Fragility Analysis of Pangasinan State University Urdaneta City
Campus Buildings
RUTH ANN D. MANINGDING1,*, JAN NICHOLAS S. BALDO2
1College of Engineering and Architecture,
Pangasinan State University,
Urdaneta City, Pangasinan, 2428,
PHILIPPINES
2College of Engineering and Architecture,
University of Cordilleras,
Baguio City, Benguet, 2600,
PHILIPPINES
*Corresponding Author
Abstract: – The Philippines is a seismically susceptible region because of its unique geographical location within the
Pacific Ring of Fire. Over the years, the Philippines has had several destructive earthquakes that have caused
building collapses and numerous fatalities. Schools are particularly vulnerable to earthquakes because most were
built using outdated building rules that do not comply with current seismic design criteria. The age of PSU buildings,
which ranges from 10 to 40 years, may increase their vulnerability to seismic hazards. Buildings designed with the
old codes might lack the necessary reinforcement and structural elements to withstand the forces generated by an
earthquake, potentially leading to catastrophic failures and endangering the lives of students, faculty, and other
occupants. Moreover, the San Manuel Fault Line near PSU buildings in Urdaneta City, Pangasinan, heightens the
risk of a significant seismic event affecting these structures. Given the critical role that schools play in communities,
both in providing education and serving as essential facilities during and after disasters, assessing their seismic
vulnerability becomes crucial. So far, no research has been conducted to evaluate the PSU building’s vulnerability to
seismic activity in the event of a significant earthquake. In line with this, this study aimed to perform a seismic
fragility analysis of PSU buildings in Urdaneta City, Pangasinan, Philippines, to determine how susceptible they
were to earthquakes. This study applies preliminary assessment using Rapid Visual Screening of FEMA P-154 to
filter out which buildings need further evaluation and detailed assessment using fragility analysis. The screened
buildings are evaluated using fragility curves to assess if the building could endure an earthquake with a 0.4 g PGA
and a 10% probability of exceedance, following the National Structural Code of the Philippines (NSCP)
requirements for seismic zone 4. The structural model of PSU buildings was created using SAP 2000, and the non-
linear static analysis, specifically the ATC40 Capacity Spectrum Method, was performed to determine the data
required to develop the fragility curves. The results demonstrate that the seismic scores of Engineering Buildings 1,
2, and 3 are below the RVS FEMA P-154 standard of 2.0, indicating that further investigation is required to evaluate
their vulnerability thoroughly. Additionally, these buildings were observed to withstand a maximum peak ground
acceleration of 0.60 g PGA at a 10% probability of exceedance based on the developed fragility curves,
corresponding to earthquake intensity up to VIII, indicating "severe shaking.” Furthermore, analysis of the fragility
curves demonstrates that none of the structures exceed the 10% probability of exceedance at 0.4 g PGA, aligning
with NSCP standards for Seismic Zone 4. As a result, these buildings are considered safe for occupancy without
requiring retrofitting measures.
Key-Words: - earthquake, rapid visual survey, seismic vulnerability assessment, fragility analysis, capacity spectrum
method, probability of exceedance.
Received: June 13, 2024. Revised: October 25, 2024. Accepted: November 21, 2024. Published: December 12, 2024.
1 Introduction
1.1 Background of the Study
Earthquakes can occur unexpectedly and without
warning, [1]. When tectonic plates move along a
fault line in the crust of the earth, it can generate
intense shaking of the ground, which is known as an
earthquake, [2]. The Philippines is prone to seismic
activity due to its distinctive geographic position
within the Pacific Ring of Fire. Over time, the
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.75
Ruth Ann D. Maningding, Jan Nicholas S. Baldo
E-ISSN: 2224-3496
789
Volume 20, 2024
nation has experienced devastating earthquakes,
resulting in building failures and significant loss of
life, [3].
Schools are particularly vulnerable to
earthquakes because most were built using outdated
building rules that do not comply with current
seismic design criteria, and others were constructed
without seismic design. Despite the inclusion of
seismic provisions in modern construction, older
buildings might still pose risks to occupants due to
their age and deterioration, [4], [5]. The seismic
vulnerability assessment of schools needs special
attention due to its essential role in the community,
both in education and post-earthquake function.
Schools play a valuable additional role in disaster
recovery efforts. When there is an emergency,
schools are usually utilized as evacuation centers,
[6]. The Department of Education in the Philippines
has designated schools as evacuation centers,
ensuring they are readily available during disasters,
[7]. During emergencies, a more secure and resilient
school may save children's lives and help get things
back to normal in the community, [8].
Due to its location close to the San Manuel
Fault Line, PSU buildings will be affected by a
significant seismic event. The San Manuel fault line
is 12.7 kilometers away from the structures,
according to the Phivolcs Fault Finder App, [9].
Moreover, the age range of most PSU buildings is
10 to 40 years old. The age of the structures may
contribute to their susceptibility to earthquake
hazards. The structural reliability and integrity of
the buildings may decrease over time due to normal
wear and tear, environmental conditions, and
structural deterioration, [10]. These structures were
designed and built using outdated building codes
that fail to meet current seismic design standards.
This study analyzed the seismic fragility of
PSU buildings in Urdaneta City, Pangasinan,
Philippines, to determine their susceptibility to
earthquakes. It applied preliminary assessment
using Rapid Visual Screening and detailed
assessment using fragility analysis of the buildings.
Rapid Visual Screening (RVS) of FEMA P-154
provides an initial assessment of the buildings
against earthquake hazards to filter out which
buildings need further evaluation. The seismic
susceptibility of the screened structures is evaluated
using fragility curves, which depict the likelihood of
collapse or damage at different levels of ground
shaking.
One of the primary distinctions of this study is
its specific focus on PSU Urdaneta City Campus
Buildings. By developing fragility models that are
tailored to the unique construction practices,
materials, geological conditions, and environmental
factors of Urdaneta City, this research directly
addresses the specific needs of the region. This
localization means that the findings directly apply to
the city's specific context, making them highly
relevant for local stakeholders, including city
planners, engineers, policymakers, and school
administration. They can use these precise insights
to develop more effective mitigation strategies and
improve the overall seismic resilience of PSU
Urdaneta City Campus buildings and other public
buildings within the region.
1.2 Objectives of the Study
This study was conducted to carry out a fragility
analysis of PSU buildings in Urdaneta City,
Pangasinan, Philippines, to evaluate their
susceptibility to earthquake damage. Specifically:
1. To identify those buildings that need to be
prioritized for further detailed seismic
evaluation based on the cut-off score of FEMA
P-154.
2. To determine the damage level of the selected
PSU buildings when subjected to a peak ground
acceleration of 0.4 g.
3. To obtain the seismic vulnerability of the
selected PSU buildings based on the developed
fragility curves.
4. To determine the maximum possible intensity
that PSU buildings can withstand.
1.3 Conceptual Framework
The conceptual framework for this study applies a
preliminary and detailed assessment of the
buildings, as depicted in Figure 1. Rapid Visual
Screening (RVS) is one of the valuable techniques
for rapidly determining the seismic risk of
significant structures, such as school buildings.
These procedures will initially assess a building's
vulnerability to seismic hazards. The buildings that
have undergone screening with scores less than the
cut-off of RVS FEMA P-154 are assessed for
seismic vulnerability using fragility analysis, which
shows the likelihood of failure or damage at various
intensities of ground shaking.
Fig. 1: Conceptual Framework
Rapid Visual
Screening Fragility
Analysis
Seismic
Vulnerability
Assesment
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.75
Ruth Ann D. Maningding, Jan Nicholas S. Baldo
E-ISSN: 2224-3496
790
Volume 20, 2024
1.4 Scope and Delimitation
The study focuses on conducting a fragility analysis
of PSU buildings in Urdaneta City, Pangasinan,
Philippines. The Rapid Visual Screening (RVS)
method from FEMA P-154 prioritizes buildings and
identifies those requiring additional evaluation. The
PSU Urdaneta Campus has eleven (11) buildings
considered for the study, as presented in Table 3.
Actual measurements were conducted to verify the
plans' dimensions, and the rebound hammer test was
performed to assess the on-site compressive strength
of the concrete. Also, to ensure accurate data for
generating the fragility curves, the researchers
conducted additional tests, including a tensile test,
to determine the actual tensile strength of the rebars.
This approach differentiates this study from others
that rely solely on the specifications provided in the
plans. The exposed rebars at the top of columns
were acquired, and a tensile test was done using the
Universal Testing Machine (UTM) on steel bars to
assess the tensile strength of reinforcing bars.
SAP 2000 will create and simulate structure
models subject to earthquakes. The major structural
elements of the building are all included in the
structural modeling, which is limited only by the
concrete works. Microsoft Excel will run the
simulation, analyze the results, and produce the
fragility curves using the data collected. Nonlinear
static "pushover" analysis will be carried out to
generate capacity curves. The pushover parameters
will follow the provisions of the ATC-40 Capacity
Spectrum Method (CSM). The Incorporated
Research Institutions for Seismology (IRIS)
databases were used to collect the ground motion
data (GMD). The structural models were subjected
to seven (7) local earthquakes and five (5)
international earthquakes. The seven (7) earthquakes
that were recorded in the Philippines from 1988 to
2023 are displayed in Table 1 (Appendix). The five
(5) earthquakes that occurred in foreign countries
from 1996 to 2023 are presented in Table 2
(Appendix).
2 Methodology
The descriptive analytical case study method is
selected to align with the necessary methodology,
blending descriptive and analytical approaches to
assess the building's susceptibility to seismic events
thoroughly. A descriptive approach was employed
for the preliminary assessment of the PSU buildings
using Rapid Visual Survey forms. The analytical
component evaluated the building's structural
behavior and response to seismic events. A
computational model of the PSU buildings was
created using SAP 2000 software, and non-linear
static analysis, specifically the ATC40 Capacity
Spectrum Method, was performed to determine the
data required to generate the fragility curves. These
models simulated the building's response to seismic
forces and helped assess its vulnerability and
potential damage.
2.1 Research Design
Figure 2 is an illustration of the proposed procedure.
PSU buildings will undergo a preliminary
assessment utilizing Rapid Visual Screening to
determine whether buildings require additional
investigation. The obtained structural plan will be
modeled using SAP 2000 software, and pushover
analysis and response spectrum analysis will be
carried out using the ground motion data gathered.
Pushover curves and response spectra, in turn,
determine an earthquake's demand and a building's
capacity. These first two components serve as inputs
for the Capacity Spectrum Method used in structural
assessments. Creating a fragility curve includes
statistical analysis of the structural response data.
Fig. 2: Research Design
2.2 Population and Locale of the Study
The PSU Urdaneta City Campus is located along
McArthur Highway in Barangay San Vicente,
Urdaneta City, Pangasinan, Philippines, covering an
area of 25,499 square meters. The city's precise
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.75
Ruth Ann D. Maningding, Jan Nicholas S. Baldo
E-ISSN: 2224-3496
791
Volume 20, 2024
geographical position is 15.9835° N latitude and
120.6334° E longitude. Figure 3 depicts the PSU
Urdaneta Campus' site development plan. Eleven
(11) school buildings, including the Graduate
School building, were involved in the case study.
The 11 buildings included in the study are listed in
Table 3.
Fig. 3: PSU Urdaneta City Campus Site
Development Plan
Table 3. Name of buildings in PSU Urdaneta City
Campus
Building Number
Building Name
BLDG 1
Academic Building 1
BLDG 2
Academic Building 2
BLDG 3
Academic Building 2 (extension)
BLDG 4
Engineering Building 1
BLDG 5
Engineering Building 2
BLDG 6
Engineering Building 3
BLDG 7
Educational Building
BLDG 8
General Education Building
BLDG 9
PTBI (Pangasinan Technology
Business Incubator Center)
BLDG 10
School of Advanced Studies
BLDG 11
Student Activity Center
2.3 Data Gathering Procedure
This study aims to create analytical fragility curves
for the buildings at Pangasinan State University
(PSU) Urdaneta Campus using a four-phase
technique indicated in Figure 4.
.
Fig. 4: Flow of Activity
2.3.1 Phase 1: Rapid Visual Screening (RVS)
Rapid Visual Screening (RVS) is an initial filter to
identify buildings with potential seismic
vulnerabilities, allowing for further evaluation and
prioritization of retrofitting efforts. RVS is a simple
yet effective method, [11]. The methodology
employed in the RVS procedure relies on
conducting a walkway inspection of a structure and
completing a Data Collection Survey Form by the
surveyor after visually examining the structure
externally and, when possible, internally, without
engaging in complex computations, [12]. The first
step was research planning, which involved pre-
field planning activities, pre-field data collection,
and preliminary site investigation. The pre-field
planning activities include selecting the data
collection forms, training the screeners to conduct
the RVS, and the review of the acquired data. The
next step is the execution of rapid visual screening.
The final section involves analyzing the collected
data, [13]. Using the FEMA P-154 RVS checklist,
the researchers determine each building's
vulnerability score. Buildings with a score of two or
below are advised to undergo a more thorough
evaluation, [14]. While Rapid Visual Screening
(RVS) may lack the precision of detailed modeling,
it provides a simple and effective approach for
identifying parts of a city prone to earthquake
vulnerability, [15], [16], [17].
2.3.2 Phase 2: Structural Modelling of PSU
Buildings
Acquire architectural and structural plans from the
university's physical plant office and verify
dimensions through on-site measurements. A
rebound hammer test will be used to estimate the
compressive strength of the concrete materials used
in the structures. This involves selecting
Phase 4
Development of Seismic Fragility Curves
Phase 3
The Capacity Spectrum Method as a Structural
Assessment Methodology
Phase 2
Structural Modelling of PSU Buildings
Phase 1
Rapid Visual screening of PSU Buildings
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.75
Ruth Ann D. Maningding, Jan Nicholas S. Baldo
E-ISSN: 2224-3496
792
Volume 20, 2024
representative samples of critical exterior columns
and beams, with three test locations chosen for each
beam and column. The Schmidt hammer, commonly
known as a Swiss hammer or rebound hammer, is a
specialized tool to evaluate the rebound or elasticity
of materials like concrete and rock, [18]. Its
primary function lies in determining surface
hardness and resistance to penetration. Acquire
exposed rebars at the top of columns and conduct a
tensile test using a Universal Testing Machine
(UTM) to assess the tensile strength of reinforcing
bars. The following are the procedures for creating a
three-dimensional frame model of the structure in
SAP 2000, [19].
1. Specify the grid system and develop a three-
dimensional frame model of the structure,
including the dimensions, shapes, and
arrangements of various structural elements.
2. Define material and geometric properties for
each structural element, such as material
properties of concrete and rebars, section
properties, and boundary conditions.
3. Define the load pattern and assign a gravity
load.
2.3.3 Phase 3: The Capacity Spectrum Method
The Capacity Spectrum Method (CSM), a
performance-based seismic evaluation approach,
determines a structure's seismic capacity, [20]. This
process involves generating the structure's capacity
curve by performing a pushover analysis. Response
spectra illustrate the demands of the seismic ground
motion, [21]. The intersection of the capacity and
demand curves is known as the performance point.
The capacity curve was derived using pushover
curves employing traditional pushover analysis, and
the demand curve was derived from the inelastic
response spectra of specific ground motion
recordings, [22]. The Capacity Spectrum Method
(CSM) was conducted using the following
methodologies:
1. Perform pushover analysis on the PSU building
structural model using SAP 2000.
2. Acquire ground motion data from the
Incorporated Research Institutions for
Seismology (IRIS). Find the normalized Peak
Ground Acceleration (PGA). The normalized
ground motion data is produced by scaling the
original ground motion data up and down, [23].
Convert normalized PGA into a response
spectrum through SeismoSignal.
PGANormalized=(GMD) 󰇡PGAExcitation
PGAMaximum 󰇢 (1)
where:
GMD = Ground Motion Data
= the level of excitation ranges
from peak ground acceleration (PGA) values,
starting from 0.1 g up to 3.0 g.
=absolute maximum ground
motion data
3. Determine the performance points by analyzing
the intersection of the capacity and response
spectrum curves. Several values of performance
points are obtained by scaling the demand curve
into various excitation levels, ranging from 0.1
g to 3.0 g.
2.3.4 Phase 4: Development of Seismic Fragility
Curves.
Fragility curves show the risk of structural damage
caused by varying levels of ground shaking. An
analytical method was used to produce these
fragility curves, and the following is a thorough
step-by-step procedure, [24], [25].
1. Determine the damage state threshold using
Table 4 and the data (yield displacement and
ultimate displacement) from the pushover curve
to calculate the limits of spectral displacement
for each damage state.
2. Establish the damage ranking by applying the
calculated damage state thresholds.
3. Determine the damage states corresponding to
all displacement performance points obtained
from the capacity curve and response spectrum.
Table 4. Damage State Threshold Values, [26],[27]
Damage
Rank
Spectral Displacement
D
C

B

A
󰇟󰇛󰇜󰇠
As

Where:
dy = yield displacement, du = ultimate displacement
4. Count the frequency of occurrence for each
damage state threshold with the performance
points.
5. Compute the damage ratio by obtaining the total
cumulative probability of occurrence. This ratio
is calculated by dividing the number of
occurrences by the total number of records at a
specific Peak Ground Acceleration (PGA) level.
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.75
Ruth Ann D. Maningding, Jan Nicholas S. Baldo
E-ISSN: 2224-3496
793
Volume 20, 2024
6. Determine the standard deviation and mean
accordingly. The mean and standard deviation
were obtained for every damage state using
equations (2) and (3). The lognormal values of
PGAs were then multiplied by each frequency
to provide lognormal values for the standard
deviation and mean.
λ=x
N (2)
ξ=(x- λ)2
N-1 (3)
Where:
x = individual ground motion data obtained
N= sample size of ground motion data
λ = mean of the ground motion data
ξ = standard deviation
7. The probability of exceedance (Pr) may be
calculated by applying equation (4).
󰇣󰇝󰇛󰇜󰇞
󰇤 (4)
Where:
Pr=probability of exceedance
x = peak ground acceleration
λ = mean
= standard deviation
=standard normal deviation
8. Plot the likelihood of exceedance versus the
peak ground acceleration (PGA) excitation level
to create seismic fragility curves for the PSU
buildings.
3 Results and Discussion
3.1 Rapid Visual Screening (RVS) Results
The RVS scores of all 11 school buildings are
displayed in Figure 5. Three of the buildings have
scores below 2.0, which is the FEMA P-154 cutoff
score. Eight of the eleven school buildings had
ratings higher than 2, indicating better seismic
performance. Engineering Buildings 1 (BLDG4),
Engineering Buildings 2 (BLDG5), and Engineering
Buildings 3 (BLDG6) had scores lower than the
cutoff of 2. This suggests that a more thorough
examination of these structures is required to
precisely determine their vulnerability to seismic
risks.
Fig. 5: RVS Score
3.2 Structural Modelling of PSU Buildings
Out of the eleven school buildings, Engineering
Buildings 1 (BLDG4), Engineering Buildings 2
(BLDG5), and Engineering Building 3 (BLDG6)
had scores that were below the cutoff of 2,
indicating the need for a more thorough
investigation using fragility analysis. The structural
model was created based on architectural and
structural plans, and material properties were
incorporated using SAP 2000. All the buildings'
main structural elements are included in the
structural modeling, with the focus of the research
being on the beams, girders, and columns that make
up the structure. Accordingly, the finished structural
models of school buildings are shown in isometric
perspective in Figure 6, Figure 7 and Figure 8.
Fig. 6: Structural Model of Engineering Building 1
Fig. 7: Structural Model of Engineering Building 2
BLD
G 1 BLD
G 2 BLD
G 3 BLD
G 4 BLD
G 5 BLD
G 6 BLD
G 7 BLD
G 8 BLD
G 9 BLD
G 10 BLD
G 11
Σειρά1 2,5 3,4 2,8 0,9 1,5 0,9 3,4 3,4 2,5 2,5 2,5
0
1
2
3
4
RVS SCORE
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.75
Ruth Ann D. Maningding, Jan Nicholas S. Baldo
E-ISSN: 2224-3496
794
Volume 20, 2024
Fig. 8: Structural Model of Engineering Building 3
3.2.1 Rebound Hammer Test
The rebound hammer test was conducted on the
school buildings to estimate the on-site compressive
strength of the concrete structures. The procedure
involves pressing the device against the concrete
surface until the hammer impacts the surface a
designated number of times at each test location.
The resulting readings from these impacts are
interpreted using a rebound hammer test chart, and
the average of these readings is used to determine
the compressive strength. A thorough explanation of
the methodologies for performing a rebound
hammer test is available in the American Standards
for Testing and Materials (ASTM C805) journal,
[28]. Tests were carried out for one column and
beam in each building at three locations close to
supports and at midspan. Instead of conducting tests
on each column and beam, a sampling strategy was
implemented. This method focused on selecting a
representative sample of each building's critical
exterior columns and beams. Specifically, three test
locations were carefully chosen for each beam and
column. Ten readings should be taken from each
test area and averaged to determine the compressive
strength. Table 5 (Appendix) summarizes the
compressive strength used for each building
obtained from the rebound hammer test.
3.2.2 Tensile Test
The exposed rebars at the top of columns were
obtained for each building to assess the tensile
strength of reinforcing bars. The sample steel bars
were tested using the Universal Testing Machine at
the authorized materials testing and geotechnical
engineering laboratory facility. During the tensile
test, the material specimen is subjected to a
gradually increasing axial load until it reaches the
point of failure or fracture, as shown in Figure 9.
The yield point and ultimate tensile strength
summary were determined as shown in Table 6
(Appendix) and used in the structural model as
material properties.
Fig. 9: Fractured Specimen of Engineering Building
1
3.3 The Capacity Spectrum Method as a
Structural Assessment Methodology
3.3.1 Pushover Curve
Figure 10, Figure 11 and Figure 12 display the
pushover curves for Engineering Building 2,
Engineering Building 1, and Engineering Building 3
along X- and Y-directions. The graphs depict the
relationship between displacement (x-axis) and base
shear force (y-axis). The peak point on the pushover
curve indicates the structural material's capacity to
withstand maximum lateral loads. In contrast, the
yield point indicates the structure’s boundary within
the elastic range, [29]. By utilizing the yielded
displacement (dy) and ultimate displacement (du)
and inputting these values into the equations
outlined in Table 2 (Appendix), one can calculate
the damage thresholds for each damage state
accordingly. In Engineering Building 2, as depicted
in Figure 10, the yield point was reached in the x-
direction at a base shear force of 3026.783 kN,
corresponding to a displacement of 20.079 mm. The
maximum displacement, on the other hand, was
recorded at 173.476 mm, with the structure enduring
a base shear force of 3402.297 kN. Similarly, in the
y-direction, the yield point was observed at a base
shear force of 1796.186 kN with a displacement of
26.437 mm, while the maximum displacement of
263.273 mm occurred at a base shear force of
2089.338 kN. The maximum base shear force of
3402.297 kN in the x-direction is higher than
2089.338 kN in the y-direction. This result suggests
that Engineering Building 2 can sustain a greater
base shear force in the x-direction, indicating that
this is the building's strongest axis. In contrast,
Engineering Buildings 1 and 3 can tolerate higher
base shear force in the y-direction, as shown in
Figure 11 and Figure 12, respectively, implying that
the y-direction is the strongest axis for these
buildings.
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.75
Ruth Ann D. Maningding, Jan Nicholas S. Baldo
E-ISSN: 2224-3496
795
Volume 20, 2024
Fig. 10: Pushover Curve of Engineering Building 2
in the X and Y directions
Fig. 11: Pushover Curve of Engineering Building 1
in the X and Y-direction
Fig. 12: Pushover Curve of Engineering Building 3
long X and Y-direction
3.3.2 Ground Motion Inputs
All ground motion records are obtained from the
Incorporated Research Institutions for Seismology
(IRIS), as shown in Figure 13. For these records, 12
actual local and foreign earthquake events are
considered, as shown in Appendix in Table 1 and
Table 2. Using Microsoft Excel, each ground
motion has been normalized and scaled accordingly
from 0.1g to 3.0g, as shown in Figure 14. There are
720 scaled earthquake recordings when the elastic
response spectra of the East-West and North-South
directions are considered for each event. As seen in
Figure 15, the response spectra are obtained using
SeismoSignal software. The response spectra were
run in SAP 2000 and converted to demand
spectrum.
Fig. 13: Ground Motion Records from Incorporated
Research Institutions for Seismology (IRIS)
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.75
Ruth Ann D. Maningding, Jan Nicholas S. Baldo
E-ISSN: 2224-3496
796
Volume 20, 2024
Fig. 14: Normalized Peak Ground Acceleration of
Luzon Earthquake (02-24-1988) of magnitude of 7.2
Fig. 15: Response Spectra based on Luzon
Earthquake (02-24-1988) of magnitude 7.2 with
0.1g PGA along East-West direction using
SeismoSignal Software
3.3.3 Capacity Spectrum Method (CSM)
The capacity spectrum approach defines the
performance point as the intersection of the capacity
and response spectrum curves. Following the ATC-
40 Capacity Spectrum Method (CSM), the capacity
curve obtained from the pushover curve can be
superimposed onto the demand spectrum curve,
[30]. Several performance point values are obtained
by scaling the demand curve into various excitation
levels between 0.1 g and 3.0 g. A total of 1440
performance points were determined for a single
model in the east-west and north-south directions,
corresponding to 12 ground motion records.
Fig. 16: Capacity Curve based on July 16, 1990
Luzon, Philippine’s earthquake with 1.0g PGA
based on magnitude 6.5 in the east-west direction at
X-axis of Engineering Building 2
Figure 16 illustrates the superimposed capacity
curve sample for Engineering Building 2 at 1.0 g
PGA, corresponding to a magnitude 6.5 earthquake
in Luzon, Philippines, on July 16, 1990. A summary
of performance points achieved when the
Engineering Buildings 2 model in SAP 2000 was
exposed to an earthquake of magnitude 6.5 that
occurred on July 16, 1990, in Luzon, Philippines, is
shown in Table 7 (Appendix).
3.4 Development of Seismic Fragility Curves
3.4.1 Damage State Threshold Limits
The spectral displacement for the buildings on both
x and y axes is determined in this study by applying
the approach, shown in Table 4. The data presented
in Table 8 (Appendix) outline the structural damage
state thresholds for both the y-axis and x-axis
directions of each school building. In addition to the
no-damage condition (D), the damage thresholds
used in this study are categorized into four groups:
slight damage (C), moderate damage (B), extensive
damage (A), and complete damage (As).
Table 8 (Appendix) shows that the spectral
displacement values for C and B are small in both
the x and y directions. Damage thresholds with
small values have a higher likelihood of being
exceeded. The damage thresholds for A and As are
comparatively greater than the spectral
displacements for C and B. For example,
Engineering Building 2 displayed a damage
threshold of 173.476 mm for As in the x-direction.
This indicates that to achieve "complete damage,” a
spectral displacement of 173.476mm is required.
This data was then used to determine the damage
rank and generate a set of fragility curves.
3.4.2 Damage Ranking
The provided tabulated data for the damage ranking
was generated utilizing the acquired performance
points and the damage state threshold limits. Table 9
(Appendix) depicts the performance points
associated with scaled peak ground acceleration
values and the corresponding damage rankings for
Engineering Building 2 along the East-West
direction (x-axis), derived from data recorded
during the magnitude 8.0 earthquake near the east
coast of Peru on August 15, 2007. The data
indicates higher damage states correspond to higher
peak ground acceleration values. Specifically, Table
9 (Appendix) shows that Engineering Building 2
sustained complete damage (As) at a peak ground
acceleration of 2.0 g during this seismic event.
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.75
Ruth Ann D. Maningding, Jan Nicholas S. Baldo
E-ISSN: 2224-3496
797
Volume 20, 2024
3.4.3 Probability of Occurrence
The Occurrences of earthquakes for each damage
state threshold were tallied using the performance
points. Table 10 (Appendix) summarizes the
occurrence frequency for different damage
conditions observed in Engineering Building 2
along the y-axis in the East-West direction based on
varying levels of peak ground acceleration (PGA)
measured in gravitational units (g). This table
generates a graph showing the probability of
occurrence for each damage situation. The
cumulative damage ratios, representing the
correlation between each damage rank and its
corresponding PGA value, were the basis for
constructing the probability of occurrence graphs.
Figure 17, Figure 18 and Figure 19 illustrate the
percentage of each damage rank associated with the
PGA. These graphs indicate that the damage rank
consistently rises as the PGA increases from 0.1 g to
3.0 g.
For Engineering Building 2, as shown in Figure
17, the graph illustrates the probability of
occurrence of different damage conditions for
Engineering Building 2 along the y-direction as a
function of peak ground acceleration (PGA). At
PGA’s up to 0.4 g, the probability of experiencing
no damage to slight damage (state D) is 100%,
indicating that the building remains mostly intact.
As the PGA increases from 0.5 g to 1.2 g, the
building will experience slight to moderate damage
(states C and B). This range indicates that the
building suffers from minor structural issues that
gradually escalate as the seismic intensity increases.
Fig. 17: Probability of Occurrence for Engineering
Building 2 in the East-West Direction(y-axis).
The building sustains moderate to extensive
damage as we approach 1.3 g-1.9 g (states B and A).
This suggests that at these levels of seismic activity,
the building's structural integrity is compromised to
a greater extent, resulting in substantial damage that
may require significant repairs or lead to partial
structural failure. Finally, the structure will
experience extensive to complete damage in the 2.0
g–3.0 g range (states A and As). This indicates that
the building will likely to experience catastrophic
structural failure, leading to complete damage.
Figure 18 illustrates the probability that
Engineering Building 1 will receive slight to
moderate damage, ranging from 0.5 g to 1.3 g, along
the east-west direction along the y-axis. The
building sustains moderate to extensive damage as
we get closer to 1.4 -2.8g. Finally, the structure will
undergo extensive damage, ultimately leading to
complete damage within the 2.9–3.0g range.
Fig. 18: Probability of Occurrence for Engineering
Building 1 in East-West Direction (Y-axis)
Figure 19 depicts the likelihood of Engineering
Building 3 experiencing slight to moderate damage
between 0.2 g and 1.1 g in the east-west direction
(y-axis). The building sustains moderate to
extensive damage as we approach 1.2-2.6 g.
Finally, the structure will experience extensive to
complete damage in the 2.7–3.0 g range.
Fig. 19: Probability of Occurrence for Engineering
Building 3 in East-West Direction (y-axis)
3.4.4 Fragility Curves
The fragility curves evaluate the likelihood of
reaching or exceeding each damage condition
(slight, moderate, extensive, and complete) over
peak ground accelerations (PGA) from 0.1 g to 3.0
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.75
Ruth Ann D. Maningding, Jan Nicholas S. Baldo
E-ISSN: 2224-3496
798
Volume 20, 2024
g. The probabilities (Pr) for each damage condition
were shown on a single graph, resulting in a clear
and interpretable representation in Appendix in
Figure 20, Figure 21 and Figure 22.
In practical terms, the "D" or "No Damage"
condition should be excluded from the fragility
curves, [29]. The figures show the probabilities for
exceeding a damage condition with its associated
scaled peak ground acceleration.
For Engineering Building 2, depicted in Figure
20 (Appendix), the likelihood of surpassing the
specified ground acceleration of 0.4 g is as follows:
in the east-west direction along the X-axis, it's
23.41%, 5.49%, 5.20%, 5.22%, and 5.47% for no
damage (D), slight damage (C), moderate damage
(B), extensive damage (A), and complete damage
(As), respectively. Along the Y-axis, it's 43.30%,
7.50%, 5.61%, 5.20%, and 5.29% for the same
corresponding levels of damage. Likewise, in the
north-south direction along both the X and Y axes,
the probabilities of surpassing a peak ground
acceleration of 0.4 g are as follows: 17.52%, 5.55%,
5.19%, 5.21%, and 5.39% for no damage (D), slight
damage (C), moderate damage (B), extensive
damage (A), and complete damage (As),
respectively. Along the Y-axis, the probabilities are
40.09%, 6.26%, 5.34%, 5.19%, and 5.27% for the
same levels of damage.
As illustrated in Figure 21 (Appendix), the
probabilities of exceeding a peak ground
acceleration of 0.4 g for Engineering Building 1 in
the east-west direction along the X and Y axes are
40.17%, 6.53%, 5.46%, 5.25%, and 5.23%, and
18.89%, 5.35%, 5.32%, 5.19%, and 5.44%,
respectively. These probabilities correspond to
damage levels of no damage (D), slight damage (C),
moderate damage (B), extensive damage (A), and
complete damage (As). Similarly, in the north-south
direction along the X-axis, the probabilities of
exceeding a peak ground acceleration of 0.4g are
39.98% for no damage (D), 6.74% for slight damage
(C), 5.38% for moderate damage (B), 5.19% for
extensive damage (A), and 5.33% for complete
damage (As). Along the Y-axis, the probabilities are
18.15% for no damage (D), 5.31% for slight damage
(C), 5.25% for moderate damage (B), 5.23% for
extensive damage (A), and 0% for complete damage
(As).
As illustrated in Figure 22 (Appendix), the
probabilities of exceeding a peak ground
acceleration of 0.4 g for Engineering Building 3 in
the east-west direction along the X-axis are 55.46%,
21.98%, 8.07%, 5.31%, and 5.24%. Along the Y-
axis, the probabilities are 53.88%, 21.07%, 6.84%,
5.19%, and 5.39%. These values correspond to
damage levels of no damage (D), slight damage (C),
moderate damage (B), extensive damage (A), and
complete damage (As), respectively. Similarly, in
the north-south direction along the X-axis, the
probabilities of exceeding a peak ground
acceleration of 0.4 g for Engineering Building 3 are
53.96% for no damage (D), 18.18% for slight
damage (C), 6.24% for moderate damage (B),
5.22% for extensive damage (A), and 5.24% for
complete damage (As). Along the Y-axis, the
probabilities are 54.23% for no damage (D), 17.62%
for slight damage (C), 6.82% for moderate damage
(B), 5.19% for extensive damage (A), and 5.33% for
complete damage (As). The resulting fragility
curves indicate that as peak ground acceleration
increases, the likelihood of exceeding all damage
states also rises.
The study considered the seismic requirements
outlined in the National Structural Code of the
Philippines (NSCP), focusing on the probability of
exceedance (Pr) for the "complete damage (As)"
state. This assessment involved evaluating the
impact of a 0.4 g PGA earthquake along the east-
west and north-south axes of each building.
According to the NSCP, structures in the Philippines
are designed to endure a peak ground acceleration of
0.4g with a 10% probability of exceedance,[31],
which is also included in the CSIRO handbook
entitled "Designing Resilient”, [32]. An extract
from the book mentioned above is presented in
Figure 23.
Fig. 23: Basic Design PGA
In addition to the NSCP provision, the Structural
Engineers Association of California (SEAOC) has
the following excerpt: “A structure with 30 or more-
year lifespan is not SAFE when subjected to a
seismic event with a 10% probability of exceeding
the collapse or complete damage. The structure,
being more than 50 years old, is vulnerable to large-
magnitude earthquakes”, [33].
Table 11.A (Appendix) shows that under the
condition of complete damage (As), the
probabilities of exceedance for Engineering
Buildings 1, 2, and 3 in the East-West direction
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.75
Ruth Ann D. Maningding, Jan Nicholas S. Baldo
E-ISSN: 2224-3496
799
Volume 20, 2024
remain relatively consistent, with maximum values
of approximately 5%. Similarly, as summarized in
Table 11.B (Appendix), the probabilities of
exceedance at 0.4g PGA for complete damage (As)
in the North-South direction are approximately 5%
for Engineering Buildings 2 and 3. However,
Engineering Building 1 demonstrates a 0%
probability of exceedance under the complete
damage (As) condition in the Y-Axis, indicating no
vulnerability to this level of seismic loading for this
building. Furthermore, fragility curve analysis,
which evaluates the likelihood of structural damage
under seismic events, revealed that none of these
three buildings exceeded the 10% probability of
exceedance at 0.4 g PGA under “complete damage
(As).” This analysis conforms to the criteria
established by the National Structural Code of the
Philippines (NSCP) for buildings in Seismic Zone 4,
demonstrating that the structures are safe for
occupancy and do not require retrofitting measures.
The maximum peak ground acceleration (PGA)
values that Engineering Building 1 can withstand
with a 10% likelihood of exceedance are displayed
by the generated fragility curves shown in Table
12.A and Table 12.B in Appendix. In the east-west
direction(x-axis), the building can endure slight
damage of up to 0.50 g, moderate damage of up to
0.53 g, extensive damage of up to 0.55 g, and
collapse damage of up to 0.60 g. Along the y-axis in
the same direction, the building can withstand slight
damage of up to 0.55 g, moderate damage of up to
0.56 g, extensive damage of up to 0.59g, and
collapse damage of up to 0.60g. Similarly, in the
north-south direction along the x-axis, the analysis
revealed that Engineering Building 1 could
withstand the following maximum peak ground
accelerations (PGA) for potential earthquakes: 0.49
g for 'slight damage', 0.54 g for 'moderate damage',
0.58 g for 'extensive damage', and 0.60 g for
'collapse damage. Along the y-axis in the north-
south direction, the building can withstand slight
damage of up to 0.56g, moderate damage of up to
0.57g, and extensive damage of up to 0.58g.
The analysis of Engineering Building 2 indicates
its ability to withstand various levels of peak ground
acceleration (PGA) with a 10% probability of
exceedance. Along the east-west direction (x-axis),
the building can sustain slight damage of up to 0.55
g, moderate damage of up to 0.58 g, extensive
damage of up to 0.58 g, and collapse damage of up
to 0.60g. Along the y-axis in the same direction, it
can endure slight damage of up to 0.46 g, moderate
damage of up to 0.53 g, extensive damage of up to
0.57 g, and collapse damage of up to 0.60g.In the
north-south direction (x-axis), Engineering Building
2 can withstand slight damage of up to 0.55 g,
moderate damage of up to 0.59 g, extensive damage
of up to 0.60 g, and collapse damage of up to 0.60 g.
Along the y-axis in this direction, it can sustain
slight damage of up to 0.50 g, moderate damage of
up to 0.56 g, extensive damage of up to 0.59 g, and
collapse damage of up to 0.60 g.
The seismic fragility curves for Engineering
Building 3 in the east-west orientation show its
ability to withstand various levels of peak ground
acceleration (PGA) with a 10% probability of
exceedance. Along the x-axis, the building can
endure slight damage of up to 0.25 g, moderate
damage of up to 0.43 g, extensive damage of up to
0.56 g, and collapse damage of up to 0.60 g. Along
the y-axis in the same orientation, it can sustain
slight damage of up to 0.25 g, moderate damage of
up to 0.48 g, extensive damage of up to 0.58 g, and
collapse damage of up to 0.60 g. In the north-south
(x-axis), Engineering Building 3 can withstand
slight damage of up to 0.29 g, moderate damage of
up to 0.50 g, extensive damage of up to 0.58 g, and
collapse damage of up to 0.58 g. Along the y-axis in
this direction, it can endure slight damage of up to
0.29 g, moderate damage of up to 0.48g, extensive
damage of up to 0.58 g, and collapse damage of up
to 0.60g.
The maximum peak ground acceleration (PGA)
associated with a 10% probability of exceedance for
every damage condition for all structures is
displayed in Appendix in Table 12.A and Table
12.B. The results indicate that all three buildings
would sustain "complete damage (As)" at a
maximum PGA of 0.60 g. Consequently, it was
determined that these buildings can withstand a
maximum of 0.60 g, which exceeds the Philippine
National Structural Code's (NSCP) minimum
standard of 0.4g. According to the Modified
Mercalli Scale, a PGA of 0.60 g corresponds to
Intensity VIII, characterized as "severe" shaking,
[34]. This level of shaking can cause significant
structural damage and pose serious risks to
occupants and safety. If seismic activity exceeds a
PGA of 0.60 g, the structures will likely sustain
irreparable damage, rendering them unsafe for
occupancy. Additionally, the findings are in
Appendix in Table 12.A and Table 12.B align with
the pushover curve results shown in Figure 10,
Figure 11 and Figure 12, validating the directional
strengths of the buildings and confirming the
accuracy of the pushover analysis. When a structure
can withstand a higher PGA in the x-direction than
the y-direction, it signifies that the structural
integrity, reinforcements, and design along the x-
axis are more robust. This is often observed through
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.75
Ruth Ann D. Maningding, Jan Nicholas S. Baldo
E-ISSN: 2224-3496
800
Volume 20, 2024
higher base shear forces and greater peak ground
accelerations in that direction during analysis, [35].
Specifically, for Engineering Building 2, the base
shear force in the x-direction is 3402.297 kN,
significantly higher than the 2089.338 kN in the y-
direction, as shown in Figure 10. This disparity
indicates that Engineering Building 2 can withstand
a greater base shear force in the x-direction,
identifying it as the building's strongest axis. This
conclusion is further supported by consistently
higher peak ground acceleration values for all
damage states in the x-direction compared to the y-
axis, as shown in Appendix in Table 12.A and
Table 12.B. Similarly, for Engineering Buildings 1
and 3, the results show larger peak ground
accelerations in the y-direction than in the x-
direction, demonstrating that the y-direction is the
strong axis, confirming the results obtained from the
pushover curve as shown in Figure 11 and Figure
12.
4 Conclusion
Three buildings, namely Engineering Buildings 1
(BLDG4), Engineering Buildings 2 (BLDG5), and
Engineering Buildings 3 (BLDG6), with seismic
scores below the RVS FEMA P-154 cutoff of 2.0
underscore the importance of conducting further
detailed assessments that go beyond simple RVS
scores. These comprehensive evaluations are crucial
for precisely determining the level of vulnerability
of these buildings to seismic risks using fragility
analysis. By integrating comprehensive evaluation
data, such as detailed structural modeling, site-
specific ground motion characteristics, and
historical earthquake data, the fragility analysis
becomes more robust. This allows for a more
precise determination of the seismic performance of
the buildings, leading to more accurate predictions
of potential damage.
Furthermore, the analysis of the fragility curves
showed that Engineering Building 1, Engineering
Building 2, and Engineering Building 3, under the
condition of "complete damage (As)," demonstrated
a maximum probability of exceedance (Pr) of 5% at
0.40 g PGA. Because no value was more than 10%,
this result fulfills the minimum requirement set by
the National Structural Code of the Philippines
(NSCP) for structures in Seismic Zone 4.
Consequently, these buildings are declared safe for
occupancy and do not require retrofitting measures.
Moreover, the fragility curve analysis indicates
that Engineering Building 1, and Engineering
Building 2 have a low probability, below 10%, of
experiencing any damage from slight to complete
during seismic events. However, Engineering
Building 3 shows a higher likelihood of sustaining
slight damage, with a lower probability of
experiencing moderate to complete damage. This
suggests that while Engineering Building 3 may
suffer minor damage like hairline cracks and non-
structural component failures, it remains safe for
occupancy during most seismic events.
Finally, the results showed that Engineering
Building 1, Engineering Building 2, and
Engineering Building 3 would sustain "complete
damage (As)" at a maximum PGA of 0.60 g PGA.
Consequently, it was discovered that these buildings
can tolerate a maximum of 0.60 g, surpassing the
minimum standard of 0.4 g set by the National
Structural Code of the Philippines . According to the
Modified Mercalli Scale, a peak ground acceleration
(PGA) of 0.60g corresponds to Intensity VIII,
described as "severe" shaking, [36]. As peak ground
acceleration (PGA) increases, the probability of
exceeding each damage rank concurrently rises, as
shown in the resulting fragility curves. The fragility
curve indicates that the structure becomes highly
vulnerable during high PGA events. In the event of
an earthquake with a recorded PGA exceeding
0.60g, the structure has a high probability of
collapsing. Additionally, the analysis reveals that
Engineering Building 2 experienced higher peak
ground acceleration (PGA) along its y-axis. This can
be attributed to the y-axis being structurally
stronger, as confirmed by the results from the
pushover analysis. Therefore, the structural
components along the x-axis of Engineering
Building 2 were deemed more significant. Similarly,
Engineering Buildings 1 and 3 show larger peak
ground accelerations in the y-direction than in the x-
direction, demonstrating that the y-direction is the
strong axis.
Acknowledgement:
The authors would like to thank Pangasinan State
University (PSU) Urdaneta Campus for permitting
them to conduct the study.
Declaration of Generative AI and AI-assisted
Technologies in the Writing Process
During the preparation of this work the authors used
QuillBot in order to improve the readability and
language of the manuscript. After using this tool, the
authors reviewed and edited the content as needed
and take full responsibility for the content of the
publication.
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.75
Ruth Ann D. Maningding, Jan Nicholas S. Baldo
E-ISSN: 2224-3496
801
Volume 20, 2024
References:
[1] D. Paton, E. Anderson, J. Becker, and J.
Petersen, “Developing a comprehensive
model of hazard preparedness: lessons from
the Christchurch earthquake,” International
Journal of Disaster Risk Reduction, vol. 14,
no. 1, pp. 37-45, 2015. doi:
10.1016/j.ijdrr.2014.11.011.
[2] C. Baclig, “Dying in an earthquake: Experts
say it’s not just the tremor,” Inquirer.Net, July
28, 2022, [Online].
https://newsinfo.inquirer.net/1636640/dying-
in-an-earthquake-experts-say-its-not-just-the-
tremor (Accessed Date: July 27, 2024).
[3] M. Medina, “Devastating ‘faults’ of the ‘Big
One,’” Inquirer Business, Apr. 27, 2019,
[Online].
https://business.inquirer.net/269314/devastatin
g-faults-of-the-big-one/amp (Accessed Date:
July 27, 2024).
[4] A. W. Oreta, R. De Jesus, M. E. Grio, R.
Mendoza Jr., L. Garciano, R. Ibabao, G.
Sevieri, and C. Galasso, "Seismic screening
and structural investigation of heritage
buildings for adaptive reuse: a survey study at
Iloilo City, Philippines," The 9th AUN/SEED-
Net Regional Conference on Natural Disaster
(RCND 2021), IOP Conf. Series: Earth and
Environmental Science, Canada, 2022, pp.1-
10. doi: 10.1088/1755-1315/1091/1/012008.
[5] R.V. Garrote, R.C. Ilumin, “Comparative
Analysis on Seismic Provisions of the
National Structural Codes of the Philippines
(NSCP) 1992 and 2010 as applied to the
Design of Reinforced Concrete Public School
Building,” Proc. Of Sixth International
Conference on Advanced in Civil Structural
and Mechanical Engineering, Thailand, 2017,
pp. 62-67, [Online].
https://ewr1.vultrobjects.com/ired/papers/352
2/2ee800e32c0981d9f9919a05db3534ea.pdf
(Accessed Date: July 27, 2024).
[6] R.C. Ilumin, A.W. Oreta, “A Post-Disaster
Functional Asset Value Index for school
buildings,” Procedia Engineering, Bangkok,
Thailand, Jan. 2018, pp. 230–237. doi:
10.1016/j.proeng.2018.01.030.
[7] M. A. Llego, “2019 DePED guidelines on
schools used as evacuation center,”
TeacherPH, Mar. 15, 2019, [Online].
https://www.teacherph.com/guidelines-
schools-evacuation-center/ (Accessed Date:
July 27, 2024).
[8] R. Gentile, C. Galasso, Y. Idris, I. Rusydy, E.
Meilianda, “From rapid visual survey to
multi-hazard risk prioritisation and numerical
fragility of school buildings,” Natural
Hazards and Earth System Sciences, vol. 19,
no. 7, pp. 1365–1386, Jul. 2019. doi:
10.5194/nhess-19-1365-2019.
[9] R. De Guzman, “DOST’S Fault Finder”,
Department of Science and Technology, Oct.
05, 2024, [Online]
https://www.dost.gov.ph/knowledge-
resources/news/48-2017-news/1223-dost-s-
fault-finder.html (Accessed Date: July
27,2024)
[10] M.B. Baylon, “Seismic vulnerability
assessment of Adamson University buildings
As-built using fragility curves,” Global
Journal of Research in Engineering, vol.18,
no.1, pp. 29-49, 2018, [Online]
https://engineeringresearch.org/index.php/GJ
RE/article/view/1785 (Accessed Date: July
27, 2024)
[11] S. U. Khan, M.I. Qureshi, I. A. Rana, A.
Maqsoom, “Seismic vulnerability assessment
of building stock of Malakand (Pakistan)
using FEMA P-154 method,” SN Applied
Sciences/SN Applied Sciences, vol. 1, no. 12,
pp. 1-14, Nov. 2019. doi: 10.1007/s42452-
019-1681-z.
[12] J. Dapito, D. Barrozo, R. Panis, A. Cañete, R.
Orbeso, K. Yagi, R. Salazar, “Two-
Dimensional Analysis Vulnerability
Assessment of Public Secondary Buildings in
District II of Manila,”8th IEEE International
Conference Humanoid, Nanotechnology,
Information Technology Communication and
Control, Environment and Management
(HNICEM), Cebu, Philippines, Dec. 2017,
pp.1-7. doi: 10.1109/HNICEM.2015.7393217
[13] I.V.P. Perez, L.J. David, S.P. Rodrigo, C.J.
Sibug, G.H. Arcilla, C.B. Nava, L. Manuntag,
C.J. Coronel, “Assessment of the seismic
vulnerability of San Guillermo parish church
and San Agustin parish church in Pampanga,”
International Journal of Advance Research,
Ideas and Innovations in Technology, vol. 9,
no. 2, pp. 487-515, June 2023, [Online]
https://www.ijariit.com/manuscripts/v9i2/V9I
2-1340.pdf (Accessed Date: July 27, 2024)
[14] S. A. El-Betar, “Seismic vulnerability
evaluation of existing R.C. buildings,” HBRC
Journal/HBRC Journal, vol. 14, no. 2, pp.
189–197, Aug. 2018. doi:
10.1016/j.hbrcj.2016.09.002.
[15] S.F. Shah, A. Ahmed, O.K. Kegyes-B, “A
case study using rapid visual screening
method to determine the vulnerability of
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.75
Ruth Ann D. Maningding, Jan Nicholas S. Baldo
E-ISSN: 2224-3496
802
Volume 20, 2024
buildings in two districts of Jeddah, Saudi
Arabia,” In Proceedings of the 15th
International Symposium on new technologies
for urban safety of megacities in Asia,
Tacloban, Philippines, Nov. 2016, pp. 7-9,
[Online].
https://www.researchgate.net/publication/309
736872_A_Case_Study_Using_Rapid_Visual
_Screening_Method_to_Determine_the_Vuln
erability_of_Buildings_in_two_Districts_of_J
eddah_Saudi_Arabia (Accessed Date: July 27,
2024).
[16] O. Kaplan, Y. Güney, A. Topçu, Y.
Özçelikörs, “A rapid seismic safety
assessment method for mid-rise reinforced
concrete buildings,” Bulletin of Earthquake
Engineering, vol. 16, no. 2, pp. 889–915, Sep.
2017. doi: 10.1007/s10518-017-0229-0.
[17] E. Harirchian, T. Lahmer, S. Buddhiraju, K.
Mohammad, A. Mosavi A, “Earthquake
Safety Assessment of Buildings through
Rapid Visual Screening,” Buildings, vol. 10,
no. 3, pp. 51, Mar. 2020. doi:
10.3390/buildings10030051.
[18] A. Aydin, A. Basu, “The Schmidt hammer in
rock material characterization, Engineering
Geology, vol. 81, no. 1, pp. 1–14, Sep. 2005.
https://doi.org/10.1016/j.enggeo.2005.06.006.
[19] R. Surva, R. Shobana, “Design & Seismic
Analysis of Single Storey 3-D Frame Using
Sap2000, International Research Journal of
Engineering and Technology (IRJET), vol. 18,
no. 4, pp. 5582-5087, April 2021. [Online].
https://www.irjet.net/archives/V8/i4/IRJET-
V8I4973.pdf (Accessed Date: July 27, 2024).
[20] M. Zameeruddin, K. Sangle, “Performance-
based Seismic Assessment of Reinforced
Concrete Moment Resisting Frame,” Journal
of King Saud University-Engineering
Sciences, vol 33, no.3, pp. 153-165,
Mar.2021.
https://doi.org/10.1016/j.jksues.2020.04.005.
[21] S. Freeman, “Review of the development of
the Capacity Spectrum Method,” ISET
Journal of Earthquake Technology, vol. 41,
no. 1, pp. 1–13, Mar. 2004, [Online].
https://engineering.purdue.edu/~ce573/Docum
ents/SAF_Review%20of%20Capacity%20Spe
ctrum%20Method.pdf (Accessed Date: July
27, 2024).
[22] I.F. Moschonas, A. J. Kappos, P. Panetsos, V.
Papadopoulos, T. K. Makarios, P.
Thanopoulos, “Seismic fragility curves for
Greek bridges: methodology and case
studies,” Bulletin of Earthquake Engineering,
vol. 7, no. 2, pp. 439–468, Sep. 2008. doi:
10.1007/s10518-008-9077-2.
[23] K. R. Karim, F. Yamazaki, “A simplified
method of constructing fragility curves for
highway bridges,” Earthquake Engineering
and Structural Dynamics/Earthquake
Engineering & Structural Dynamics, vol. 32,
no. 10, pp. 1603–1626, May 2003. doi:
10.1002/eqe.291.
[24] M. Shinozuka, M. Feng, J. Lee, T.
Naganuma, “Statistical Analysis of Fragility
Curves,” Journal of Engineering Mechanics,
vol. 126, no. 12, pp. 1224–1231, Dec. 2000.
doi: 10.1061/(ASCE)0733-
9399(2000)126:12(1224).
[25] M.B. Baylon, C. S. Lictaoa, I.J.C. Castillo,
C.K.P. Dungog, L.R. Jerusalem, “Seismic
vulnerability assessment of Sta. Lucia high
school in Pasig City Philippines using rapid
visual assessment and fragility curves,”
Zenodo (CERN European Organization for
Nuclear Research), vol. 1, no. 2, pp. 119–135,
Nov. 2021. doi: 10.5281/zenodo.5726470.
[26] U.Y. Vazurkar, D.J. Chaudhari,
“Development of Fragility Curves for RC
Buildings,” International Journal of
Engineering Research, vol. 5, no. 3, pp. 591-
594, Feb. 2016, [Online].
https://www.academia.edu/27583791/Develop
ment_of_Fragility_Curves_for_RC_Buildings
(Accessed Date: July 27,2024)
[27] M. Vasavada, V.R. Patel, “Development of
Fragility Curves for RC Buildings using
HAZUS method,” International Research
Journal of Engineering and Technology
(IRJET), vol. 3, no. 5, pp. 2256–2262, May
2016, [Online].
https://www.irjet.net/archives/V3/i5/IRJET-
V3I5479.pdf (Accessed Date: July 27, 2024).
[28] T. Gehlot, S. S. Sankhla, A. Gupta, “Study of
Concrete Quality Assessment of Structural
Elements Using Rebound Hammer Test,”
American Journal of Engineering Research
(AJER), Vol.5, no.8, pp.192- 198, 2016,
[Online].
https://www.ajer.org/papers/v5(08)/Z0508019
20198.pdf (Accessed Date: July 27, 2024).
[29] A. J. Pierre, “Seismic performance of
reinforced concrete structures with pushover
analysis,” IOP Conference Series: Earth and
Environmental Science, Indonesia, 2020, pp.
1-10. doi: 10.1088/1755-1315/426/1/012045.
[30] Y. Menasri, M. S. Nouaouria, M. Brahimi,
"Probabilistic Approach to the Seismic
Vulnerability of RC Frame Structures by the
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.75
Ruth Ann D. Maningding, Jan Nicholas S. Baldo
E-ISSN: 2224-3496
803
Volume 20, 2024
Development of Analytical Fragility Curves",
International Journal of Engineering, vol. 30,
no. 7, pp. 945-954, July 2017, [Online].
https://www.ije.ir/article_72967.html
(Accessed Date: July 27, 2024).
[31] A. Balingit, N. Almario, B.J. Antonio, J. Bon,
J. Cada, T. J. Soriano, J.R. Gabriel, J.V.
Tongol, “Seismic Assessment of Phil Health
Regional Office III Building in the City of
San Fernando, Pampanga, Philippines Using
Fragility Curves,” International Journal of
Scientific Research and Engineering
Development, vol. 7, no.3, pp.1446-1473,
June 2024, [Online].
https://ijsred.com/volume7/issue3/IJSRED-
V7I3P155.pdf (Accessed Date: July 27, 2024)
[32] X. Wang, C.H. Wang, Y.B. Khoo, C. Morga,
M.G. Stewart, “Designing resilient structures:
Mainstreaming disaster risk reduction and
climate change adaptation in local design
practices handbook,” Commonwealth
Scientific and Industrial Research
Organization (CSIRO), Australia, 2016,
[Online].
https://philippines.embassy.gov.au/files/mnla/
Designing%20Resilient%20Structures%20Ha
ndbook%202016%20(12).pdf (Accessed
Date: July 27, 2024).
[33] M. J. Gemoya, R. R. Ordanez, T. Paulino, J.
Razon, A. Santos, J.M. Songco, M. Tongol,
J.C. Zoleta, J.R. Gabriel, M. B. Baylon,
“Seismic Vulnerability Assessment of College
of Engineering and Architecture (CEA)
Extension Building in Don Honorio Ventura
State University (DHVSU) Main Campus
Using Fragility Curves”, International
Journal of Progressive Research in Science
and Engineering, vol. 4, no. 6, pp. 279-286,
2023, [Online].
https://journal.ijprse.com/index.php/ijprse/arti
cle/view/917 (Accessed Date: July 27, 2024)
[34] S. A. Freeman, A. Irfanoglu, & T. F. Paret,
“Earthquake engineering intensity scale: A
template with many uses,” In 13th World
Conference on Earthquake Engineering,
Vancouver, BC, Canada, Aug. 2004, pp. 1-15,
[Online].
https://www.iitk.ac.in/nicee/wcee/article/13_1
667.pdf (Accessed Date: July 27, 2024).
[35] M.B. Baylon, H. D. Caranto, A. P. Espino, B.
D. Nilo, “Developing Seismic Fragility
Curves of Old Hospital in Manila,
Philippines,” In 3rd International Conference
on Protection of Historical Constructions.
Lisbon, Portugal, 2017, pp. 1-13, [Online].
https://www.researchgate.net/publication/319
187969_DEVELOPING_SEISMIC_FRAGIL
ITY_CURVES_OF_OLD_HOSPITAL_IN_
MANILA_PHILIPPINES (Accessed Date:
July 24, 2024)
[36] T. Zera, M. Nafian, “Comparing Two models
of mapping the peak ground acceleration
(PGA) in western Java,” International
Conference on Science and Technology
(ICOSAT 2017), Jakarta, 2017, pp.140-142.
DOI: 10.2991/icosat-17.2018.30.
Contribution of Individual Authors to the
Creation of a Scientific Article (Ghostwriting
Policy)
All authors are equal contributors to this work.
Sources of Funding for Research Presented in a
Scientific Article or Scientific Article Itself
No funding was received for conducting this study.
Conflict of Interest
The authors declare that there are no conflicts of
interest.
Creative Commons Attribution License 4.0
(Attribution 4.0 International, CC BY 4.0)
This article is published under the terms of the
Creative Commons Attribution License 4.0
https://creativecommons.org/licenses/by/4.0/deed.en
_US
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.75
Ruth Ann D. Maningding, Jan Nicholas S. Baldo
E-ISSN: 2224-3496
804
Volume 20, 2024
APPENDIX
Table 1. Local Ground Motion Data
Location
Magnitude
Latitude
Longitude
Depth
Date
Luzon, Philippines
Mw 7.2
13.45
124.63
18.02 km
24/02/1988
Luzon, Philippines
Mw 6.5
10.81
126.83
44.4 km
17/07/1990
Mindanao, Philippines
Mw 6.5
8.33
126.52
63.2 km
25/10/1990
Samar, Philippines
Mw 7.1
12.07
125.28
20.7 km
21/04/1995
Samar, Philippines
Mw 7.0
12.63
125.58
5.3 km
05/05/1995
Philippine Islands Region
Mw 7.6
10.81
126.83
44.4 km
31/08/2012
Luzon, Philippines
Mw 7.0
17.55
120.80
46.0 km
27/07/2022
Table 2. Foreign Ground Data Motion
Location
Magnitude
Latitude
Longitude
Depth
Date
Kuril Islands
Mw 7.1
45.26
149.89
36.6 km
07/02/1996
Near the Coast of Peru
Mw 8.0
-13.38
-76.56
41.2 km
15/08/2007
Near East Coast of Japan
Mw 9.1
38.29
142.50
19.7 km
11/03/2011
Near the Coast of Central Chile
Mw 8.3
-31.57
-71.67
22.4 km
16/09/2015
Central California
Mw 7.1
35.77
-117.60
8.0 km
06/07/2019
Table 5. Summary of Compressive Strength Used for Each Building
Location of
Test Area
Part of
Structure
Rebound Reading Number
Average
Rebound
Reading
Correction
Factor
Corrected
Compressive
Strength
(MPa)
A
B
C
D
E
F
G
H
I
J
Engineering
Building 2
Column
1
38.30
37.10
34.00
34.50
36.20
37.50
36.60
41.20
37.40
35.50
36.83
1
17.80
Beam 1
38.10
39.30
31.30
32.70
37.10
32.80
31.20
35.50
37.30
37.80
35.31
1
16.00
Engineering
Building 1
Column
2
34.60
44.10
36.80
39.20
34.90
40.00
39.90
37.00
36.70
40.50
38.37
1
19.10
Beam 2
39.40
43.20
44.40
49.10
42.00
42.90
39.00
40.70
37.10
46.80
42.46
1
21.80
Engineering
Building 3
Column
3
37.20
39.60
41.10
43.90
39.60
42.50
37.50
41.80
40.40
43.10
40.67
1
21.60
Beam 3
41.20
42.80
3480
42.40
37.30
39.60
41.20
40.70
44.20
40.23
40.23
1
21.30
Table 6. Summary of the tensile strength used for each building.
Test
Engineering Building 1
Engineering Building 2
Engineering Building 3
Mechanical Properties
Yield point (psi)
53.650
49,300
47,850
Tensile Strength (psi)
72,500
72,500
66,700
Elongation, %
20
20
20
TS/YS Ratio
1.35
1.47
1.39
Description of fracture
IRR
IRR
IRR
Bending Properties
Degree bent, 180 degrees
No crack
No crack
No crack
Physical Properties
Actual Unit Mass, kg/m
1.506
1.522
1.555
Variation in Mass, %
-2.96
-1.93
0.19
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.75
Ruth Ann D. Maningding, Jan Nicholas S. Baldo
E-ISSN: 2224-3496
805
Volume 20, 2024
Table 7. Performance Points summary tabulation based on the July 16, 1990, earthquake in Luzon, Philippines
(6.5)
Table 8. Damage State Threshold Limits of each School Building
Building Name
X Direction
Y Direction
Spectral Displacement (mm)
Spectral Displacement (mm)
D
C
B
A
As
D
C
B
A
As
Engineering Building 1
Engineering Building 2
Engineering Building 3
0
0
0
12.892
14.055
33.162
18.417
20.079
47.347
34.585
58.428
103.765
83.088
173.476
272.939
0
0
0
11.754
18.506
10.348
16.792
26.437
14.783
33.344
85.646
79.024
83.001
263.273
271.748
PGA
East-West Direction
North-South Direction
X-Axis
Y-Axis
X-Axis
Y-Axis
V(KN)
D(mm)
V(KN)
D(mm)
V(KN)
D(mm)
V(KN)
D(mm)
0.1
161.629
1.013
145.137
1.238
86.033
0.560
116.230
0.846
0.2
323.265
1.981
290.276
3.203
172.059
1.075
232.456
2.420
0.3
484.893
2.948
435.413
5.168
258.092
1.590
348.686
3.994
0.4
646.522
3.916
580.549
7.133
344.124
2.106
464.910
5.567
0.5
808.158
4.884
725.688
9.097
430.157
2.621
581.142
7.141
0.6
969.787
5.852
870.825
11.062
516.183
3.136
697.366
8.714
0.7
1131.422
6.820
1015.964
13.027
602.216
3.651
813.596
10.288
0.8
1285.740
7.770
1161.101
14.992
688.249
4.166
929.820
11.861
0.9
1422.885
8.678
1306.238
16.957
774.275
4.681
1046.052
13.435
1
1559.975
9.585
1439.106
18.833
860.307
5.196
1162.286
15.008
1.1
1696.997
10.492
1538.897
20.454
946.340
5.712
1278.506
16.582
1.2
1833.959
11.399
1636.117
22.033
1032.373
6.227
1394.732
18.155
1.3
1970.866
12.305
1703.246
23.553
1118.399
6.742
1487.995
19.627
1.4
2107.706
13.211
1750.163
25.009
1204.432
7.257
1577.081
21.074
1.5
2244.484
14.117
1795.410
26.413
1285.415
7.768
1665.592
22.512
1.6
2381.208
15.022
1822.421
28.557
1362.316
8.277
1719.107
24.045
1.7
2517.866
15.927
1837.381
31.196
1439.689
8.789
1768.599
25.581
1.8
2654.468
16.831
1852.200
33.810
1517.549
9.304
1806.816
27.254
1.9
2791.004
17.735
1862.637
36.569
1595.897
9.823
1826.522
29.280
2
2914.284
18.581
1873.854
39.112
1674.729
10.345
1839.610
31.589
2.1
2954.486
19.063
1881.266
41.991
1754.063
10.870
1852.621
33.884
2.2
2993.651
19.533
1885.446
45.888
1833.898
11.399
1862.885
36.274
2.3
3027.965
20.153
1887.717
50.417
1914.238
11.931
1871.386
38.603
2.4
3048.454
21.432
1896.816
53.621
1995.082
12.466
1878.564
41.169
2.5
3068.735
22.698
1909.032
56.313
2076.447
13.004
1885.539
46.073
2.6
161.629
1.013
145.137
1.238
2158.331
13.547
1925.582
60.114
2.7
323.265
1.981
290.276
3.203
2240.735
14.092
1927.960
63.099
2.8
484.893
2.948
435.413
5.168
2323.673
14.641
1930.357
66.108
2.9
646.522
3.916
580.549
7.133
2407.149
15.194
1932.774
69.141
3.0
808.158
4.884
725.688
9.097
2491.162
15.750
1935.209
72.199
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.75
Ruth Ann D. Maningding, Jan Nicholas S. Baldo
E-ISSN: 2224-3496
806
Volume 20, 2024
Table 9. Sample Performance Points and Damage Ranking for Engineering Building 2 in
East-West Direction (y-direction)
PGA
D (mm)
DST
PGA
D (mm)
DST
0.1
3.773
D
1.6
99.976
A
0.2
8.273
D
1.7
103.798
A
0.3
12.773
D
1.8
107.318
A
0.4
17.273
D
1.9
110.699
A
0.5
21.318
C
2
271.975
As
0.6
25.133
C
2.1
289.502
As
0.7
30.173
B
2.2
302.407
As
0.8
36.878
B
2.3
310.541
As
0.9
44.605
B
2.4
318.423
As
1.0
67.657
B
2.5
326.042
As
1.1
67.657
B
2.6
332.870
As
1.2
81.329
B
2.7
339.473
As
1.3
89.378
A
2.8
345.862
As
1.4
93.742
A
2.9
354.010
As
1.5
96.798
A
3
370.062
As
Table 10. Sample Total Number of Occurrence in the East-West
Direction (y-axis) for Engineering Building 2
PGA (g)
Y - Direction
PGA (g)
Y - Direction
D
C
B
A
As
D
C
B
A
As
0.1
12
0
0
0
0
1.6
2
3
3
4
0
0.2
12
0
0
0
0
1.7
2
2
4
4
0
0.3
12
0
0
0
0
1.8
2
2
4
4
0
0.4
12
0
0
0
0
1.9
0
2
6
4
0
0.5
10
2
0
0
0
2.0
0
2
6
3
1
0.6
10
2
0
0
0
2.1
0
2
5
4
1
0.7
8
2
2
0
0
2.2
0
2
3
6
1
0.8
8
2
2
0
0
2.3
0
2
3
5
2
0.9
8
2
2
0
0
2.4
0
2
3
5
2
1.0
5
4
3
0
0
2.5
0
2
3
4
3
1.1
4
4
4
0
0
2.6
0
2
2
4
4
1.2
4
4
4
0
0
2.7
0
1
3
4
4
1.3
3
5
3
1
0
2.8
0
0
4
3
5
1.4
2
5
4
1
0
2.9
0
0
4
3
5
1.5
2
4
4
2
0
3.0
0
0
3
4
5
Table 11.A: Summary of Probability of Exceedance for PSU Buildings at 0.4g Peak Ground Acceleration in the
East-West Direction for Each Damage Rank
Building Name
East West
X Axis
Y - Axis
C
B
A
As
C
B
A
As
Engineering Building 1
Engineering Building 2
Engineering Building 3
6.53%
5.49%
21.98%
5.46%
5.20%
8.07%
5.25%
5.22%
5.31%
5.23%
5.47%
5.24%
5.32%
7.50%
21.07%
5.32%
5.61%
6.84%
5.19%
5.20%
5.19%
5.44%
5.29%
5.39%
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.75
Ruth Ann D. Maningding, Jan Nicholas S. Baldo
E-ISSN: 2224-3496
807
Volume 20, 2024
Table 11.B: Summary Of Probability of Exceedance for PSU Buildings at 0.4g peak ground acceleration in the
North-South Direction for every damage rank.
Building Name
North - South
X Axis
Y - Axis
C
B
A
As
C
B
A
As
Engineering Building 1
Engineering Building 2
Engineering Building 3
6.74%
5.55%
18.18%
5.38%
5.19%
6.54%
5.19%
5.21%
5.22%
5.33%
5.39%
5.24%
5.31%
6.26%
17.62%
5.25%
5.34%
6.82%
5.23%
5.34%
5.19%
0%
5.27%
5.33%
Table 12.A: Maximum Peak Ground Acceleration in the East-West Direction with a 10% Probability of
Exceedance
Building Name
East West
X Axis
Y - Axis
C
B
A
As
C
B
A
As
Engineering Building 1
Engineering Building 2
Engineering Building 3
0.50 g
0.55 g
0.25 g
0.53 g
0.58 g
0.43 g
0.55 g
0.58 g
056 g
0.60 g
0.60 g
0.60 g
0.55 g
0.46 g
0.25 g
0.56 g
0.53 g
0.48 g
0.59 g
0.57 g
0.58 g
0.60 g
0.60 g
0.60 g
Table 12.B: Maximum Peak Ground Acceleration in the North-South Direction with a 10% Probability of
Exceedance
Building Name
North - South
X – Axis
Y - Axis
C
B
A
As
C
B
A
As
Engineering Building 1
Engineering Building 2
Engineering Building 3
0.49 g
0.55 g
0.29 g
0.54 g
0.59 g
0.50 g
0.58 g
0.60 g
0.58 g
0.60 g
0.60 g
0.58 g
0.56 g
0.50 g
0.29 g
0.57 g
0.56 g
0.48 g
0.58 g
0.59 g
0.58 g
0.60 g
0.60 g
Fig. 20: Seismic Fragility Curves of Engineering Building 2
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.75
Ruth Ann D. Maningding, Jan Nicholas S. Baldo
E-ISSN: 2224-3496
808
Volume 20, 2024
Fig. 21: Seismic Fragility Curves of Engineering Building 1
Fig. 22: Seismic Fragility Curves of Engineering Building 3
WSEAS TRANSACTIONS on ENVIRONMENT and DEVELOPMENT
DOI: 10.37394/232015.2024.20.75
Ruth Ann D. Maningding, Jan Nicholas S. Baldo
E-ISSN: 2224-3496
809
Volume 20, 2024