (2016).
118 L. LANDI ET AL.
In the Y direction, all the frames have deep beams, while in the X direction the outer
frames have strong beams and the interior frames have flat beams. The column positions
in plan show that the larger dimension is orientated in the Y direction, so this, in
combination with larger beam sections in the Y direction, causes a stiffer structure in
this direction compared to the X direction. The structure was designed for some seismic
action but without more recent seismic design requirements (such as reliable reinforce-
ment details or capacity design principles). Considering the construction period, the
design was performed according to the admissible stress design approach. Details about
the original calculations were not available, so information about the design assumptions
and modelling criteria are not known. No structural walls are present aside from masonry
infill walls. They represent a traditional unreinforced masonry configuration constructed
in full contact with the surrounding load-bearing structure, consisting of a clay brick
configuration widely used in Italian building practice for the era of construction: two brick
leaves divided by an intermediate cavity.
As explained previously, the building was damaged during the 2009 L’Aquila earth-
quake. The observed damage was widespread in the non-structural elements such as infill
walls, partitions and floors, as well as some RC structural elements. The damage, in
particular, concentrated in the storeys above the basement, being more evident in the
upper four storeys. At these storeys several partitions presented full-width X-shaped
cracks, some partitions collapsed, the external infills presented cracks along the perimeter
and between the windows and plaster detachment was observed, also from the ceilings. At
the third and fourth levels, two external beams presented significant damage due to
flexure, while in the last three storeys a number of columns presented slight to moderate
damage. The level of damage sustained by the building can be associated to a performance
level, according to Italian code, between the damage control and life safety levels. The
building, subsequently, was repaired and retrofitted. The decision to repair the damaged
existing structure rather than to replace it with a new building was mainly linked to the
estimate of the repair and retrofit cost, which was estimated to be lower than the cost for
demolition and reconstruction. This is an issue of particular importance, since the
decision of repairing and upgrading rather than replacing is challenging and requires
particular attention. This aspect should also be the topic of future research, in order to
provide clear criteria regarding reparability and residual capacity. The repair and retrofit
intervention was based mainly on the introduction of new RC walls around the central
elevator and on the reinforcement of several single structural elements and it was designed
in order to achieve 60% of the design seismic action for new buildings. The final costs
were about 20% of the replacement cost for the repair intervention, 30% for the retrofit
interventions and 20% for the adjustment intervention on the facilities.
3.2. Location, Site Hazard and Ground Motions

Since the building is located in L’Aquila, the hazard of this city was considered and the
data were taken from the Italian website INGV (Istituto Nazionale di Geofisica
e Vulcanologia). For each of the nine return periods taken into account by the website
just mentioned (30, 50, 72, 101, 140, 201, 475, 975, 2475 years), 10 pairs of horizontal
ground motion components were selected and scaled. In particular, the ground motions
were selected from the catalogue of SHARE (2013) and scaled following the procedure of
Ay, Fox, and Sullivan (2017) to be compatible with the conditional spectra for the site
(conditioned at a period T = 1.5 s), using the maximum spectral acceleration of the two
orthogonal components of the record (denoted as SamaxDir) (Haselton et al. 2012). The
considered spectrum for the record selection is for soil type A. A total of 90 records has
been selected and used to run non-linear time history analyses. For more details about the
site hazard and ground motions selection see (O’Reilly et al. 2019).
3.3. Structural Modelling and Analysis Results

Two different three-dimensional structural models were created using the software
Ruaumoko3D (Carr 2013) namely: bare frame (BF) and infilled frame (IF) models. The
nonlinear model is characterized by cracked section properties assigned to all the
elements, which have been modelled as Giberson one component members to create
a lumped plasticity model, and allowing the plastic hinges to form at both the ends of the
beams and at the top and bottom of the columns. The cracked section properties have
been defined according to the results from the moment-curvature analysis performed
using Cumbia (Montejo and Kowalsky 2007). Cumbia was applied using the Berry and
Eberhard (2005) buckling model and the plastic hinge length evaluated as described in
Priestley, Calvi, and Kowalsky (2007). The calculations have been performed assuming
material properties according to mean values from diagnostic investigations on the
structure: concrete compressive strength of 20 MPa and steel yield stress of 343 MPa.
The moment capacities of all elements are listed in Saborío-Romano (2016). The plastic
hinges behaviour under cyclic loading was represented by adopting the modified “fat”
and “thin” Takeda hysteresis rule, implemented in Ruaumoko, for the beams and
columns, respectively. These rules account for stiffness degradation but strength dete-
rioration was neglected assuming it had a limited influence for most considered earth-
quake records.
The contribution of the masonry infill walls (strength and stiffness) was considered
only in the infilled frame model for walls that do not have any opening and were
surrounded in all contact surfaces by an RC frame (see Fig. 1 left). This assumption was
introduced considering that most of the panels with openings are characterized by the
presence of two openings and by an area of the openings similar or larger than 50% of the
area of the panel. In such cases, the effects of the infill panel are limited (Al-Chaar 2002)
and in general may be neglected. The details of these walls are two 12 cm thick solid clay
brick leaves each separated by a centre cavity of 14 cm. The numerical model for these
elements is two equivalent diagonal struts connected to the RC frames at the centrelines
intersection (see Fig. 2 right, Fig. 3), using a spring with the cyclic response of the
Crisafulli hysteretic model (Crisafulli and Carr 2007).
The peak axial stress of these elements was derived from the lateral resistance force
model of Decanini et al. (2004) and the axial stiffness was obtained from the displacement-
based model of Sassun et al. (2016), which associates limiting inter-storey drift values with
corresponding damage states (DS) on the masonry walls. The material properties of the
masonry infill walls and the elements properties are described in Saborío-Romano (2016).
Notably, the adverse effects of frame–infill interaction (i.e. induced shear failures) are
neglected in the structural model. Damage associated with the shear failure of columns
due to the effects of the infills, in fact, was not observed in the building.
120 L. LANDI ET AL.
Figure 2. Ruaumoko model of the bare frame (left) and of the frame with infill walls (right).
Figure 3. Equivalent strut geometry from Hak et al. (2012) (left) and backbone stress–strain curve
(right).
The eigenvalue analysis provided the fundamental periods of the two models, bare (BF)
and infilled frames (IF), in the two principal directions, X and Y: T1,X = 1.83 s and T1,Y
= 1.43 s for the BF and T1,X = 1.41s, and T1,Y = 1.21 s for the IF. This is the reason for
considering 1.5 s as a period of interest for the building in the definition of the conditional
spectra, being the closest available value between the initial period of the two structural
models.
Two different types of analyses were used to determine the structural response under
seismic actions: nonlinear pushover analysis and nonlinear time-history analysis (NLTH).
In the pushover analysis, the inelastic mechanism formation sequence and force–displace-
ment curve were determined under a sustained monotonic inverse triangular lateral force
vector. In order to know when the elements reached specific damage states, the limit states
according to the Italian code (NTC 2018) were considered. These limit states were defined
in terms of peak inter-storey drift at each storey, which are Operational (SLO), Damage
Control (SLD), Life Safety (SLV), and Collapse Prevention (SLC) limit states. The drift
limits for the SLO and SLD limit states were 0.33% and 0.5%, respectively, corresponding
to a damage limitation for the non-structural elements. The chord rotation capacities for
the SLV and SLC limit states were calculated as a function of the properties of the
structural elements and were 2.0% and 2.6%, respectively, for the critical storey in the
X-direction and 2.0% and 2.7% for the Y-direction. The obtained capacity curves are
shown in Fig. 4 in terms of base shear normalized to the estimated total weight and roof
displacement to the total height. At the SLC limit states the shear collapse of the structural
10 18
15
8
B as e She ar/Total We ight [%]
B as e She ar/To tal We ight [%]

12
6
Bare Frame Bare Frame
Infilled Frame 9 Infilled Frame
4 Out of SLC Out of SLC
SLO 6 SLO
SLD SLD
2
SLV 3 SLV
SLC SLC
0 0
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2 0 0.25 0.5 0.75 1 1.25 1.5 1.75 2
Top Displacement/Total Height [%] Top Displacement/Total Height [%]
(a) (b)
Figure 4. Capacity curves in X (a) and Y (b) directions including the identifications of the different limit
states.
elements should be considered also, especially in presence of a structure not designed to

more recent seismic criteria, with limited shear reinforcement in the columns. However, it
was verified that the shear capacity of the columns which developed plastic hinges was
sufficient to allow the development of the flexural plastic hinges.
Comparing the capacity curves on the left and on the right of Fig. 4, the structure is
stiffer and stronger, in terms of base shear, in the Y direction for both models with respect
to the X direction, which is expected due to the structural geometry and configuration.
The displacement capacity of the structure, based on the top (roof) displacement, seems to
be very similar for both models and directions, which can be estimated as 0.3 m. The
curves of the infilled frame models have a higher initial stiffness compared to the bare
frame models, due to the contribution of the masonry infill walls. However, for higher
displacements, the capacity of the masonry infill walls decreases due to the strength
degradation of the infills, so the lateral force capacity tends to be similar to that of the
bare frame.
Figure 5 displays the plastic hinge formation in the structure when the Collapse
Prevention limit state (SLC) was reached in the pushover analysis in the X direction.
Displacement profiles at the centre of mass are provided in Fig. 6 for each of the four limit
states considered from NTC (2018). At the imminent collapse condition in Fig. 5, the
Figure 5. Locations of plastic hinge formation in the X direction for the bare (a) and infilled (b) frame
models at SLC.
122 L. LANDI ET AL.
6 6
5 5
4 4
Sto rey
Sto rey
3 SLO 3 SLO
SLD SLD
2 2
SLV SLV
1 1
SLC SLC
0 0
0 0.1 0.2 0.3 0 0.1 0.2 0.3
Displacement [m] Displacement [m]
(a) (b)
Figure 6. Displacement profiles at the centre of mass in X (a) and Y (b) directions at the different limit
states for the bare frame (continuous line) and infilled frame (dash line).
development of the plastic hinges in the structure occurs principally between the third and
the fifth storey in both directions in a mixed inelastic mechanism. In these storeys, all
beams present plastic hinges in both ends, while the bottom part of some outer columns of
the third storey and the top part of some outer columns of the fifth storey have hinging.
Additionally, in the first two and the last storeys, the beams develop very few plastic
hinges and, at the base level, there are not plastic hinges in the columns, showing that the
inelastic mechanism of the structure does not follow the capacity design principles and
a desired plastic deformation (i.e., beam-sway mechanism). This behaviour of the struc-
ture can be explained due to the higher stiffness of the first two storeys, described in
Section 3.1, being rigid enough to produce a cantilever wall effect in these columns,
causing the third storey columns to hinge. This collapse mechanism can be also observed
in Fig. 6, where the displacements at the SLC and SLV limit state are proportionally large
between the third and the fifth storey compared to the other storeys. These displacements
are much lower in the first two storeys, showing the effect of increased stiffness of the base
columns previously described.
Moreover, in Fig. 6, the difference between the displacement profiles of the bare and
infilled frame models is small because the limit states SLO and SLD are defined in the
NTC (2018) to minimize the damage on non-structural components, such as the masonry
infill walls, even for the bare frame model. For the limit states SLV and SLC, the threshold
is defined for the RC columns, for which the masonry infill walls in the infilled frame
model would have extensive damage, so the infilled structure tends to behave as the bare
frame model. The obtained displacement profiles are in agreement with damage trends
observed after the L’Aquila earthquake, with less damage in the first few stories above the
basement and more apparent damage in the last four storeys.
Following the pushover analyses, the structural models were analysed with the 90
selected ground motion records in order to perform the nonlinear time-history analyses.
The purpose was to obtain the Engineering Demand Parameters (EDPs) that are required
for the refined loss estimation. The considered EDPs are the maximum inter-storey drift,
the peak floor acceleration (response quantity relevant to acceleration-sensitive non-
structural elements) and the residual inter-storey drift (response quantity relevant to the
repairability after a major earthquake). A single value of these EDPs was obtained for each
storey and each direction. The principal ground motion component of each record was
oriented in the X direction and the secondary component in the Y direction.
Notwithstanding the hazard definition for the site, this approach is likely to provide
slightly conservative estimates of the expected loss for the considered typology of building,
considering that the X direction was the weaker direction. Figure 7 represents the inter-
storey drifts as a function of seismic intensity in the X direction at the fourth storey
(between level three and four), where for several records the maximum values along the
height were obtained. Figure 7 shows also the values of median and dispersion.
The median values of inter-storey drift show a nearly linear increasing trend with
seismic intensity, with larger values for the bare frame than for the infilled frame. The
median values for the bare frame are larger than those of the infilled frame for all intensity
levels, with differences of about 28% on average over all intensities and storeys. The
dispersion shows instead a variable trend with seismic intensity, with a tendency to
decrease for the infilled frame and to increase for the bare frame. At the fourth storey,
for a return period of 2475 years, the median and dispersion for the bare frame are,
respectively, 2.01% and 0.31 (Fig. 7(a,c)), for the infilled frame 1.71% and 0.21 (Fig. 7
(b,d)).
Also, the median values of peak floor acceleration are characterized by a nearly linear
increasing trend with the seismic intensity level. For the peak floor accelerations,
however, the median values for the infilled frame are for all intensity levels slightly
larger than those of the bare frame, with differences of about 18% on average over all
intensities and storeys. The dispersion of the peak floor acceleration has a variable trend
with seismic intensity, with a tendency to decrease for both the infilled and the bare
frames. At the fourth storey, for a return period of 2475 years, the median and disper-
sion for the bare frame are, respectively, 0.27 g and 0.47, for the infilled frame 0.29 g
and 0.31.
4.00% 4.00%
3.50% 3.50%
3.00% 30yrs 3.00% 30yrs

Inter-storey drift [%]
Inter-storey drift [%]
50yrs 50yrs
2.50% 72yrs 2.50% 72yrs
101yrs 101yrs
2.00% 2.00%
140yrs 140yrs
1.50% 201yrs 1.50% 201yrs
475yrs 475yrs
1.00% 975yrs 1.00% 975yrs
2475yrs 2475yrs
0.50% 0.50%
Median Median
0.00% 0.00%
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
SamaxDir(1.5 s) [g] SamaxDir(1.5 s) [g]
(a) (b)
0.35 0.35
Dispersion of inter-storey drift
Dispersion of inter-storey drift
0.3 0.30
0.25 0.25
0.2 0.20
0.15 0.15
0.1 0.10
0.05 0.05
0 0.00
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
(c) SamaxDir(1.5 s) [g] (d) SamaxDir(1.5 s) [g]
Figure 7. Inter-storey drift in the X direction at the fourth storey for the bare frame (a) and the infilled
frame (b), dispersion of inter-storey drift at the fourth storey for the bare (c) and infilled frame (d).
124 L. LANDI ET AL.
3.4. Direct Displacement-Based Assessment Procedure (DDBA)

The application of the DDBA procedure for the case study building is developed in this
section as an alternative to the computer-based structural analysis described in the previous
section. In general, the displacement-based design or assessment procedure is based on the
assumption of a displacement profile, which should correspond to the expected inelastic
mechanism. For example, the displacement shape proposed by Priestley, Calvi, and Kowalsky
(2007) is usually considered as an inelastic first mode shape for a beam-sway mechanism that
can be achieved through the detailing of a newly designed frame structure. Nevertheless, this
displacement shape assumes the plastic hinge formation at column base level and the relative
stiffness between adjacent floors very similar, which is not necessarily accurate for existing
buildings, as is the situation for the case study. For this reason, Sullivan et al. (2018) proposed
a new displacement shape for RC frames which is able to account for the relative stiffness of
the storeys, the variation of yield drifts between adjacent floor levels and the yield drift of the
columns at the base level. The DDBA was then applied according to this procedure, which is
here very briefly recalled. For more details see the work of Sullivan et al. (2018). The results
obtained with this procedure were compared with the non-linear pushover analysis and used
to obtain the EDPs for the DDBA simplified loss estimation procedure proposed by Welch,
Sullivan, and Calvi (2014).
Firstly, the likely mechanism and lateral resistance of each storey are assessed by
comparing, at each joint, the flexural strengths of beams and columns. The storey shear
resistance VR,i is obtained by summing up the shear expected in the columns with the
predicted mechanism, through the moments at the top and bottom part of each column in
the storey i that can be developed after reaching joint equilibrium:
P P
Mcol;b;i þ Mcol;a;i1
VR;i ¼ (1)
hs;i
where hs,i is the storey height and the terms ΣMcol,b,i and ΣMcol,a,i-1 are the sum of the
column end moments immediately below the joint centrelines at level i and above the
joints at level i-1 respectively. The moments at joint equilibrium are taken as the mini-
mum between the sum of the column moments and beam moments extrapolated to the
joint centroid as well as the moment resistance of the joint itself.
The yield drift at each level is calculated according to the expected mechanism. For
beam-sway flexural mechanism Equation (2) can be used. For a column-sway flexural
mechanism the yield drift can be obtained with Equation (3) (Glaister and Pinho 2003)
while for columns yielding in flexure at the base level it can be derived with Equation (4)
(Sullivan et al. 2018):
εy Lb;i
θy;i ¼ 0:5 (2)
hb;i
hs;i
θy;i ¼ 0:43εy (3)
hc
hcf
θy;0 ¼ 0:7εy (4)
hc
where εy is the yield strain of the longitudinal reinforcement in the beams, Lb,i is the length
of the beams between column centres, hb,i is the section depth of the beams, hc is the
section height of columns and hcf is the contraflexure height of the base columns. Since the
yield drifts for beams or base columns are mainly referred to levels rather than storeys, the
determination of the storey yield drift is performed by an average in which the yield drifts
at each adjacent level can be weighted according to strain-energy proportions:
P P
Mj;i θy;i þ Mj;i1 θy;i1
θy;sys;i ¼ P P (5)
Mj;i þ Mj;i1
where θy,sys,i is the storey drift required to yield storey i, Mj,i and Mj,i-1 are the total flexural
resistances (for the governing mechanism) provided at joint centres and θy,i and θy,i-1 are
the drifts at yield at levels i and i-1 respectively.
In this procedure, the determination of the displacement profile is based on the
knowledge of the storey stiffness. By obtaining the shear resistance for each storey,
calculated with Equation (1) and the yield drift, determined with Equation (5), the storey
stiffness can be estimated as follows:
VR;i
ky;i ¼ (6)
θy;sys;i hs;i
The displacement profile is then computed with an iterative procedure where a distribution
of lateral force is applied to the structure, determining, at each storey, a storey shear demand
and a consequent storey drift demand, from which it is possible to derive the displacement
profile. The procedure is iterative since the lateral force profile, determined according to
Priestley et al. (2007), provides forces proportional, at each storey, to the storey displace-
ment. Moreover, since an elastic behaviour is assumed up to the first storey yield, the
specific lateral forces associated with this condition are determined.
Regarding the post-yield response, different assumptions are made according to the
type of mechanism. For a beam-sway mechanism, it is possible to assume a fixed displaced
shape corresponding to the one derived at the first storey yield and subsequently to
increase the values of displacement until all storeys have yielded. For a column-sway
mechanism it is expected that post-yield deformations tend to concentrate at the soft-
storey level, so it is assumed to increase the drift only at that level. Since mixed mechanism
may be expected, this choice could be facilitated by calculating a sway potential index Si
(Priestley, Calvi, and Kowalsky 2007) for each level according to Equation (7):
sum of beam strengths at level i
Si ¼ (7)
sum of column strengths at level i
When Si is greater than 1.0 a column-sway mechanism is expected, while if Si is lower than
0.85 a beam-sway mechanism is assumed. When Si is less than 1.0 a beam-sway mechan-
ism could be expected, but Priestley, Calvi, and Kowalsky (2007) point out that if Si is
greater than 0.85 then it would be prudent to assume a column-sway mechanism. The
resulting sway potential indices for the case study (Saborío-Romano 2016) were lower
than 0.85 at all storeys in both directions, except in the last storey where the sway potential
indices were higher than 1.0. Therefore, according to the assumption for a beam-sway
mechanism, the displaced shape derived at the first storey yield was assumed for the post-
yield response.
126 L. LANDI ET AL.
Once the displacement profile is fixed, the storey displacement values are increased
until the full mechanism is developed. For given storey displacement values it is possible
to calculate the drift demand θi at each storey i as θi = (Δi – Δi-1)/hs,i, where Δi are the floor
displacements, and the displacement ductility µi as µi = θi/θy,sys,i. If the storey displacement
ductility is lower than 1.00, the storey shear Vi can be calculated using Vi = µiVR,i.
Otherwise, when the ductility demand becomes equal or larger than 1.00, the storey
shear remains set to the storey shear resistance VR,i. By knowing the storey shear Vi, it
is possible to calculate the base overturning moment MO/T = ∑ Vi hs,i and the base shear Vb
= MO/T/He where He is the effective height of the equivalent SDOF system (Priestley,
Calvi, and Kowalsky 2007).
The procedure was applied also for the case in which the masonry infills are included.
In this case, the displaced shape obtained for the bare frame is used also for the infilled
frame. For this case study, in fact, it was verified that accounting for the masonry infills in
the derivation of the displacement profile produced results very similar to those obtained
considering for the infilled frame the same displacement profile derived for the bare frame
(Saborío-Romano 2016). So the masonry infills were included only in the calculation of
the storey shear resistances and OTM resistance of the structure.
Figure 8 illustrates the comparison between the first mode shape obtained with the BF
model, the traditional displacement profile of Priestley, Calvi, and Kowalsky (2007) and
the elastic-displaced shape estimated with the new approach proposed by Sullivan et al.
(2018): it is evident that the new procedure produced better results for the frame system in
question. Figures 9 and 10 show a comparison of the force-displacement response
obtained for structural models BF and IF with the DDBA and with the computer-based
structural analysis. In general, it is possible to observe a good agreement between the
capacity curves estimated with the proposed DDBA and those provided by the pushover
analysis performed with Ruaumoko software.
For the BF model in the X direction, the estimated capacity curve tends to have a slightly
lower stiffness in the elastic branch. The transition between elastic-inelastic behaviour is nicely
captured with the new displacement profile used. In the Y direction for the BF model, the
initial stiffness and the elastic–inelastic transition shape are similar to the structural model
Figure 8. Comparison of the new displaced shape proposed by (Saborío-Romano 2016; Sullivan et al.
2018), the first mode shape and the traditional displacement profile of (Priestley, Calvi, and Kowalsky
2007), for the bare frame model in (a) X and (b) Y directions.
3500 6000
3000
5000
2500
Base Shear [kN]
Base Shear [kN]

4000
2000
3000
1500
2000
1000
Pushover Pushover
500 DDBA (New DS) 1000 DDBA (New DS)
Out of SLC Out of SLC
0 0
0 0.1 0.2 0.3 0.4 0 0.1 0.2 0.3 0.4
Top Displacement [m] Top Displacement [m]
(a) (b)
Figure 9. Comparison between the capacity curve obtained with the new displaced shape proposed by
(Sullivan et al. 2018) and the curve obtained with the computer-based structural analysis, for the BF
model, X (a) and Y (b) direction.
Figure 10. Comparison between the capacity curve obtained with the new displaced shape proposed
by (Sullivan et al. 2018) and the curve obtained with the computer-based structural analysis, for the IF
model, X (a) and Y (b) direction.
curve, however, the elastic branch is longer. On the other hand, for the IF model (Fig. 10) the
capacity curves obtained show good results, capturing the overall behaviour of the structure,
with a good estimation of the elastic branch and a slight overestimation of the shear resistance
at ultimate capacity. The transition between the elastic-inelastic behaviour in the X direction is
slightly underestimated, while for the Y direction is slightly overestimated. In all cases, the
overestimation of the base shear is lower than 14.0%.
The results obtained with the DDBA applied according to the procedure proposed by
Sullivan et al. (2018) have then been used to perform the displacement-based loss
estimation. In order to perform this loss estimation, it was necessary to determine
the seismic demand for the considered nine levels of seismic intensity and the corre-
sponding groups of earthquake records. To this purpose, the capacity curves obtained
with the DDBA procedure were converted into spectral acceleration-displacement (i.e.
ADRS format) curves using the typical relationships for determining the properties of
the equivalent SDOF system in the displacement-based procedure (Priestley, Calvi, and
128 L. LANDI ET AL.
Kowalsky 2007). The capacity curves in the ADRS format were superimposed to the
nine average spectra corresponding to the arithmetic mean of the 10 maximum
component spectra in each of the groups of earthquake records relative to the nine
levels of seismic intensity. For each average spectrum, the inelastic spectrum and its
intersection with the capacity spectrum were determined according to the procedure
proposed by Sullivan, Welch, and Calvi (2014). According to this method, a point of
the inelastic spectrum associated to a value of effective period Te larger than the initial
period Ti of the structure can be determined by calculating the ductility demand μ
corresponding to that value of effective period:
2
Te
Ti ð1 rÞ
μ¼ 2 (8)
1 r TTei
where r is the post-yield stiffness ratio (typically r = 0.05 for RC structures) that
characterizes a bi-linear pushover curve. For this ductility demand, it is possible to
determine the equivalent hysteretic damping with literature formulations, and the
corresponding damping reduction factor to be applied to the spectral acceleration
and displacement values at the period Te of the elastic spectrum for 5% damping. For
the bare frame model Equation (9) is used to calculate the equivalent viscous damping:

μ1
eq;RCframe ¼ 0:05 þ 0:565 forμ 1 (9)
μπ
For the IF model, the equivalent viscous damping ξeq,syst was obtained with a weighted
average based on the energy dissipated by the different structural elements at each storey
ξeq,i, as shown in Equation (10), using the storey displacement ductility demand µi = θi/θy,i.
For the RC frames without infills, Equation (9) is used, for the RC frames with infills, two
approaches were used: the equivalent damping proposed in (Landi, Tardini, and Diotallevi
2016), shown in Equation (11) (approach EVD1) and a value of equivalent damping based
on experimental information (Morandi, Hak, and Magenes 2014; Ozkaynak et al. 2014)
and equal to 10% (approach EVD2).
P
Vi θi eq ;i
eq;syst ¼ P (10)
Vi θi

μ 0:07
eq;infillframe ¼ 0:05 þ 0:83 forμ 0:07 (11)
μπ
As an example, the demand and capacity spectrum in the ADRS format for the X direction
and the nine obtained values of displacement demand are illustrated in Fig. 11. From the
Figure, it can be observed that the combination of the seismic hazard at different intensity
levels and the structure capacity in both directions show that the inelastic behaviour of the
structure is expected to occur mainly for the high-intensity levels. For this reason, the
sources of seismic loss in the lower and mid-intensity levels are expected to be from
damage to the non-structural components of the building.
Figure 11. Inelastic spectral ordinates for the equivalent SDOF system for the bare frame in X direction.
4. Loss Estimates for the Case Study Existing Building

The loss analysis was executed using the Performance Assessment Calculation Tool
(PACT) software, presented in the FEMA P-58 guidelines (FEMA 2012) and it was
based only on the economic loss due to damage to the building components, neglecting
casualties, or downtime losses. The total replacement cost of the building was assumed as
3.064.500€, considering a replacement cost per square meter of 1.500€, while the core and
shell replacement cost was considered as 1.500.000€.
4.1. Damageable Inventory and Costing Information

The likely damage that could occur in a component of the case study building is evaluated
through fragility functions, which are typically cumulative lognormal distributions defined
by a median response (EDPmed) and dispersion (β), associated with specific damage states.
For the RC frame elements and elevator components of the building, the fragility func-
tions recommended in FEMA P-58 (2012) and described in Table 1 were selected from the
PACT library, while for the glazing windows and masonry components (infill walls and
partitions) the fragility curves and repair costs of (Sassun 2014) and (Sassun et al. 2016)
were selected, respectively, since they are consistent with the type of elements and
construction practice in Italy. Among the non-structural components considered, the
masonry components and the windows are sensitive to inter-storey drift, the elevator,
instead, is sensitive to accelerations. Other common acceleration sensitive non-structural
components were not included since they were either not present, such as suspended
ceilings, or were deemed rugged for the case study typology. Additionally, a building
residual inter-storey drift fragility curve recommended in FEMA P-58 (2012) and pro-
vided in PACT was selected to determine if the repair of the building is practical with the
residual deformation after the seismic actions. Table 1 presents the damageable assemblies
with the corresponding fragility functions, unit quantities, and repair costs considered for
each damage state and component type.
130
Table 1. Damageable components and corresponding fragility functions, unit quantities, and repair cost.
Quantities X:# in X dir., Y:# in Y dir.
ND:# in Non-Directional
L. LANDI ET AL.
Damage First Second Third Fourth Fifth Sixth

Description state EDP EDPmed β Unit St. St. St. St. St. St. Repair Cost [€ per unit]
Windows 1.0 × 1.0 m DS1 IDR 0.0224 0.30 ea. - X:5 X:2 X:2 X:2 - 190
DS2 IDR 0.0244 0.30 520
Windows 1.0 × 2.5 m DS1 IDR 0.0153 0.30 ea. - - X:3 X:3 X:3 X:3 265
DS2 IDR 0.0184 0.30 Y:1 Y:1 Y:1 Y:1 1151
Windows 1.5 × 1.5 m DS1 IDR 0.0136 0.30 ea. - - X:14 X:13 X:13 X:9 250
DS2 IDR 0.0148 0.30 Y:3 Y:3 1045
Windows 1.0 × 0.5 m DS1 IDR 0.0279 0.30 ea. - - Y:5 X:3 X:3 X:1 164
DS2 IDR 0.0301 0.30 Y:3 Y:3 Y:5 309
Masonry partitions DS1 IDR 0.0014 0.36 m2 X:66 X:79 X:83 X:87 X:88 X:95 62.9
DS2 IDR 0.0033 0.48 Y:90 Y:107 Y:114 Y:119 Y:120 Y:130 142.4
DS3 IDR 0.0096 0.21 296.5
DS4 IDR 0.02 0.28 476.3
Masonry infill walls DS1 IDR 0.0014 0.36 m2 X:118 X:122 X:130 X:136 X:137 X:149 62.9
DS2 IDR 0.0033 0.48 Y:60 Y:63 Y:67 Y:69 Y:70 Y:76 142.4
DS3 IDR 0.0096 0.21 296.5
DS4 IDR 0.02 0.28 476.3
RC Non-conforming MF, beam one side DS1 IDR 0.015 0.40 ea. X:10 X:10 X:10 X:10 X:10 X:10 Cost functions defined in PACT [15]
(B1014.111a) DS2 IDR 0.02 0.40 Y:16 Y:16 Y:16 Y:16 Y:16 Y:16 library
DS3 IDR 0.025 0.40
RC Non-conforming MF, beam both side DS1 IDR 0.015 0.40 ea. X:22 X:22 X:22 X:22 X:22 X:22
(B1014.111b) DS2 IDR 0.02 0.40 Y:16 Y:16 Y:16 Y:16 Y:16 Y:16
DS3 IDR 0.025 0.40
Elevator (D1014.012) DS1 PFA 0.31 0.45 ea. ND:1 - - - - -
Residual Drift DS1 R-IDR 0.01 0.3 ea. ND:1 - - - - - TRC
IDR, Inter-Storey Drift Ratio; PFA, Peak Floor Acceleration; R-IDR, Residual Inter-Storey Drift Ratio; TRC, Total Replacement Cost.
4.2. Refined Loss Analysis

The refined loss analysis was performed according to the PEER-PBEE methodology. In
this case, the structural analysis results were taken for the nonlinear time-history analyses
(NLTH) described in Section 3.3, performed with the considered nine groups of earth-
quake records associated with nine levels of seismic intensity. The loss analysis was
performed by applying the software PACT, inserting the hazard information in Section
3.2, the results of the mentioned NLTH analyses and the damageable inventory and
costing information described in Section 4.1 and listed in Table 1. In particular, for
each intensity level, the values of the EDPs obtained for each earthquake record were
introduced in the software PACT for the loss assessment. While more than 10 ground
motion pairs could have improved the likely dispersion in demands, the process followed
in this work does permit the performance of simplified and more rigorous loss assessment
approaches to be compared.
The economic losses shown in this section are reported as the Expected Annual Loss
(EAL), which is the expected losses in any given year expressed in percentage of building
replacement cost, in order to show the annualized repair costs considering the entire range
of seismic hazard at the site, in a probabilistic manner. Four intensity-based assessments at
81%, 63%, 10%, and 5% probability of exceedance in 50-year hazard level were selected
based on the values provided by the Italian Code (NTC 2018) to define the seismic
intensity for the evaluation at the different limit states.
Additionally, a sensitivity analysis was performed in order to evaluate the change in the
loss estimation with different assumptions done in the damage analysis, which are
described as follows:
● Basic model: model based on the fragility functions and repair cost information
previously described in Section 4.1 and Table 1. These models are referred to as BF.1
and IF.1.
● Variation of the damage states of RC columns: The fragility curves for the RC
elements were selected originally from the PACT library. However, according to
the Italian Code (NTC 2018), the drift limits calculated for each column (taking the
NTC prescriptions as median values) for the SLV and SLC are slightly different
compared with the values of the assumed fragility functions. In the first two models,
referred to as BF.2 and IF.2, only the DS3 is changed to the limit state SLC. In the
models BF.3 ad IF.3 both damage states DS2 and DS3 are taken as the limit state of
the code, SLV and SLC, respectively.
● Variation of the residual drift fragility function: The median value of the fragility
curve of PACT for the residual drift was changed to 0.8% in the models BF.4 and
IF.4, and to 1.2% in the models BF.5 and IF.5. The dispersion was kept at the value of
0.3, recommended in FEMA P-58 (2012) and provided in PACT.
A summary of the median repair cost estimates for each model is shown in Table 2, while
Fig. 12 illustrates the influence of the different components in the losses obtained for the
intensity levels considered in the basic models. As shown in Table 2, for the basic models
the largest losses are obtained for the BF model, with an EAL of 0.306%, where the
masonry infill walls were not considered in the structural analysis but were still included
132
L. LANDI ET AL.
Table 2. Comparison of key loss metrics for the different models, including the different options of sensitivity study.
Model BF.1 BF.2 BF.3 BF.4 BF.5 IF.1 IF.2 IF.3 IF.4 IF.5
EAL € 9368 9440 9450 9389 9349 7719 7793 7825 8110 7218
[%] [0.306%] [0.308%] [0.308%] [0.306%] [0.305%] [0.252%] [0.254%] [0.255%] [0.265%] [0.236%]
Annualized Probability to Complete Loss due to Residual Drift [%] 0.0235% 0.0235% 0.0235% 0.0251% 0.0219% 0.0569% 0.0569% 0.0569% 0.0842% 0.0366%
Losses at limit states of NTC2008 € [%] 81% in 57600 22600
50 years (SLO) [1.88%] [0.74%]
63% in 103462 61500
50 years (SLD) [13.5%] [8.38%]
10% in 39000 350000
50 years (SLV) [12.73%] [11.4%]
5% in 1577500 1700000 1705000 1600000 1572500 1025000 1045000 1050000 1035000 970000
50 years (SLC) [51.5%] [55.5%] [55.6%] [52.2%] [51.3%] [33.4%] [34.1%] [34.3%] [33.8%] [31.7%]
Figure 12. Influence of the components on the repair cost at intensity level of (a) 81% in 50 years, (b)
63% in 50 years, (c) 10% in 50 years and (d) 5% in 50 years, with the EDPs of the NLTH analysis of BF
and IF models.
in the repair cost analysis. For the IF model, where the infill walls were accounted for in
the structural response, the EAL has a value of 0.252%. These results show the variation of
the EAL by considering or not the masonry infill walls as part of the structural system
during the analysis phase. In this case, the infills caused a reduction of seismic demand
and a consequent reduction of EAL. It is also seen in Fig. 12 that the highest contributor
to economic losses in the lower three intensity levels is damage on the masonry infill walls
and partitions, while for the higher intensity level masonry infills become the second most
important contributor to losses, where the first is the contribution of the damage to the
structural elements. These results, in addition to the corresponding MAFE at the intensity
levels (higher MAFE in the lower intensity levels and vice versa), confirm that masonry
components are significant to the loss analysis results.
The results of the sensitivity analysis also indicate a small variation in the EAL value
due to different assumptions of the RC element damage states and the building residual
inter-storey drift fragility curve. This result shows that the fragility curves selected for the
RC columns are in accordance with the SLV and SLC limit states of the Italian Code (NTC
2018): in particular, the median values recommended in FEMA P-58 (2012) for DS2 and
DS3 are very similar to the drift limits of the columns of the critical storeys for the limit
states SLV and SLC. This result shows also that the residual drifts obtained, due to the
structure response under the seismic actions, did not have a relevant influence in the loss
estimation for the current case study. The reason is that the obtained residual drifts were
very low to produce significant economic losses. The only variation in the repair cost due
to the changes of the sensitivity analysis is apparent at the largest intensity level (see Table
2) but the MAFE associated to this level is small, obtaining the same EAL in practical
terms as the basic models.
The values in Table 2 can be compared with the real-sustained costs described in
Section 3.1 for the repair costs, which were reported earlier as being around 20% of the
replacement cost. Considering the infill model, note that a repair cost of around 20% the
replacement cost is predicted to occur for intensity levels having between 10% and 5%
probability of exceedance in 50 years. In order to evaluate if the predicted levels of loss
are similar to those observed, these intensity levels can be compared with the intensity of
the shaking at the site during the L’Aquila earthquake. For this purpose, reference is
made to the records obtained at four stations located close to the site and the epicentre
(within a distance of around 4–5 km) and their associated response spectra (Iervolino
134 L. LANDI ET AL.
et al. 2010). In particular, the acceleration response spectra of the average of the two
horizontal components of these records, given in the aforementioned reference, are
examined due to the uncertainty as to the real ground motions at the base of the building
and along its main axes. Considering the spectral acceleration just defined at a period of
1.5 s, it is noted that one station provided a value close to a 10% probability of
exceedance, two stations provided values between those relative to 10% and 5% prob-
ability of exceedance (0.14 g and 0.19 g, respectively), and one station recorded shaking
that was more intense than a 5% probability of exceedance. As these intensity levels are
similar to that predicted to cause similar levels of repair costs, this data indicates that the
loss-assessment procedure adopted here is reasonably accurate for the building
considered.
4.3. Simplified DBA Loss Estimation

In this case, the structural analysis results were taken from the DDBA procedure, and the
seismic demand was determined as explained in Section 3.4 for the nine intensity levels.
Initially, it was applied the software PACT, in order to have an estimate of the different
values of loss for the nine considered intensity levels.
The simplified assessment procedure of PACT requires a single vector of EDPs for both
directions at each intensity level, assuming them as the median response values, with
dispersion factors incorporated in the loss analysis. In order to estimate the individual
inter-storey drift profiles for each intensity level, the displacement at each storey is
obtained by equating the displacement of the equivalent SDOF system with the inelastic
spectral displacement demand, considering the assumed displacement shape and
a multiplier factor obtained in an iterative process. In order to estimate the maximum
residual inter-storey drift of the models, 20% of the value of the maximum inter-storey
drift in both directions for all storeys was assumed, for each intensity level. This assump-
tion was made instead of other proposals in the literature (Ruiz-García and Chora 2015)
in order to have a conservative and simple definition, considering also the influence of
residual inter-storey drifts in the previous sensitivity analysis. The dispersion for the
residual drifts was taken as 0.3, consistent with the assumption of Section 4.1. The
acceleration demands were estimated using the approach described in FEMA P-58
(2012). Also, the dispersions of the EDPs were determined using the values suggested in
FEMA P-58 (2012), which attributes to the median inter-storey drift and floor acceleration
responses the total dispersion factors βSD and βFA, respectively, as shown in Equations (12)
and (13):
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
βSD ¼ βaΔ 2 þ βm 2 (12)
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
βFA ¼ βaa 2 þ βm 2 (13)
where βaΔ and βaa are the analysis record-to-record dispersion for drift and accelerations,
respectively, which consider random uncertainty in the seismic demands, and βm is the
modelling dispersion that incorporates the uncertainty related to the structural response.
These three dispersion factors are taken from FEMA P-58.
Subsequently, the simplified loss analysis of Welch, Sullivan, and Calvi (2014) was
applied, which considers a tri-linear loss curve, where only in the two intermediate
points the loss estimation must be performed to obtain the EAL estimate. With regard to
the two bounding limit states, the Zero-Loss limit state is considered with a MAFE of the
lowest intensity level (30 year return period) and an MDF of zero. For the other
boundary point, the MDF of 1.0, corresponding to the total replacement cost, has
been associated with the MAFE of the highest intensity level (2475 yr return period).
The two intermediate limit states are called operational and damage control limit states
and are defined by the reaching of given values of inter-storey drift. The operational
limit state is considered as peak inter-storey drift of 0.33%, being this value the damage
state DS2 of the masonry infill walls. According to (Sassun et al. 2016) below this value
some repairs will be necessary, however, not leading to significant repair effort or
disruption in the facility. For the damage control limit state, Welch, Sullivan, and
Calvi (2014) suggest to consider this threshold as the first yielding of any storey, being
the first transition point from linear to non-linear behaviour in the capacity curve. From
the results obtained from the DDBA procedure, this transition in the capacity curve
occurs for an intensity level relative to 10% probability of exceedance in 50 years for
both BF and IF models, with a peak inter-storey drift of 0.92% at the fourth storey. The
loss values associated with the intensity levels relative to the reaching of the two
intermediate states were taken from the results of PACT. Table 3 presents the four
points considered for the tri-linear loss curve. Table 3 reflects the assumptions of the
procedure of Welch, Sullivan, and Calvi (2014) for the zero-loss and near collapse limit
states. Comparing Table 3 with Table 2 for an intensity with 10% probability of
exceedance in 50 years, it is possible to notice similar values of loss for the BF structure
and slightly lower values for the IF structure. Table 4 shows the comparison between the
EAL values obtained with the refined loss analysis (Table 2) and those found with the
simplified approach for both models.
Table 3. Four limit states considered for the tri-linear loss curve according to the procedure of Welch,
Sullivan, and Calvi (2014).
Bare Frame (BF) Infilled Frame (IF)
Limit states for Tri-linear Int. Economic Loss Int. Economic Loss [€]
curve Level MAFE [€] Level MAFE EVD1 EVD2
Zero-Loss 81% in 0.0333 0 81% in 0.0333 0 0
50 years 50 years
Operational 50% in 0.0139 127941 22% in 0.0050 131875 156667
50 years 50 years
Damage Control 10% in 0.0021 377500 10% in 0.0021 249412 289167
50 years 50 years
Near Collapse 2% in 50 years 0.0004 3064500 2% in 50 years 0.0004 3064500 3064500
Table 4. Comparison between the values of EAL obtained with the refined procedure and the
DBA simplified estimation of Welch, Sullivan, and Calvi (2014).
EAL euros [%]
BF IF-EVD1 IF-EVD2
Refined loss estimation 9368 [0.306%] 7719 [0.252%]
DBA simplified loss estimation 8388 [0.274%] 6474 [0.211%] 6952 [0.227%]
136 L. LANDI ET AL.
From Table 4, the EAL for the BF model obtained using the DDBA simplified approach
has a value of 0.274%, being 0.895 times the EAL of the refined loss assessment, showing
a good prediction of this parameter. For the IF model, the EAL obtained for damping
approaches EVD1 and EVD2 are 0.211% and 0.227%, which are 0.853 and 0.901 times the
EAL obtained with the refined loss analysis. These results show that the DDBA simplified
loss estimation gives reasonable loss approximations, being for all cases the EAL lower
than the value obtained with the refined loss model. Comparing the two damping
approaches for the IF model, they did not produce significant differences, with
a percentage variation of EAL of about 5%. Among them, in this case study, the EVD2
determined values more close to the refined loss estimation.
5. Retrofitting Schemes and Cost-benefit Analyses

Different retrofit schemes are considered in order to improve the seismic performance of
the structure. These retrofit schemes are then compared in terms of their effects on the
loss assessment. In particular, three main options are considered: two of them based on
the addition of an inter-storey damping system, the third on the insertion of new RC walls
(RCW). One of the two damping systems is made of nonlinear fluid-viscous dampers
(VD), the other is made of T-ADAS hysteretic damping devices (HD). An important issue
is the definition of proper design criteria for the different retrofit options and in particular
for the two different damping systems, in order to make them comparable by choosing the
same performance objective.
The seismic demand for the design of the retrofit schemes has been evaluated con-
sidering the elastic response spectrum defined by the Italian code (NTC 2018) for the site
of the building and for the same soil type (A). The design of the retrofit schemes has been
performed by defining the objectives with reference to the response of the bare frame
structure. Since the previous analyses were performed by applying the main seismic
component along the X direction, which is the weakest direction, the design of the retrofit
options has been performed in order to improve the response of the structure along this
direction. Therefore, the starting point of the design was the pushover curve of the bare
frame along the X direction. In the design procedure, this pushover curve has been
transformed into an equivalent elastic-plastic curve according to code criteria (NTC
2018): equal area under the pushover curve and the elastic-plastic curve up to a top
displacement corresponding to attainment of the SLC limit state, elastic stiffness equal to
the secant stiffness at 60% of the lateral strength at the yielding point of the elastic-plastic
curve (Fig. 13).
5.1. Design Criteria for the Considered Retrofit Options

For the two considered damping systems, the devices are positioned in plan at the corners
of the structure, as is illustrated in Fig. 14. The design of both the retrofit options with
damping devices is based on the same performance objective, which is to achieve a target
reduction of the displacement demand of the bare frame under the seismic intensity level
associated with a return period of 475 years. The spectral displacement demand of the
bare frame, called here Df, is determined in the ADRS format at the intersection of the
capacity curve and the reduced demand spectrum corresponding to an equivalent
3000
2500
Base shear [kN]

2000
1500
1000
500
0
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
Top Displacement [m]
Figure 13. Elastic-plastic idealization of the pushover curve of the bare frame in X direction.
Figure 14. Position of the damping devices in plan and along the height.
damping given by the sum of the inherent and hysteretic damping: ξeq = ξi + ξhyst. Then,
the target spectral displacement demand of the retrofitted structure is obtained by apply-
ing a target reduction factor ηdes to the spectral displacement demand of the bare frame:
D = ηdes Df. For both VD and HD damping systems, it was assumed ηdes = 0.55 (Fig. 15) as
this value was considered a practical target considering the added damping that could be
provided to such a system.
It should be recalled that fluid-viscous dampers provide mainly supplemental damp-
ing, while the hysteretic devices provide supplemental damping and also increase of
stiffness and strength of the structure. The design of the two damping systems is
performed according to the two different procedures briefly described below. The pro-
cedure for the fluid-viscous dampers is similar to the one proposed by Landi, Lucchi, and
Diotallevi (2014). The procedure for the hysteretic dampers is similar to that of Benedetti,
Landi, and Merenda (2014), with some differences as those related to the distribution of
the device properties between the storeys. For more details see the aforementioned
references.
Once the target spectral displacement D is determined, the procedure for the design of
the fluid-viscous dampers is made by the following steps:
138 L. LANDI ET AL.
Figure 15. Reduction of displacement demand with viscous dampers (on the left) and hysteretic
devices (on the right).
(1) The elastic-plastic system is replaced with an equivalent elastic system, using the
secant stiffness at the displacement demand, and the effective period Te correspond-
ing to the secant stiffness is obtained.
(2) The elastic spectral acceleration Ael is obtained as a function of Te and the damping
reduction factor (η in the Italian code) is calculated knowing the yield spectral
acceleration Ay: η = Ay/Ael.
(3) The effective damping is obtained inverting the expression of the damping reduc-
tion factor of Italian code (NTC 2018) (with the damping ratio specified as
a percentage): ξeq = 10/η2-5.
(4) The required supplemental viscous damping ratio for an elastic behaviour ξve is
calculated from the following expression, where ξve is the only unknown since the
hysteretic damping and the frame ductility demand μ, given by the ratio of the
target displacement demand and the yield displacement, are known (α is the
damper exponent and is set equal to 0.5):
α
eq ¼ i þ ve μ12 þ hyst (14)
The damping coefficients CN of each nonlinear viscous damper are obtained using the
expression of the first mode supplemental damping proposed by Ramirez et al. (2000) and
assuming a constant distribution of the damping coefficients. In this expression, given the
inter-storey positioning of the dampers and the damper exponent, the only unknown
remains CN. Considering α = 0.5, the supplemental damping under elastic behaviour was
ξve = 23%, while the damping coefficient of the single devices was CN = 576.1 kN(s/m)0.5.
The procedure for the design of the hysteretic devices requires the determination of an
equivalent hysteretic damper which represents the damping system in the ADRS format.
This equivalent damper is defined by three parameters, yield acceleration, yield displace-
ment, and peak displacement: Ayd, Dyd, and Dd. Assuming that the maximum damper
displacement Dd is equal to that of the frame structure and adopting an acceptable level of
ductility demand μd for the device, it is possible to derive Dyd = Dd/μd. The design
procedure is then aimed at defining Ayd in terms of the ratio r = Ayd/Ay, where Ay is
the yield spectral acceleration of the bare frame, in order to provide a target spectral
displacement demand. In particular, this requires the following iterative steps:
(1) A tentative value of the ratio r = Ayd/Ay to define the yield force of the equivalent
damping system is fixed. Then, having prefixed the ductility demand and peak
damper displacement, the elastic-plastic capacity curve of the equivalent damper is
obtained.
(2) By combining the elastic-plastic capacity curve of the equivalent damper and that of
the bare frame it is possible to derive a trilinear capacity curve of the retrofitted
structure. This trilinear curve is then transformed into an elastic-plastic curve.
(3) The equivalent damping of the whole system can then be determined knowing the
ductility demand μ of the bare frame, the ductility demand μd of the equivalent
damper and the ratio r (Ramirez et al. 2000):

hyst 2r 1
eq ¼ i þ þ 1 (15)
1 þ r πð1 þ rÞ μd
(1) The spectral displacement demand of the retrofitted structure is then determined
with the capacity spectrum method and compared with the target displacement D.
If they are different, a new value of r is fixed and steps 1 to 4 are repeated until
convergence.
Once the properties of the equivalent damper have been defined, the properties of the
single devices to be placed at the different storeys should be derived. The properties of the
single devices are determined assuming that they sustain the same ductility demand μd of
the equivalent damper and considering that their maximum displacement corresponds to
the inter-storey drift. This drift can be obtained knowing the spectral displacement
demand and considering the associated displacement profile. The yield force of each
device can be derived in two ways: by distributing the yield force obtained for the
equivalent damper, or, alternatively, assuming that the ratio between the single device
stiffness and the storey elastic stiffness is the same as the ratio between the equivalent
damper stiffness and the frame elastic stiffness. This second way has been assumed in this
case study, and this ratio was equal to about 1.8, while the obtained value of r was 0.32.
The following values of lateral strength Vdj and stiffness Kdj were obtained for the devices
at the different storeys: first storey, Vd1 = 317 kN, Kd1 = 141750 kN/m; second storey, Vd2
= 246 kN, Kd2 = 73500 kN/m; third storey, Vd3 = 158 kN, Kd3 = 47250 kN/m; fourth
storey, Vd4 = 88 kN, Kd4 = 26250 kN/m; fifth storey, Vd5 = 69 kN, Kd5 = 15103 kN/m; sixth
storey Vd6 = 59 kN, Kd6 = 13289 kN/m.
A retrofit option with RC walls positioned in the lateral bays of the two outer frames in
the X direction has been considered as an alternative to supplemental damping devices
(same position of damping devices, see Fig. 14). Since the dimensioning of the wall, to
achieve the same displacement reduction obtained with dissipative devices, would have
140 L. LANDI ET AL.
required large values of stiffness and strength, and consequently of costs, the walls have
been designed on the basis of the geometrical and reinforcement rules provided by the
Italian Building Code (NTC 2018).
5.2. Structural Modelling and Analysis Results for the Considered Retrofit Options
The nonlinear fluid-viscous dampers and the T-ADAS devices have been added in the
same concentrated plasticity model used for the nonlinear analyses of the original struc-
ture according to the parameters derived from the design procedures. The RC walls have
been modelled with concentrated plasticity beam-column element connected to the frames
with rigid-links. Then, pushover analyses and the nonlinear time-history analyses have
been performed for the retrofitted options. In particular, the seismic demand for the nine
considered intensity levels was determined based on the results of pushover analyses. The
time-history analyses were performed only for the four intensity levels corresponding to
the four code limit states and were used for comparison with the results of pushover
analyses.
The seismic demand for the nine intensity levels was determined, with the pushover
analyses, in the ADRS format at the intersection between the capacity curve and the
reduced spectral demand curve. In this case, however, differently from what described in
Section 3.4 and according to what described at the beginning of this section, the capacity
curve has been transformed in a bilinear elastic-plastic curve, and the demand spectra, in
order to be consistent with the retrofit design approach, were obtained from the Italian
code spectra (NTC 2018). Figure 16 shows the maximum inter-storey drift obtained for
each intensity level and for the different retrofit options with pushover and nonlinear
time-history analyses.
Figure 16. Maximum inter-storey drift derived with the pushover analyses for the nine intensity levels
(a) and comparison with results from time-history analyses (b).
With regard to nonlinear time-history analyses, the maximum drift along the height
was obtained for each ground motion and then the values relative to the different ground
motions of the same intensity level have been averaged. Since the time-history analyses
have been performed only for the aforementioned four intensity levels, the points related
to the highest intensity level (return period of 2475 years) have been obtained for
comparison purposes by extending the last branch of the curve. From Fig. 16 it is possible
to notice that fluid-viscous dampers provided lower values of drift than hysteretic dampers
at the lower intensity levels and that the seismic demand determined in the ADRS format
with the pushover analyses was conservative in comparison with the nonlinear time-
history analyses. The lower values obtained with fluid-viscous dampers at the lower
intensity levels in comparison with the other retrofit options and in particular with the
one with hysteretic dampers are probably due to this reason. At the lower intensity levels,
the hysteretic dampers were probably in the elastic range or did sustain little excursion
beyond the yield point. In this condition, the reduction of displacement was due mainly to
the increase of stiffness and not to the added damping, and for the considered hysteretic
devices, the increase of stiffness produced a lower reduction of displacement than that
obtained with fluid-viscous dampers. Similar comparisons can be made with reference to
the considered intervention with RC walls.
It should be observed that with increasing intensity level, the curve of the maximum
drift obtained for the retrofit with HD tends to the curve relative to the option with VD.
This is due to the adopted design criterion. The values in Fig. 16a relative to a return
period of 475 years, obtained for the solutions with HD and VD were similar but not equal
since the design criterion was aimed at equating the spectral displacement, i.e. the
displacement of the equivalent SDOF system, not the inter-storey drift, which can be
different if the displaced shapes are different.
5.3. Simplified Loss Estimates for the Retrofit Solutions and Cost-benefit Analyses
As explained in the previous section, the structural analysis results for the different retrofit
solutions were taken from the pushover analyses and from the intersection in the ADRS
format between the capacity curve and the demand spectra relative to the nine considered
intensity levels. Unlike the simplified loss estimates based on DDBA performed for the
existing structure, these analyses assumed the code spectra as representative seismic
demand in place of the average spectra of selected records. Further, the elastic-plastic
idealization of the capacity curve was considered in place of the obtained pushover curve.
Therefore, in these conditions, the loss estimates were performed again for the bare frame,
in order to have a comparison between the different retrofit options and the existing
structure for the same analysis criteria.
The inter-storey drift profiles corresponding to the nine intensity levels were extracted
from the analysis results. The other relevant response quantities such as the floor accel-
erations were obtained according to the simplified approach outlined in FEMA P-58
(2012). The same was done for the dispersion values. Subsequently, the simplified loss
analysis of Welch, Sullivan, and Calvi (2014) was applied, which considers a tri-linear loss
curve, where only in the two intermediate points the loss estimation must be performed to
obtain the EAL estimate. The same criteria and intermediate limit states considered for the
existing building and previously described were adopted. Table 5 shows the obtained
142 L. LANDI ET AL.
Table 5. Comparison between the values of EAL obtained with the DBA simplified estimation of Welch,
Sullivan, and Calvi (2014) for the bare frame and the different retrofit options.
EAL euros [%]
BF VD HD RCW
DBA simplified loss estimation 10266 [0.335%] 4957 [0.162%] 6167 [0.201%] 6835 [0.223%]
Table 6. Benefit-cost ratios and break even times for the

different retrofit options.
Structure VD HD RCW
Benefit-cost ratio 1.47 1.14 1.02
tbreakeven (years) 34 44 49
values of EAL for the bare frame and the different retrofit options. It should be observed
that the value obtained for the bare frame in Table 5 is larger than those in Table 4. This is
due to the previous-mentioned differences adopted in the structural analysis, which
provided more conservative values. As expected, the three considered retrofit options
returned a significant reduction of EAL. As a consequence of the differences observed in
the analysis results, the solution with VD provided larger reduction of EAL in comparison
with the other retrofit options.
By making a rough estimate of the cost of the retrofit intervention as 6% of the
replacement cost, limited only to the introduction of dampers and not including the
strengthening of the existing structural members, the values of benefit-cost ratios and
break even times (Calvi 2012) illustrated in Table 6 were obtained (neglecting the infla-
tion). The differences between these values therefore reflect those obtained for the EAL in
Table 5.
6. Conclusions
Nonlinear static and time-history analyses have been performed on a model of a real
existing RC building damaged during the L’Aquila earthquake, 2009. Then, a new version
recently proposed for the DDBA procedure has been applied in order to provide estimates
of the results of the pushover curves. With regard to loss assessment, the refined proce-
dure based on the PEER-PBEE approach has been compared with a simplified displace-
ment-based for loss assessment. By comparing the repair costs reported for the building in
the L’Aquila earthquake with those predicted for similar intensity levels, it is concluded
that the loss assessment procedure is reasonably accurate. Finally, different retrofit options
have been developed and their effectiveness in reducing losses has been examined.
The nonlinear analyses highlighted the differences caused by considering or neglecting
the infill panels in the model. The obtained displacement profiles were compatible with
the damage observed after the L’Aquila earthquake, which was concentrated mainly in the
storeys above the basement, and it was more evident in the last four storeys. The
application of the DDBA procedure, recently refined by Sullivan et al. (2018), provided
good estimates of the displacement profiles and pushover curves obtained from advanced
non-linear analyses. As such, this paper has demonstrated it to be particularly suitable for
the seismic assessment of existing RC frame buildings.
The loss analyses highlighted the significant effect that damage to non-structural
elements at the lower intensity levels has on EAL values. The DDBA simplified loss
estimation approach of Welch, Sullivan, and Calvi (2014) proved to be effective in
predicting EAL values when an appropriate displaced shape was used to determine the
structural capacity. Additionally, the importance of appropriately selecting two intermedi-
ate limit states of the simplified tri-linear curve used in the Welch, Sullivan, and Calvi
(2014) approach was highlighted. For the bare and infilled frame models, the simplified
DDBA loss analysis provided good estimates of the EAL obtained from the refined loss
assessment, with an acceptable level of approximation for a quick loss analysis, without the
use of computer structural models and complex analysis.
The simplified loss analysis that was applied to different retrofit options permitted
evaluation of the effectiveness of different interventions to be gauged using a single
parameter, the EAL, as a measure of performance. From this study, it was evident that
the sensitivity of the EAL in Italian RC buildings to drift demand is high, with the most
effective reduction of EAL obtained using fluid-viscous dampers as a retrofit solution since
this provided the largest reduction of drift demands at all intensity levels.
Acknowledgments
The study presented in this article was developed within the activities of the ReLUIS-DPC
2014–2018 research program. The support of the RELUIS consortium is gratefully acknowledged.
The authors would like to thank also Prof. A. Benedetti of the University of Bologna and the design
office Benedetti and Partners for providing detailed information about the case study.
Funding
This work was funded by Presidenza del Consiglio dei Ministri—Dipartimento della Protezione
Civile.
ORCID
Luca Landi http://orcid.org/0000-0001-7259-1186
Timothy J. Sullivan http://orcid.org/0000-0002-6997-5434
References
Al-Chaar, G. 2002. Evaluating strength and stiffness of unreinforced masonry infill structures. US
Army Corps of Engineers, Engineer Research and Development Center.
Ay, B. O., M. J. Fox, and T. J. Sullivan. 2017. Technical note: practical challenges facing the selection
of conditional spectrum-compatible accelerograms. Journal of Earthquake Engineering 21 (1):
169–80. doi: 10.1080/13632469.2016.1157527.
Benedetti, A., L. Landi, and D. G. Merenda. 2014. Displacement-based design of an energy
dissipating system for seismic upgrading of existing masonry structures. Journal of Earthquake
Engineering 18 (4): 477–501. doi: 10.1080/13632469.2014.897274.
144 L. LANDI ET AL.
Berry, M. P., and M. O. Eberhard. 2005. Practical performance model for bar buckling. ASCE
Journal of Structural Engineering 131 (7): 1060–70. doi: 10.1061/(ASCE)0733-9445(2005)
131:7(1060).
Calvi, G. M. 2012. Alternative choices and criteria for seismic strengthening. Proceedings of the
15th World Conference on Earthquake Engineering, Lisbon, Portugal. doi: 10.1094/PDIS-11-11-
0999-PDN.
Carr, A. J. 2013. RUAUMOKO inelastic time-history analysis of three-dimensional framed structures.
Christchurch, New Zealand: University of Canterbury.
Crisafulli, F. J., and A. J. Carr. 2007. Proposed macro-model for the analysis of infilled frame
structures. Bulletin of the New Zealand Society of Earthquake Engineering 40 (2): 69–77. doi:
10.5459/bnzsee.40.2.69-77.
Decanini, L., F. Mollaioli, A. Mura, and R. Saragoni. 2004. Seismic performance of masonry infilled
R/C frames. Proc. of the 13th World Conference on Earthquake Engineering, Vancouver,
Canada.
FEMA. 2012. Seismic performance assessment of buildings: Vol. 1 – Methodology, FEMA P-58-1.
Prepared by the Applied Technology Council for the Federal Emergency Management Agency,
Washington, DC.
Glaister, S., and R. Pinho. 2003. Development of a simplified deformation-based method for seismic
vulnerability assessment. Journal of Earthquake Engineering 7 (Special Issue 1): 107–40. doi:
10.1080/13632460309350475.
Hak, S., P. Morandi, G. Magenes, and T. J. Sullivan. 2012. Damage control for clay masonry infills
in the design of RC frame structures. Journal of Earthquake Engineering 16 (Supp 1): 1–35. doi:
10.1080/13632469.2012.670575.
Haselton, C. B., A. Whittaker, A. Hortacsu, J. Baker, J. Bray, and D. Grant. 2012. Selecting and
scaling earthquake ground motions for performing response-history analyses NIST GCR 11-917-
15. Proceedings of the 15th World Conference on Earthquake Engineering, Lisbon, Portugal.
Iervolino, I., F. De Luca, E. Chioccarelli, and M. Dolce 2010. L’azione sismica registrata durante il
mainshock del 6 aprile 2009 a L’Aquila e le prescrizioni del DM 14/ 01/2008(in italian). http://
www.reluis.it.
Landi, L., S. Lucchi, and P. P. Diotallevi. 2014. A procedure for the direct determination of the
required supplemental damping for the seismic retrofit with viscous dampers. Engineering
Structures 71: 137–49. doi: 10.1016/j.engstruct.2014.04.025.
Landi, L., A. Tardini, and P. P. Diotallevi. 2016. A procedure for the displacement-based seismic
assessment of infilled RC frames. Journal of Earthquake Engineering 20 (7): 1077–103. doi:
10.1080/13632469.2015.1112324.
Montejo, L. A., and M. J. Kowalsky. 2007. Cumbia: Set of codes for the analysis of reinforced concrete
members. Raleigh: North Carolina State University.
Morandi, P., S. Hak, and G. Magenes. 2014. In-plane experimental response of strong masonry
infills. 9th International Masonry Conference, International Masonry Society, Guimarães.
NTC. 2008. Norme Tecniche Per Le Costruzioni. DM 14/ 01/2008. Rome, Italy (in Italian).
NTC. 2018. Aggiornamento delle Norme Tecniche Per Le Costruzioni. DM 17/ 01/2018. Rome,
Italy (in Italian).
O’Reilly, G. J., R. Monteiro, A. M. Nafeh, T. J. Sullivan, and G. M. Calvi. 2019. Displacement-based
framework for simplified seismic loss assessment of existing buildings. Journal of Earthquake
Engineering 1–22. doi: 10.1080/13632469.2020.1730272.
OPCM n. 3274. 2003. Primi elementi in materia di criteri generali per la classificazione sismica del
territorio nazionale e di normative tecniche per le costruzioni in zona sismica, G.U. 8 maggio
2003 (in Italian).
Ozkaynak, H., E. Yuksel, C. Yalcin, A. A. Dindar, and O. Buyukozturk. 2014. Masonry infill walls in
reinforced concrete frames as a source of structural damping. Earthquake Engineering &
Structural Dynamics 43 (7): 949–68. doi: 10.1002/eqe.2380.
Porter, K. A. 2003. An overview of PEER’s performance-based earthquake engineering methodol-
ogy. Proceedings of the Ninth International Conference on Applications of Statistics and
Probability in Civil Engineering, San Francisco, CA.
Priestley, M. J. N. 1993. Myths and fallacies in earthquake engineering - conflicts between design
and reality. Bulletin of the New Zealand National Society for Earthquake Engineering 26 (3):
329–41.
Priestley, M. J. N. 2000. Performance based seismic design. Proceedings of the 12th World
Conference on Earthquake Engineering, Auckland, New Zealand. doi: 10.5459/bnzsee.33.3.325-
346.
Priestley, M. J. N. 2003. Myths and fallacies in earthquake engineering, revisited, 2003. Pavia: IUSS
Press.
Priestley, M. J. N., G. M. Calvi, and M. J. Kowalsky. 2007. Displacement based seismic design of
structures. Pavia, Italy: IUSS Press.
Ramirez, O. M., M. C. Constantinou, C. A. Kirche, A. S. Whittaker, M. W. Johnson, J. D. Gomez,
and C. Z. Chrysostomou. 2000. Development and evaluation of simplified procedures for analysis
and design of buildings with passive energy dissipation systems. Report MCEER-00-0010,
Multidisciplinary Center for Earthquake Engineering Research. Buffalo: State University of
New York at Buffalo.
Ruiz-García, J., and C. Chora. 2015. Evaluation of approximate methods to estimate residual drift
demands in steel framed buildings. Earthquake Engineering and Structural Dynamics 44 (15):
2837–54. doi: 10.1002/eqe.2611.
Saborío-Romano, D. 2016. Performance based and simplified displacement based assessment of
infilled RC frame building in L’Aquila, Italy. MSc Thesis, IUSS Pavia, Italy.
Sassun, K. P. 2014. A parametric investigation into the seismic behaviour of window glazing
systems. Dissertation, European School for Advanced Studies in Reduction of Seismic Risk
(ROSE School), University of Pavia, Italy.
Sassun, K. P., T. J. Sullivan, P. Morandi, and D. Cardone. 2016. Characterizing the in-plane seismic
performance of infill masonry. Bulletin of the New Zealand Society for Earthquake Engineering 49
(1): 98–115. doi: 10.5459/bnzsee.49.1.98-115.
SEAOC. 1995. Vision 2000, a framework for performance-based engineering. Sacramento: Structural
Engineers Association of California.
SHARE. 2013. EU-FP7 project: Seismic hazard harmonization in europe, collaborative project
coordinated by the swiss seismological service. Zurich, Switzerland.
Sullivan, T. J., M. J. N. Priestley, and G. M. Calvi, editors. 2012. A model code for the displacement-
based seismic design of structures, DBD12. Pavia: IUSS Press.
Sullivan, T. J., D. Saborio-Romano, G. J. O’Reilly, D. P. Welch, and L. Landi. 2018. Simplified
pushover analysis of moment resisting frame structures. Journal of Earthquake Engineering. 1–28.
forth comming. doi: 10.1080/13632469.2018.1528911.
Sullivan, T. J., D. P. Welch, and G. M. Calvi. 2014. Simplified seismic performance assessment and
implications for seismic design Earthquake Engineering and Engineering Vibration 13 (1): 95–122.
doi: 10.1007/s11803-014-0242-0.
Welch, D. P., T. J. Sullivan, and G. M. Calvi. 2014. Developing displacement-based procedures for
simplified loss assessment in performance-based earthquake engineering. Journal of Earthquake
Engineering 18 (2): 290–322. doi: 10.1080/13632469.2013.851046.

Displacement Based Simplified Seismic Loss Assessment of Post 70s RC Buildings - 2020

Uploaded by

Copyright:

Available Formats

You might also like

Displacement Based Simplified Seismic Loss Assessment of Post 70s RC Buildings - 2020

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Displacement Based Simplified Seismic Loss Assessment of Post 70s RC Buildings - 2020

Uploaded by

Copyright:

Available Formats

Journal of Earthquake Engineering

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/ueqe20

Displacement-Based Simplified Seismic Loss

Luca Landi , Daniel Saborio-Romano , David P. Welch & Timothy J. Sullivan

To link to this article: https://doi.org/10.1080/13632469.2020.1735577

Published online: 08 Jun 2020.

Submit your article to this journal

View related articles

View Crossmark data

Citing articles: 1 View citing articles

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

Displacement-Based Simpliﬁed Seismic Loss Assessment of

ABSTRACT ARTICLE HISTORY

Kowalsky 2007) is a Direct Displacement-Based Design (DDBD) procedure based on the

2. Displacement-Based Loss Assessment (DBLA) Framework and Speciﬁc

3. The Case Study: An Existing RC Frame Building in L’Aquila

3.2. Location, Site Hazard and Ground Motions

3.3. Structural Modelling and Analysis Results

B as e She ar/Total We ight [%]

B as e She ar/To tal We ight [%]

elements should be considered also, especially in presence of a structure not designed to

3.00% 30yrs 3.00% 30yrs

3.4. Direct Displacement-Based Assessment Procedure (DDBA)

Base Shear [kN]

4. Loss Estimates for the Case Study Existing Building

4.1. Damageable Inventory and Costing Information

Damage First Second Third Fourth Fifth Sixth

4.2. Reﬁned Loss Analysis

4.3. Simpliﬁed DBA Loss Estimation

5. Retroﬁtting Schemes and Cost-beneﬁt Analyses

5.1. Design Criteria for the Considered Retroﬁt Options

Base shear [kN]

Table 6. Beneﬁt-cost ratios and break even times for the

You might also like