Keynote Paper

PBEE in a Moderate-Seismicity Region: South Korea

*Han Seon Lee1)

1)
School of Civil, Environmental and Architectural Engineering, Korea University,
Seoul 02841, Korea
1)
hslee@korea.ac.kr

ABSTRACT

Performance-Based Earthquake Engineering (PBEE) has been developed mainly

for the region of high seismicity for the last three decades. Though abundant
information on PBEE is available throughout the world, the application of this PBEE to
the moderate-seismicity regions such as their maximum considered earthquake being
less than magnitude 6.5 is not always straightforward because some portion of the
PBEE may not be appropriate in these regions due to the environment different from
the high-seismicity regions. This paper reviews the state-of-art in PBEE briefly. Then,
the seismic hazard in moderate-seismicity regions including Korean Peninsula is
introduced with its unique characteristics. With this seismic hazard, representative low-
rise RC MRF structures and high-rise RC residential wall structures are evaluated by
using PBEE approach. Also, the range of forces and deformations of the representative
building structures in Korea is given. Based on these reviews, some ideas for the use of
PBEE to improve the state-of-practice in moderate-seismicity regions are proposed.

1. INTRODUCTION

The conventional engineering design in many fields has been conducted by

satisfying all the requirements in the codes corresponding to the field of engineering.
We call this design procedure the design to the prescriptive codes.
Prescriptive earthquake design codes currently in use are based on the traditional
design philosophy preventing structural and nonstructural elements of buildings from
any damage in low-intensity earthquakes, limiting the damage in these elements to
reparable levels in medium-intensity earthquakes, and preventing the overall or partial
collapses of buildings in high-intensity earthquakes. After the 1994 Northridge and
1995 Kobe earthquakes, the structural engineering community realized that the almost
of damage, economic loss due to downtime, and repair cost of the structure was
unacceptably high even though those structures complied with available seismic codes
based on the above traditional philosophy.
1)
Professor

1
 This realization led to the development of the concept of the first-generation
performance-based earthquake engineering (PBEE) through Vision 2000 report
(SEAOC 1995) in the US where the performance-based earthquake design is defined
as a design framework by designating the desired system performance at various
intensity levels of seismic hazard. The designer and owner consult to select the desired
combination of performance and hazard levels to use as design criteria. In subsequent
documents of the first-generation PBEE such as ATC 40 (1996), FEMA 273(1996),
FEMA 356(2000), and ASCE 41-13(2013), the element deformation and force
acceptability criteria corresponding to the performance are specified for different
structural and non-structural elements for linear, nonlinear, static, and dynamic analysis.
These criteria do not possess probability distributions on the both sides of demand and
supply. Also, the element performance evaluation is not tied to the global performance.
Considering the shortcomings of the first-generation procedures that are incapable of
probabilistic calculation of system performance measures, such as monetary losses,
downtime, and causalities, which are expressed regarding the direct interest of various
stakeholders, the second generation PBEE has been developed by Pacific Earthquake
Engineering Research Center (PEER) in the US. The key feature of the methodology is
the calculation of performance in a rigorous probabilistic manner. Accordingly,
uncertainty in earthquake intensity, ground motion characteristics, structural response,
physical damage, and economic and human losses are explicitly considered in this
approach.
The second-generation PBEE (such as PEER PBEE) methodology consists of four
successive analyses: hazard, structural, damage, and loss. However, those analyses
have been performed only for strong-seismicity regions such as California in the US.
This paper reviews the state-of-art in PBEE briefly. Then, the seismic hazard in
moderate-seismicity regions including Korean Peninsula is introduced with its unique
characteristics. With this seismic hazard, representative low-rise RC MRF structures
and high-rise RC residential wall structures are evaluated by using PBEE approach.
Also, the range of forces and deformations of the representative building structures in
Korea is given. Based on these reviews, some ideas for the use of PBEE to improve
the state-of-practice in moderate-seismicity regions are proposed.

2. HISTORY OF PBEE

2.1 Introduction (excerpted from Porter 2003)

PBEE implies design, evaluation, construction, monitoring the function and loads
responds to the diverse needs and objectives of owners-users and society. It is based
on the premise that performance can be predicted and evaluated with quantifiable
confidence to make, together with the client, intelligent and informed trade-offs based
on life-cycle considerations rather than construction costs alone. (Krawinkler and
Miranda 2004)
Performance-based earthquake engineering (PBEE) in one form or another may
supersede load-and-resistance-factor design (LRFD) as the framework under which
many new and existing structures are analyzed for seismic adequacy. A key distinction
between the two approaches is that LRFD seeks to assure performance primarily in
terms of failure probability of individual structural components (with some system

2
aspects considered, such as the strong-column-weak-beam requirement), whereas
PBEE attempts to address performances primarily at the system level in terms of risk of
collapse, fatalities, repair costs, and post-earthquake loss of function.
Initial efforts to frame and standardize PBEE methodologies produced SEAOC’s
Vision 2000 report (1995) and FEMA 273 (1997), a product of the ATC-33 project. The
authors of these documents frame PBEE as a methodology to assure combinations of
desired system performance at various levels of seismic excitation. The system
performance states of Vision 2000 include fully operational, operational, life safety, and
near collapse. Levels of excitation include frequent (43- year return period), occasional
(72-year), rare (475-year) and very rare (949-year) events. These reflect Poisson-
arrival events with 50% exceedance probability in 30 years, 50% in 50 years, 10% in 50
years, and 10% in 100 years, respectively. The designer and owner consult to select an
appropriate combination of performance and excitation levels to use as design criteria,
such as those suggested in Fig. 1.
FEMA 273 expresses design objectives using a similar framework, although with
slightly different performance descriptions and levels of seismic excitation. Each global
performance level is detailed regarding the performance of individual elements. The
design is believed to satisfy its global objectives if the structural analysis indicates that
the member forces or deformations imposed on each element do not exceed
predefined limits. Performance is binary and largely deterministic: if the member force
or deformation does not exceed the limit, it passes; otherwise, it fails. If the acceptance
criteria are met, the design is believed to assure the performance objective, although
without a quantified probability. Other important pioneering PBEE efforts include ATC-
32 (1996a), ATC-40 (1996b), and FEMA 356 (2000).

Fig. 1 Vision 2000 recommended seismic Fig. 2 Visualization of PBEE

performance objectives for buildings (Moehle and Deierlein 2004)
(SEAOC 1995)

3
 2.2 PEER PBEE (excerpted from Moehle and Deierlein 2004)
Performance-based earthquake engineering seeks to improve seismic risk
decision-making through assessment and design methods that have a strong scientific
basis and that express options in terms that enable stakeholders to make informed
decisions. A key feature is the definition of performance metrics that are relevant to
decision making for seismic risk mitigation. The methodology needs to be underpinned
by a consistent procedure that characterizes the important seismic hazard and
engineering aspects of the problem, and that relates these quantitatively to the defined
performance metrics. The first generation of performance-based earthquake
engineering assessment and design procedures for buildings in the United States
made important steps toward the realization of performance-based earthquake
engineering. These procedures conceptualized the problem as shown in Fig. 2. Here,
the building is visualized as being loaded by earthquake-induced lateral forces that
result in nonlinear response and resulting damage. Relations are then established
between structural response indices (interstory drifts, inelastic member deformations,
and member forces) and performance-oriented descriptions such as Immediate
Occupancy, Life Safety, and Collapse Prevention. Without minimizing the remarkable
accomplishments of these first-generation procedures, several shortcomings can be
identified:

• Engineering demands are based on simplified analysis techniques, including static

and linear analysis methods; where dynamic or nonlinear methods are used,
calibrations between calculated demands and component performance are largely
lacking.
• The defined relations between engineering demands and component performance
are based somewhat inconsistently on relations measured in laboratory tests,
calculated by analytical models, or assumed on the basis of engineering judgment;
consistent approaches based on relevant data are needed to produce reliable
outcomes.
• Structural performance is defined by component performance states, where the
overall system performance is assumed to be equal to the worst performance
calculated for any component in the building.

Although the developers widely recognized the shortcomings of the first-generation

procedures, limitations in available technologies and supporting research did not permit
further development at that time. Since then, the Pacific Earthquake Engineering
Research Center (PEER) has embarked on a research program aimed at developing a
more robust methodology for performance-based earthquake engineering. Recognizing
the complex, multi-disciplinary nature of the problem, PEER has broken the process
into logical elements that can be studied and resolved in a rigorous and consistent
manner. The process begins with a definition of a ground motion Intensity Measure (IM),
which defines in a probabilistic sense the salient features of the ground motion hazard
that affect structural response. The next step is to determine Engineering Demand
Parameters, which describe structural response regarding deformations, accelerations,
or other response quantities calculated by simulation of the building to the input ground
motions. Engineering Demand Parameters are next related to Damage Measures,
4
which describe the condition of the structure and its components. Finally, given a
detailed probabilistic description of damage, the process culminates with calculations of
Decision Variables, which translate the damage into quantities that enter into risk
management decisions. Consistent with current understanding of the needs of
decision-makers, the decision variables have been defined in terms of quantities such
as repair costs, downtime, and casualty rates (Fig. 2). Underlying the methodology is a
consistent framework for representing the inherent uncertainties in earthquake
performance assessment.
While full realization of the methodology in professional practice is still years away,
important advances are being made through research in PEER. Some specific
highlights are presented in the following text.
Given the inherent uncertainty and variability in seismic response, it follows that a
performance-based methodology should be formalized within a probabilistic basis.
Referring to Fig. 3, PEER’s probabilistic assessment framework is described in terms of
four main analysis steps (hazard analysis, structural/nonstructural analysis, damage
analysis, and loss analysis). The outcome of each step is mathematically characterized
by one of four generalized variables: Intensity Measure (IM), Engineering Demand
Parameter (EDP), Damage Measure (DM), and Decision Variable (DV). Recognizing the
inherent uncertainties involved, these variables are expressed in a probabilistic sense as
conditional probabilities of exceedance, i.e., p[A|B]. Underlying the approach in Fig. 3 is
the assumption that the performance assessment components can be treated as a
discrete Markov process, where the conditional probabilities between parameters are
independent.

Fig. 3 Underlying probabilistic framework (Moehle and Deierlein 2004)

The first assessment step entails a hazard analysis, through which one evaluates
one or more ground motion Intensity Measures (IM). For standard earthquake intensity
measures (such as peak ground acceleration or spectral acceleration) IM is obtained
through conventional probabilistic seismic hazard analyses. Typically, IM is described
as a mean annual probability of exceedance, p[IM], which is specific to the location (O)
and design characteristics (D) of the facility. The design characteristics might be
described by the fundamental period of vibration, foundation type, simulation models,
etc. In addition to determining IM, the hazard analysis involves characterization of
appropriate ground motion input records for response history analyses.
Given IM and input ground motions, the next step is to perform structural simulations
to calculate Engineering Demand Parameters (EDP), which characterize the response in
terms of deformations, accelerations, induced forces, or other appropriate quantities. For
5
buildings, the most common EDPs are interstory drift ratios, inelastic component
deformations and strains, and floor acceleration spectra. Relationships between EDP
and IM are typically obtained through inelastic simulations, which rely on models and
simulation tools in areas of structural engineering, geotechnical engineering, SSFI (soil-
structure-foundation-interaction), and non-structural component and system response.
The next step in the process is to perform a damage analysis, which relates the
EDPs to Damage Measures, DM, which in turn describes the physical damage to a
facility. The DMs include descriptions of damage to structural elements, non-structural
elements, and contents, in order to quantify the necessary repairs along with functional
or life safety implications of the damage (e.g., falling hazards, the release of hazardous
substances, etc.). These conditional probability relationships, p(DM|EDP), can then be
integrated with the EDP probability, p(EDP), to give the mean annual probability of
exceedance for the DM, i.e., p(DM).
The final step in the assessment is to calculate Decision Variables, DV, in terms
that are meaningful for decision makers. Generally speaking, the DVs relate to one of
the three decision metrics discussed above with regard to Fig. 2, i.e., direct dollar
losses, downtime (or restoration time), and casualties. In a similar manner as done for
the other variables, the DVs are determined by integrating the conditional probabilities
of DV given DM, p(DV|DM), with the mean annual DM probability of exceedance,
p(DM).
The methodology just described and shown in Fig. 3 is an effective integrating
construct for both the performance-based earthquake engineering methodology. The
methodology can be expressed in terms of a triple integral based on the total probability
theorem, as stated in Eq. 1.

ν(𝐷𝑉) = ∭𝐺⟨𝐷𝑉|𝐷𝑀⟩|𝑑𝐺⟨𝐷𝑀|𝐸𝐷𝑃⟩|𝑑𝐺⟨𝐸𝐷𝑃|𝐼𝑀⟩|𝑑𝜆(𝐼𝑀) (1)

Though this equation form of the methodology might be construed as a minimalist

representation of a very complex problem, it nonetheless serves a useful function by
providing researchers with a clear illustration of where their discipline-specific
contribution fits into the broader scheme of performance-based earthquake engineering
and how their research results need to be presented. The equation also emphasizes
the inherent uncertainties in all phases of the problem and provides a consistent format
for sharing and integrating data and models developed by researchers in the various
disciplines.
The proposed methodology is intended to serve two related purposes. The first of
these is as a performance engine to be applied in full detail to the seismic performance
assessment of a facility. As illustrated in Fig. 2, the application would result in a
comprehensive statement of the probabilities of various losses (in terms of dollars,
downtime, and casualties) for events or time frames of interest to the owner or decision
maker for that facility. Though illustrated in an apparent static loading domain in Fig. 2,
this is for illustrative purposes only; the intent is to apply the methodology using a fully
nonlinear dynamic analysis.
It leads to the second intended purpose of the methodology. Presuming it can be
used to provide reliable results for a complete facility analysis, the methodology then

6
can be used as a means of calibrating simplified procedures that might be used for the
advancement of future building codes. It is in this application that the methodology is
likely to have its largest potential impact.
3. SEISMIC HAZARD IN MODERATE-SEISMICITY REGIONS

3.1 Definition of moderate seismicity region (excerpted from Scholz 2002)

Qualitative descriptions of earthquake size (moderate, great, etc.) are often used,
which, though based roughly on the magnitude, are meant to convey the potential
destructive power of the earthquake had it occurred in a populated area. Because in
this discussion we are considering only the physics of the rupture and not its effects,
such terms are not useful. We do, however, find it necessary to divide earthquake into
two classes, called simply large earthquakes and small earthquakes. Small
earthquakes are all those events whose rupture dimensions are smaller than the width
W* of the schizosphere (Fig. 4).
They therefore propagate and terminate entirely within the bounds of the
schizosphere and their behavior may be described as a rupture in an unbounded
elastic-brittle solid. A Large earthquake, in contrast, is one in which a rupture dimension
equals or exceeds the width of the schizosphere. Once an earthquake becomes large,
it is constrained to propagate only horizontally with its aspect ratio increasing as it
grows and its top edge at the free surface and its bottom at the base of the
schizosphere.
There are two reasons for making this distinction. The first is that it is found that
small and large earthquakes, so defined, obey different scaling relationships and
produce radiation with different spectral shapes, which may reflect their different
geometries and boundary conditions, the second is that we need to consider only large
earthquakes when quantitatively considering the role of earthquakes in tectonics.
Notice that the magnitude level where earthquakes change from small to large depends
on the tectonic environment. For the San Andreas fault, say, where the schizospheric
width is only about 15 km, this occurs at about M = 6~6.5, whereas in a subduction
zone, where the downdip width of the schizosphere is much greater, it may be at about
M=7.5.
This distinction between small and large earthquakes can be applied to two
earthquakes which occurred in 2016, Gyeongju in Korea and Kumamoto, Kyushu in
Japan as shown in Fig. 5.

7
 Fig. 4 Diagram Illustrating the definitions of small and large earthquake, showing
hypocenter (H), epicenter (E), moment centroid (MC), and the dimensions of rupture
(a, L, and W) (Scholz 2002)

(a) Moderate earthquake (MW= 5.4, 2016 Gyeongju, Korea) (Hong et al. 2017)

(b) Large earthquake (MW = 7.0, 2016 Kumamoto, Japan) (USGS 2016)
Fig. 5 Fault dimension of moderate and large earthquakes

3.2 Intraplate Earthquakes (excerpted from Scholz 2002)

Although something like 95% of the global seismic moment release is produced by
plate boundary earthquakes, there are significant numbers of earthquakes that occur
well away from plate boundaries. These intraplate earthquakes are important because
they greatly expand the region of possible seismic hazard from the proximity of plate
boundaries. Their role in tectonics is poorly understood, both from the viewpoint of the
origin of the forces that generate them and what sort of structures localize them.
One way to distinguish between interplate and intraplate earthquakes is based on the
slip rate of their faults and hence their recurrence time, as explained in Table 1. Intraplate
earthquakes are classified into two types. Type II earthquakes occur in broad zones near
and tectonically related to plate boundaries or in diffuse plate boundaries. Examples are

8
earthquakes of the Basin and Range province of western North America, which very
broadly may be considered to be part of the Pacific-North America plate boundary, or
inland earthquakes in Japan, which are tectonically a part of the compressional Pacific-
Eurasian plate margin. In contrast, Type III earthquakes occur in mid-plate regions and
seem to be unrelated to plate boundaries. This classification is of course somewhat
artificial, because there is a continuous spectrum of earthquake types and slip rates.
An important reason for this classification is that intraplate and interplate
earthquakes, so defined, have distinctly different source parameters, which
systematically have stress drops higher by a factor of 3 than the interplate earthquakes.

Table 1. Classification of tectonic earthquakes (Scholz 2002)

Type Slip rate (v), mm/yrs. Recurrence time, yrs.
I. Interplate ν > 10 ~ 100
II. Intraplate, plate boundary
0.1 ≤ ν ≤ 10 102 ~104
related
III. Intraplate, midplate ν < 0.1 >104
3.3 Characteristics of the seismic hazard in moderate seismicity regions
Either very infrequent major earthquakes or infrequent moderate earthquakes are
known to have occurred in east northern America (ENA). In either case, the more
recent seismic activity is minor. An example might be a region with a single recorded
occurrence of a magnitude 7 or larger earthquake, and no damaging earthquake since
(e.g., Memphis, Charleston, or Boston) or a string of earthquakes of magnitude 5 to 5.5
and sufficient geologic evidence to imply the possibility of a rare larger event.
(Nordenson and Bell 2000)
• The localities are not expecting nor are generally prepared for an earthquake, and
the buildings are, for the most part, not earthquake resistant.
• Typically, these regions are located away from tectonic plate boundaries, and
major faults, and so the source of earthquakes is less well understood and hazards
assessments are more difficult.
• The ground shaking caused by earthquakes diminishes, or “attenuates,” much less
with increasing distance from the earthquake. That means that for a given
magnitude the “felt area” and extent of damage is much greater in most moderate
seismic regions than in high seismic regions.
• There are few active faults, for which there is an average historic slip rate of 1 mm
per year or more and evidence of seismic activity within Holocene times (the past
11,000 years), and all the faults are hidden faults.
• Charleston earthquake (1886) was an exceptionally large magnitude earthquake
without any surface rupture in Fig. 6. The shaking effects of the 31 August 1886
Charleston, South Carolina, earthquake indicate that it was a major shock.
Moment magnitude estimates range from Mw 6.9 (Bakun and Hopper 2004) to Mw
7.3 (Johnston 1996). The fault movement that was the cause of the 1886 South
Carolina earthquake did not rupture the Earth's surface, but rather was confined
to its interior. Therefore, an important piece of direct field evidence, the direction
of the trend (termed by geologists the "strike") of the fault surface was not
provided for this earthquake nor was the direction of movement on the fault; that
is, vertical, horizontal, or some combination.
9
 (a) Reported intensities (Johnston 1996) (b) Damage and collapse of buildings
Fig. 6 The 1886 Charleston earthquake
3.4 Background seismic hazard maps in moderate seismicity regions
3.4.1 CENA in U.S.A. (excerpted from Frankel 1995)
Fig. 7 diagrams the four-model method used for the hazard map in Central and
Eastern (CENA) U.S.A. Three alternative models of hazard are used for this magnitude
range (Fig. 7 left). Model 1 is based on spatially-smoothed a-values derived from the
magnitude 3 and larger earthquakes since 1924. Here a is the activity level in the
Gutenberg-Richter equation log N = a-bM, where N is the number of events with
magnitudes greater or equal to M. In this model, the magnitude 3 and greater events
are assumed to illuminate areas of faulting which can produce destructive events.

Fig. 7 Chart of four models to make seismic hazard maps in the CENA in U.S.A.
(Frankel 1995)

The areas of large ground motions in Fig. 8(a) simply indicate areas with larger
numbers of magnitude 3 and larger events since 1924. This map does not contain the
hazard from events with magnitudes larger than 7.0, so it underestimates the
probabilistic ground motions for New Madrid and Charleston.
A trial map of probabilistic ground motions for model 3 (Fig. 7) is shown in Fig. 8(b).
The 25 cm/sec 2 contour line basically follows the boundary of the source zone. The
area within the 25 cm/sec2 contour (Fig. 8(b)) has a probabilistic ground motion of
about 30 cm/ sec2 (3% g), for 10% PE in 50 years.

10
 (a) model 1 (b) model 3
Fig. 8 Trial ground-motion map for models 1 and 3, 10% probability of exceedance in
50years (Values are peak ground acceleration in cm/sec2) (Frankel 1995)
3.4.2 A simple background seismic hazard in Korean Peninsula
PSHA is composed of 4 steps as shown in Fig. 9. The first step is the identification
of all the sources of earthquakes. Second is the statistical representation of the relation
between the magnitude of the earthquake and its frequencies by the Gutenberg-Richter
recurrence law. The third is the establishment of the attenuation law between the
ground motion parameters and the rupture or epicentral distance where the median and
standard deviation of ground motion parameters are to be obtained. Finally, the fourth
step is to derive the hazard curve represented by the relation between the hazard
parameter and the probability of exceedance of the specific parameter value.

Fig. 9 Four steps of a probabilistic seismic hazard analysis (Kramer 1996)

Uniform background zone in Fig. 8(b) assumes that the probability of occurrence of
the earthquake (Fig. 10) is uniform all over the region, so the number of occurrence for
each level of the earthquake is distributed uniformly as shown in Table 2.

11
 Ri-1 Site
Ri

R
• Rmax =200 km
• Area(Total)
Area = Area =125,664 km2

Fig. 10 Probability of occurrence of earthquake

Table 2. Number of M  5 intraplate earthquake events on land in a 50 years period (Lam 2014)
Country Land Area N(M5) in 50 years N(M5) in 50 years
(km) [Recorded Number] [Recorded Number
Normalized to 1,000,000 km2]
Australia 7,692,024 45 6
Brazil 8,515,767 33 4
Eastern US 2,291,043 13 56
Eastern & Central China 1,550,974 14 9
France 674,843 4 6
Southern India 635,780 3 5
Germany 357,021 1 3
British Isles 315,134 3 9  10
Peninsular Malaysia 131,598 <1 <1
Korean Peninsula 223,348 3 13
Total  = 22,387,532  = 120 Average = 5
The earthquake recurrence relationship (Fig. 11) assuming a doubly-truncated
exponential function can be expressed as follows:
1
Constant seismicity rate

exp[−𝛽(𝑚−𝑚0 )] −exp[−𝛽(𝑚𝑚𝑎𝑥 −𝑚0 )]

0.1
𝜆𝑚 = 𝜈 (2)
1 −exp[−𝛽(𝑚𝑚𝑎𝑥 −𝑚0 )]

0.01
𝑚0 ≤ 𝑚 ≤ 𝑚𝑚𝑎𝑥 , 𝑚0 = 4
λm

0.001 𝜈 = exp⁡(𝛼 − 𝛽𝑚0 )

𝛼 = 2.303a
0.0001 mmax 6 6.5 7 8

𝛽 = 2.303𝑏
0.00001
4 5 6 7 8
Magnitude, m
Fig. 11 Earthquake recurrence relationship

where 𝜆𝑚 is the number of earthquakes with magnitude greater than M, occurring in a

fixed time interval and within the circular source area of radius Rmax. ν is the total
number of earthquakes with magnitude greater than Mmin, occurring in a fixed time

12
interval and within the circular source area. β=2.3b, in which b is the slope of the
Gutenberg-Richter relationship. The corresponding probability density function is
defined as follows:
𝛽 exp[−𝛽(𝑀 − 𝑀𝑚𝑖𝑛 )]
𝑓(𝑀) = (3)
1 − exp[−𝛽(𝑀𝑚𝑎𝑥 − 𝑀𝑚𝑖𝑛 )]
The probability distribution of PHA in Korea is assumed to be same as that given by
GMPE of Boore (2008) in Fig. 12. For each combination of epicentral distance,
magnitude, and PGA, the probability of exceedance (PGA ≥ y*) can be obtained using
Eq. (4).
ln 𝑦 ∗ =ln 𝑃𝐺𝐴
P[𝑃𝐺𝐴 > 𝑦 ∗ |𝑚, 𝑟] = 1 − 𝐹𝑌 ( ) (4)
𝜎ln 𝑃𝐺𝐴

0.4

0.35
Attenuation relation (Boore 2008)
m = 6.25
0.3

0.25
PGA (g)

0.2

0.15
±1σ = ±0.5
0.1

0.05
Mean =

0
0 50 100 150 200
Epicentral distance (km)

Fig. 12 Relationship between epicentral distance and PGA

For every combination of M and R, seismic intensities are predicted by employing
suitable ground motion prediction equations (GMPEs). Finally, the total seismic hazard
of the site encompassing all the considered M-R combinations can be computed using
the conventional Cornell-McGuire approach (Cornell 1968, McGuire 1976) which is
represented by the following integral:
𝑁𝑀 𝑁𝑅

𝜆𝑦 ∗ = 𝜈 ∑ ∑ 𝑃[𝑌 > 𝑦 ∗ |𝑚𝑗 , ⁡𝑟𝑘 ]⁡⁡𝑃[𝑚𝑗 ]⁡⁡𝑃[𝑟𝑘 ] (5)

𝑗=1 𝑘=1

The probability of exceedance, or, the annual frequency of PGA ≥ y* is shown in

Fig. 13, where PGA’s corresponding to the probability of exceedance 10% and 2% in
50 years are 0.0253 and 0.0541g, respectively. The value of 0.025g for the probability
of 10% in 50 years is similar to the value of 2.5%g in the uniform background zone in
Fig. 8(b). these PGA’s are compared with the effective PGA in KBC 2016 (Table 4)
where PGA’s in KBC 2016 are about four times larger than those based on uniform
background zone in Korean Peninsula.

13
 0.1

Mean annual rate of exceedance of PGA

(0.0253g, 0.00212)
Seismic Hazard Curve
0.01
RP. 475 yrs
(0.05g, 0.000490 ) (지진 재해도 곡선)
0.001 RP. 2043 yrs.
(0.11g, 0.0000549 )
0.0001 RP. 18200 yrs.
(0.0541g, 0.000404)
0.00001 RP. 2476 yrs.

0.000001

0.0000001

1E-08

1E-09
0 0.05 0.1 0.15 0.2
Peak Ground Acceleration (PGA, g)
Fig. 13 Mean annual rate of exceedance of PGA for Korean Peninsula

Table 3. Comparison between Effective PGA in KBC 2016 and Background Hazard
Return periods (year) KBC 2016 Background Hazard
500 0.11g 0.0253g
2500 0.22g 0.0541g

The disaggregation for PGA 0.05g and PGA 0.11g (Fig. 14) are shown as
histograms on the epicentral distance and the magnitude. It can be found in this figure
that most of the contribution comes from within the distance less than 50 km and from
the magnitude ranging from M 4.5 to 6.5. In moderate seismicity regions such as ENA
and Korean Peninsula, the hazard derived from uniform background zone serve as the
lower bound for the probabilistic seismic hazard map.

Annaual rate of exceedance of a PGA 0.05g Annaual rate of exceedance of a PGA 0.11g
5.00E-05
1.00E-05

4.00E-05
8.00E-06

3.00E-05
6.00E-06
λm

λm

2.00E-05 4.00E-06
Magnitude

Magnitude

1.00E-05 6.0~6.5
6.25 6.0~6.5
6.25
5.5~6.0 2.00E-06
5.25
5.0~5.5 5.5~6.0
5.25
5.0~5.5
4.5~5.0
4.25 4.5~5.0
0.00E+00 4.0~4.5 0.00E+00 4.25
4.0~4.5
7 16 25 35 45 55 65 75 85 95 105 115 7 16 25 35 45 55 65 75 85 95 105 115
Distance (km) Distance (km)

(a) 0.05 g (b) 0.11 g

Fig. 14 Disaggregation for PGA: 0.05g and 0.11g

3.5 Seismicity and seismic hazard map in Korean Peninsula

Historical earthquakes were recorded in various historical sources including
Samgooksagi, Koryosa, and Choseon-wangjosillog, which were listed in various studies
(e.g., Wada, 1912; Lee and Yang, 2006). The seismic intensities and source
information of historical events during 2–1904 A.D. are collected from Lee and Yang

14
(2006). The number of earthquakes during the period of the Three Kingdoms is 56,
earthquakes during the period of the Unified Silla is 33, earthquakes during the period
of the Koryo dynasty is 158, and earthquakes during the period of the Choseon dynasty
is 1938 (Fig. 15(a)).
The historical earthquake records during the Choseon dynasty comprise about 89%
of the total historical earthquake records. Large-size events with seismic intensities of
VIII to IX are recorded mostly in the periods before the Choseon dynasty. On the
contrary, earthquakes with seismic intensities greater than IV were recorded well during
the Choseon dynasty (Fig. 15).
(c)

Fig. 15 (a) Temporal distribution and seismic intensities of historical earthquake

during 2-1904 A.D.; (b) Histogram for the numbers of records for every 50-year period; and
(c) Gutenberg-Richter recurrence law between magnitude and frequency
(Houng and Hong 2013)

From these data, Gutenberg-Richter recurrence law between magnitude and

frequency is shown in Fig. 15(c). However, the maximum magnitude is estimated to be
7.45±0.04, which appears to be an excessive overestimation due to the fact that no
active fault having surface rupture regarding this maximum event could be identified,
and the range of the region of the damage and casualties were not nationwide.
The Korean hazard map given in Fig. 16 cannot be easily matched to the historical
and instrumental distribution of earthquake events in Fig. 17. the rationale for the
development of Korean Seismic Hazard map is not well explained outside of the group
of Korean seismologist in contrast to the very transparent communication between the
seismologist group and the engineers or other stakeholder’s groups in the USA.

15
 (a) Return period: 500 yrs (b) Return period: 2400 yrs
Fig. 16 National Seismic Hazard Map in Korean Peninsula

(a) Historical earthquakes (b) Instrumented earthquakes

Fig. 17. Map of epicenters of historical and instrumental earthquakes (NEMA 2012)

3.6 Gyeongju Earthquake and GMM based on Gyeongju EQ records

3.6.1 Gyeongju Earthquake (excerpted from Hong et al. 2017)
A moderate-sized earthquake with a local magnitude of ML 5.8 occurred on 12
September 2016. The events occurred around the Yangsan fault zone in the
southeastern Korean Peninsula that had been seismically inactive (Fig. 18). The M5.8

16
earthquake is the largest event in the Korean Peninsula since 1978 when national
seismic monitoring began.
The peak ground accelerations reached 4.5g in the E-W component, 4.3g in the N-
S component, and 2.3g in vertical component at station USN at an epicentral distance
of 8.2km. Ground motions at an epicentral distance of 8.2km (station USN) are stronger
than those at an epicentral distance of 5.8km (station MKL). The spectra of
displacement waveforms at three local stations (MKL, USN, and HDB) display
characteristic high-frequency energy. The responsible fault rupture was not found on
the surface.

Fig. 18 (a) Tectonic setting around the Korean Peninsula and (b) an enlarged map of
the Korean Peninsula with the presentation of major geological provinces.
(Hong et al. 2017)
The brittle shear failure occurred at short columns in the basement of a 5-story
residential building structure and under the roof of the temple as shown in Fig. 19(b)
and (a). Many nonstructural failures occurred such as falling of oriental-roof tiles, glass
breakage and falling of objects at the stores (Fig. 19(c)). Because Gyeongju is the
ancient capital of Silla during 1st to 9th centuries, many cultural heritages were
damaged or deformed as shown in Fig. 19(d).

(a) Shear failure of short column (Lee 2017) (b) Failure of column at the
base of structure (KBS)
Fig. 19. Damage and failure modes during 2016 Gyeongju Earthquake

17
 (c) Damages of non-structural element (roof tile, window, goods) (YTN, Ohmynews)

(d) Damages of artifacts - Cultural Heritage (Yonhapnews)

Fig. 19. Damage and failure modes during 2016 Gyeongju Earthquake (cont.)

3.6.2 GMM based on Gyeongju EQ records

A ground motion model (GMM) with its parameters has been determined based on
the seismograms obtained from Gyeongju earthquake (KMA 2016). Atkinson’s
attenuation model (2004) was adopted with the geometric spreading R -1.3 up to 70 km.
With this model, stress parameter was estimated to be 831 bar, which is very high
when compared to the ordinary value of 100 bar in intraplate regions. This high value of
stress parameter was used to simulate the exceptionally spectral values observed near
the source. Soil conditions for all the stations are assumed to correspond to NEHRP
B/C boundary (760 m/s). Response Spectral Accelerations (RSA g’s) of the synthetic
earthquake accelerograms are compared with RSA’s obtained from those of Gyeongju
earthquake in Fig. 20(a). Since the value of stress parameter was determined to match
the near-source accelerogram, RSA’s at 12 km, 14 km, and 25 km distances are similar
between synthetic and real accelerograms. However, the estimation by synthetic
accelerograms are very conservative at the far distance. Similar results can be found in
Fourier spectra in Fig. 20(b).

18
 (a) Response spectra

(b) Fourier spectra

Fig. 20 Comparison of spectra with actual records and simulated seismograms

Fig. 21 compares the recorded accelerograms and the synthetic at the hypocentral
distances, 12 km, 14 km, and 25 km. In some cases, such as that of 12 km distance,
the synthetic record shows a larger PGA than the recorded. Over all synthetic
accelerograms simulate well the recorded in shape and duration.

19
 0.5 0.5 0.5
Gyeongju EQ (2016.09.12) Gyeongju EQ (2016.09.12) Simulated
Acceleration (g)

Acceleration (g)

Acceleration (g)
0.25 MKL station 0.25 MKL station 0.25 RHYPO = 12 km
RHYPO = 12 km RHYPO = 12 km
0 0 0

-0.25 -0.25 -0.25

EW Component NS Component
-0.5 -0.5 -0.5
0.5 0 2 4 6 8 10 0.5 0 2 4 6 8 10 0.5 0 2 4 6 8 10
Gyeongju EQ (2016.09.12)
Time (sec) Gyeongju EQ (2016.09.12)
Time (sec) Time (sec) Simulated

Acceleration (g)
Acceleration (g)

Acceleration (g)
0.25 USN station 0.25 USN station 0.25 RHYPO = 14 km
RHYPO = 14 km RHYPO = 14 km
0 0 0

-0.25 -0.25 -0.25

EW Component NS Component
-0.5 -0.5 -0.5
0.2 0 2 4 6 8 10 0.2 0 2 4 6 8 10 0.2 0 2 4 6 8 10
Gyeongju EQ (2016.09.12)
Time (sec) Gyeongju EQ (2016.09.12)
Time (sec) Time (sec) Simulated
Acceleration (g)

Acceleration (g)

Acceleration (g)
0.1 DKJ station 0.1 DKJ station 0.1 RHYPO = 25 km
RHYPO = 25 km RHYPO = 25 km
0 0 0

-0.1 -0.1 -0.1

EW Component NS Component
-0.2 -0.2 -0.2
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
Time (sec) Time (sec) Time (sec)

(a) EW (b) NS (c) Simulated

Fig. 21 Comparison of near-source ground motions of Gyeongju earthquake and
synthetic ground motions corresponding to Gyeongju earthquake

Fig. 22(a) shows the comparison of PSA0.2s obtained from Saguenay earthquake
(MW = 5.8) which have the magnitude similar to Gyeongju earthquake, and the
attenuation curves by the several models. The model A04 presented the value of stress
parameter, 2161 bar. Fig. 22(b) compares the GMM attenuation curve with the PSA0.2s
obtained from Gyeongju earthquake, where the GMM attenuation for Gyeongju match
well those from recorded ground motion, but in the conservative side at the long
distances. With this GMM, the synthetic accelerograms for the larger magnitude
earthquake are generated.

(a) Saguenay earthquake

(b) Gyeongju earthquake
(Boore et al. 2010)
Fig. 22 Comparison PSA (T=0.2 s) distributions of actual records and GMM or GMPE

The synthetic earthquake accelerograms, when MW 6.5 earthquake would occur,

are shown in Fig. 23(a). The response acceleration spectrum (RAS) of this synthetic
record is compared with RAS of 1985 Nahanni earthquake in Canada (MW = 6.9) in Fig.
23(b).

20
 2
Simulated seismogram, Mw=6.5, RHYPO = 10 km
Acceleration (g)

-1

-2
0 5 10 15 20
Time (sec)

(a) Time history of acceleration (b) Response acceleration spectra (RAS)

Fig. 23 Synthetic earthquake accelerograms

3.7 Seismic Loads in moderate seismicity regions

3.7.1 Eastern North America
ENA is one of the representative moderate seismicity regions in the world and adopts
the methodology developed in WNA for seismic hazard analysis very competently. Two
representative cities in ENA are New York City and Charleston which experienced a
strong earthquake in 1886. The earthquake load PSA for these two cities in IBC and
ASCE 7 at the short period (0.2 s) and 1 second are SS = 0.3g and S1 = 0.06g for New
York, and SS=1.25g and S1=0.3g for Charleston. (Fig. 24) The corner period between
velocity constant and displacement constant regions are 6s and 8s for New York and
Charleston, respectively.

0.6 0.6
Acceleration Spectra ADRS Spectra
Site Class: A Site Class: A
Site Class: B Site Class: B
0.4 Site Class: C 0.4 Site Class: C
Site Class: D
Sa (g)
Sa (g)

Site Class: D
Site Class: E Site Class: E
0.2 0.2 Sd = 5.96 cm (Site: B)

TL = 6 sec

0 0
0 2 4 6 8 0 5 10 15 20 25
Period, T (sec) Sd (cm)
(a) New York
1 1
Acceleration Spectra ADRS Spectra
Site Class: A Site Class: A
0.75 Site Class: B 0.75 Site Class: B
Site Class: C Site Class: C
Site Class: D
Sa (g)

Sa (g)

Site Class: D
0.5 0.5 Site Class: E
Site Class: E

0.25 0.25
TL = 8 sec

0 0
0 2 4 6 8 10 0 20 40 60 80 100 120
Period, T (sec) Sd (cm)
(b) Charleston
Fig. 24 Design spectra corresponding to IBC and ASCE in ENA

21
 3.7.2 Australia
The Australian continent is a representative intraplate region. Australia developed
its seismic hazard map and seismic load based on the analysis of its historical and
instrumental earthquake records. Fig. 25 compares design spectrum in Melbourne,
Australia and that in Seoul, Korea, corresponding to soil condition C. The spectrum of
Australia has two corner periods, where the second corner period T2 determines
displacement-constant region. The value of T2 is 1.5s in Australia which is much shorter
than 6s in New York and 8s in Charleston. Also, it can be seen that Korean design
spectrum (ADRS) is much larger than that of Melbourne.

0.7
60 30 0.7
sign spectrum Seoul (Sc)
Design spectrum Seoul (Sc) Seoul (Sc)
Acceleration, Sa (g)

Acceleration, Sa (g)
0.6
50 Design spectrum
0.6
Seoul (Sc)
Melbourne (Sc)
Melbourne (Sc) 25 Melbourne (Sc)
0.5
40 20
0.5 Seoul (Sc)
Melbourne (Sc) Melbourne (Sc)
Sv (cm/s)

Sd (cm)
0.4 0.4
30 15
0.3 0.3
20 10
0.2 0.2 T2 = 1.5 sec
10 5
0.1 0.1
0 0
0 0
3 4 0 1 2 3 4 0 1 2 3 4
ec) 0 1 T2
(sec) 3 4 0 2 T (sec) 4 6 8 10
T (sec) Displacement, Sd (cm)
ADRS
Seoul (Sc)
Seoul Melbourne
Melbourne (Sc)
 Zone factor, S=0.22g  Hazard factor, a=0.144g (z=0.08g)
(Return period of Zone 1 : 2400yr)  Probability factor, kp=1.8 (2500yr)
 T0= 0.107, Ts=0.536  T1= 0.35, T2=1.5
 Soil factor: Sc (Vs.30=360~800m/s)  Soil factor: Be (Vs.30=360m/s ~)

15
Fig. 25 Design spectra for Seoul and Melbourne: low-to-moderate seismicity regions
20
cm) (Lam 2014)

3.7.3 Korean Building Code 2016

The current KBC design spectra are shown in Fig. 26, and compared with those of
1940 El Centro earthquake (MW = 6.9), and 1952 Taft earthquake (MW = 7.3). These
figures clearly reveal that Korean earthquake design loads are not actually for
moderate, but for strong earthquake ground motions.
The level of earthquake loads for each risk category of building structures
determines seismic design category as shown in Table 4. Zone factor 0.22 g in Seoul
has the seismic design category D for the soil condition S C, SD and SE, where SDC D
means that the seismic details (for the strong-seismicity region) be imposed with other
additional requirements. The situation in Seoul is similar to the case of Sacramento in
California as shown in Table 4. However, New York in ENA does have one D only for
soil condition SE and highest risk category IV.

22
 0.7 0.7
KBC 2016_Sc
SE
0.6 0.6 KBC 2016_Sd
El centro (Sd)
0.5 SD 0.5 taft (Sc)

Sa (g)
SC 0.4
Sa (g)

0.4 SD with depth to bedrock < 20m

SB
0.3 SC with depth to bedrock
0.8
< 20m
0.3 120 KBC 2009_Sc
60 KBC 2009_Sc
SA Design spectrum KBC 2009_Sc
KBC 2009_Sd KBC 2009_Sd
Design spectrum Design spectrum KBC 2009_Sd
100 50
0.2 0.6 El centro (Sd)0.2 El centro (Sd)
taft (Sc)
El centro (Sd)
taft (Sc)
taft (Sc) 80 40
0.1

Sd (cm)
Sa (g)

Sv (cm/s)
0.1 0.4 60 30

0 0.0 40 20
0.8 0.2 120 60
Design
0 spectrum
1 KBC 2009_Sc
2
KBC 2009_Sd 100
3 KBC 2009_Sc
KBC 2009_Sd 4Design spectrum0 20 50
1 spectrum
Design 2 3
KBC 2009_Sc
KBC 2009_Sd
10 4
0.6 T (sec)
El centro (Sd) 0.0 El centro (Sd) 0 T (sec) El centro (Sd)
0
taft (Sc) taft (Sc) taft (Sc)
080 1 2 3 4 40
0 1 2 3 4 0 1 2 3
0.7

Sd (cm)
Sa (g)

Sv (cm/s)

T (sec) T (sec) T (sec)

0.4 ADRS 0.8
KBC
60 2016_Sc 30
0.6 KBC
ADRS 2016_Sd KBC 2009_Sc
20
0.2 El40centro (Sd) KBC 2009_Sd
El centro (Sd)
0.5 0.6 taft
20 (Sc) taft (Sc) 10ElCentro earthquake Taft earthquake
0  Magnitude, M=7.3
Sa (g)

Magnitude, M=6.9
Sa (g)

0.0 0
0.4 0 0.4
1 2 3 4 0 1 2 3 4  Rupture
0 1
distance, 2 R=12.99km
3 4  Rupture distance, R=43.49km
T (sec) T (sec) T (sec)
0.3
0.8 0.2  Soil factor, Sd (VS.30=213.4m/s)  Soil factor, Sc (VS.30=385.4m/s)
ADRS KBC 2009_Sc
0.2
0.6
KBC 2009_Sd
El centro (Sd)
0.0
taft (Sc) El Centro earthquake Taft earthquake
0.1 0 10 20 30 40
 Magnitude,  Magnitude, M=7.3
Sa (g)

M=6.9
Sd (cm)
0.4
0.0  Rupture distance, R=12.99km  Rupture distance, R=43.49km
0.2  Soil factor, Sd (VS.30=213.4m/s)  Soil factor, Sc (VS.30=385.4m/s)
0 10 20 30 40
0.0
Sd (cm)
0 Fig. Sd
10 26
20 Comparison
(cm)
30 40 among KBC2016, El Centro, and Taft Spectra

Table 4. Seismic Design Categories

KBC 2016 IBC 2000 IBC 2006 (New York)
Site Seismic Design SDS SD1 Seismic Design
Sacramento,
Class SDS* SD1* Category(SDC) Category(SDC)
CA, SDC II
Special I II IV III I, II
SA 0.293 0.117 C B B B 0.160 0.032 A A A
SB 0.366 0.146 D C C C 0.200 0.040 B B C
SC 0.439 0.234 D D D D (C) 0.240 0.068 B B C
SD 0.527 0.336 D D D D 0.312 0.096 B B C
SE 0.732 0.497 D D D D 0.468 0.140 D C C
* Seismic Zone 1, S = 0.22 g

Lateral force resisting building system is classified as shown in Table 5. There are
some differences between KBC 2009 and IBC 2006. Korean Building Code follows the
classification of US codes. Some important differences are the height limit for high-rise
building structures, but KBC 2009 requires the special seismic details for the building
structures with the height exceeding 60m and belonging to the design category D. Most
of the residential buildings in Korea do not belong to this category, but more residential
buildings recently constructed exceed this height limit, therefore become subject to
special detailing requirement as shown in Fig. 27, where the congestion of
reinforcement due to this requirement cause difficulty in construction.

23
 Table 5. Design factors for RC lateral force-resisting systems (Fardis 2014)
Code KBC 2016 IBC 2006 (=ASCE 7-10)
Height limit Height limit
Seismic Force- Design factors Design Design factors Design
Resisting System Category Category
R Ω0 Cd C D R Ω0 Cd C D
Bearing wall Special RC walls 5 2.5 5 - - 5 2.5 5 - 50m
systems Ordinary RC walls 4 2.5 4 - 60m 4 2.5 4 - X
Building frame Special RC walls 6 2.5 5 - - 6 2.5 5 - 50m
systems Ordinary RC walls 5 2.5 4.5 - 60m 5 2.5 4.5 - X
Special MRF 8 3 5.5 - - 8 3 5.5 - -
Moment resisting
Intermediate MRF 5 3 4.5 - - 5 3 4.5 - -
frame (MRF)
Ordinary MRF 3 3 2.5 - X 3 3 2.5 X X
Dual systems with Special RC walls 7 2.5 5.5 - - 7 2.5 5.5 - -
special MRF Ordinary RC walls 6 2.5 5 - - 6 2.5 5 - X
Dual systems with Special RC walls 6.5 2.5 5 - - 6.5 2.5 5 - 50m
intermediate MRF Ordinary RC walls 5.5 2.5 4.5 - 60m 5.5 2.5 4.5 - X

(a) Special details of shear walls (b) Mock-up test of special shear wall
(30-story residential bldg. in Daegu, Korea)
Fig. 27 Problems of special details required for SDC D (Chung et al. 2013)

Fig. 28 compares RSA by 200 synthetic accelerograms using GMM based on

2016.9.12 Gyeongju earthquake records with design spectrum of KBC. The GMM uses
the value of stress parameter, 831 bars and it is shown in the previous section that the
RSA by this GMM matches well the RSA by the Gyeongju earthquake records at near-
source distances with RSA by this GMM being conservative at the far-source distance.
The soil condition assumed in GMM is NEHRP B/C boundary (VS,30 = 760 m/s) while
design spectrum is based on SB (VS,30 = 760 m/s ~ 1,500 m/s). RSA’s for median +1σ
by the synthetic ground motions generated for MW = 6.5 and rupture distance = 30 km
are twice larger than those by KBC 2016 site: B in the range of short periods with S d
being less than, or equal to 6 cm. The cross point between ADRS by the synthetic and
ADRS by KBC 2016 implies the second corner period for constant displacement region,
herein T2 = 1.3 sec, which is similar to T2 = 1.5 sec in the Australian code. The
characteristics of earthquake ground motions for near source distance expected in
Korean peninsula can be summarized as follows:

24
 • The accelerations at the short periods can be twice as large as the KBC
accelerations.
• The displacements at the long periods larger than 1.5 sec are bounded by some
value, such as 6cm for soil condition, SB.

Another case for MW = 6.0 and rupture distance = 20 km is shown also in Fig. 27.
The trend is that the RSA in short periods become 20% higher and the S d in a long
period much lower than the those of previous case for MW = 6.5, rupture distance = 30
km.

1 1
Mw = 6.5, R = 30 km Mw = 6.0, R = 20 km
0.75 ±σ 0.75 ±σ
KBC2016, Site: B KBC2016, Site: B
Sa (g)

Sa (g)
0.5 0.5

0.25 0.25

0 0
0 1 2 3 4 0 1 2 3 4
Period, T (sec) Period, T (sec)
(a) Acceleration response spectra
1 1
Mw = 6.5, R = 30 km Mw = 6.0, R = 20 km
0.75 ±σ 0.75 ±σ
KBC2016, Site: B KBC2016, Site: B
Sa (g)

Sa (g)

0.5 0.5
T = 0.75 sec
T = 1.3 sec
0.25 0.25

0 0
0 2 4 6 8 10 0 2 4 6 8 10
Sd (cm) Sd (cm)
(b) Acceleration-displacement response spectra (ADRS)
Fig. 28 Comparison with response spectra by 200 synthetic accelerograms using
GMM based on 2016.9.12 Gyeongju earthquake records with KBC design spectrum

4. PBEE IN A MODERATE SEISMICITY REGION: SOUTH KOREA

4.1 Example of PBEE in Moderate Seismicity Region: Evaluation of RC MRF

designed with different loads
Evaluation of a typical low-story RC MRF building structure in Korea was performed
by using PBEE. The prototype is shown in Fig. 29 with the location being Seoul
(effective ground motion factor, S = 0.22 g) floor and total area 1,215 m 2 and 4,860 m2,
respectively, used for office and Risk Category “ordinary” (Seismic grade II in Korea),
soil condition SC. This prototype was designed for the two seismic load levels
corresponding to 2/3 of the intensity of maximum considered earthquake (MCE; return
period 2,500 years) specified KBC 2009 and corresponding to the intensity of

25
earthquake with the return period of 500 years. The details for these two load cases are
shown in Table 6.
54000

9000 9000 9000 9000 9000 9000

54000

7500
9000 9000 9000 9000 9000 9000

3600 3600 3600 3600

22500
7500

14400
7500
X Z

Y X

(a) plan (b) elevation

Fig. 29 Prototype building structure: 4-story RC MRF in Korea

Table 6. Design seismic load of 4-story prototype building model according to KBC 2016
Parameter Value
Design spectrum Earthquake with return
Seismic load
(KBC) period of 500 years
Seismic zone factor S = 0.22 for Seoul S = 0.11 for Seoul
Soil type SC
SDS = 0.433 g; SDS = 0.330 g;
Design spectral accelerations at 0.2s and 1.0s
and SD1 = 0.232 g and SD1 = 0.174 g
Seismic design category D C
Response modification factor R=3
Displacement amplification factor Cd = 2.5
Importance factor IE = 1.0
Fundamental period
Ta. = 0.540 s
(empirical equation)
Seismic coefficient (Cs = SD1/(R/IE×1.5T)) Cs = 0.1431 Cs = 0.1100

Fig. 30 shows the results of unidirectional pushover analysis in the X and Y

directions for these two designs. The fragility of the collapse was developed using the
SPO2IDA tool provided in FEMA P58 (2012). Figs. 31 and 32 present the concept of
SPO2IDA and some illustrations.

26
 15000 15000
Pushover curve, X-dir Pushover curve, Y-dir

10000 10000

Vx (kN)
Vx (kN)

5000 5000

RC-MRF (RP. 500yr, Sc) RC-MRF (RP. 500yr, Sc)

RC-MRF (KBC2009, Sc) RC-MRF (KBC2009, Sc)
0 0
0 0.01 0.02 0.03 0 0.01 0.02 0.03
Roof drift ratio Roof drift ratio
Fig. 30 Pushover curves in the X and Y directions
3000
pushover

fit
2500

2000

Base shear (kips)

1500

1000

500

0
0 0.5 1 1.5 2 2.5
Roof displacement (ft)

Fig. 31 Concept of SPO2IDA using the results of Fig. 32 Illustration of SPO2IDA

pushover analysis and IDA (FEMA-P58 2012) for the prototype buildings

The fragility curves of collapse can be obtained by using IDA results in SPO2IDA as
shown in Fig. 33, where the value of abscissa, Sa, represents the spectral acceleration
at the fundamental period of the prototype, T = 0.89 s. According to this fragility curves
for the MCE represented by Sa = 0.39 g (SC) Design I for the intensity of earthquake
with the return period of 500 years shows the probability of collapse 0.916 %, while
Design II for the intensity of 2/3 of the earthquake with the return period of 2,500 years
reveals the probability of 0.0889%.

27
 1

0.8 Design for earthquake with

return period 500 years
Probability

0.6
for 2/3 of
Design per
Sa for MCE (SC) intensity of MCE
0.4

0.2
0.91%
0.089%
Collapse fragility
0
0 0.5 1 1.5 2 2.5 3
Spectral Acceleration (T = 0.89 s) (g)
Fig. 33 Fragility curves of collapse of prototype

When the developed fragility of collapse is input to PACT provided in FEMA P58,
the economic loss can be predicted. Fig. 34 describes the procedure for the loss
estimation. Three options of assessment type provided in FEMA P58 are Intensity,
Scenario, and Time-based assessment. In this study, the type of Intensity assessment
was used. Earthquake hazard is defined as that for the MCE with the return period of
2,500 years in Korea. Building response was analyzed using the simplified method
such as SPO2IDA or nonlinear response method. Building performance model can be
defined in PACT, and the results from this model are also provided in PACT. Table 7
shows the input items of structural and nonstructural elements for the prototype.

Obtain Site and Building Description

Select Assessment Type and Performance Measure

Define Earthquake Hazard

Analyze Building Response

(Simplified or Nonlinear Response method)

Assemble Building Performance Model

Review Results for Selected Performance Measures

Fig. 34 Procedure for the loss estimation (FEMA-P58 2012)

Table 7. Input items of structural and nonstructural elements for the prototype
(FEMA-P58 2012)
Component Type Quantity (unit) / story Demand

28
 X dir. Y dir. Parameter
Column & ACI 318 SMF, Concrete Column & beam
28 (EA) 28 (EA) IDR
beam joint = 24" x 24", Beam both sides
Curtain Walls - Generic Midrise Stick-Built
Curtain wall, Config: Monolithic, Lamination: 69.73 29.03
Window Unknown, Glass Type: Unknown, Details:
IDR
(SF 30) (SF 30)
Aspect ratio = 6:5, Other details Unknown

Non-monolithic precast concrete stair

Stair assembly with concrete stringers and 2 (EA) 2 (EA) IDR
treads with no seismic joint.
Suspended Ceiling, SDC A,B, Area (A): 7.27 Floor
Ceiling
1000 < A < 2500, Vert support only (SF 1800) Acceleration
Wall Partition, Type: Gypsum with metal
Wall 2.36 2.70
studs, Full Height, Fixed Below, Fixed IDR
Partition (LF 100) (LF 100)
Above
In Table 8, the results of economic loss and casualties are compared for the two
designs on the two seismic hazards represented by the earthquakes with the return
periods of 500 years and 2,500 years. The cost of new construction is assumed 8,800
million W (1US$ = 1,100 W: 8 million US$) based on the average known cost of 6
million W for 3.3 m2. When the building was subject to the earthquake with the return
period of 500 years, the repair costs are estimated less than 10% of the cost of the new
construction with the repair time being five months and with negligible casualties for
both designs. When the building was subjected to the earthquake of return period of
2,500 years, the number of casualties is still less than 1, but the repair cost occupies 55%
and 28% of the new construction with repair time being almost one year for Design I
and II, respectively. It means that though the structure would not collapse, relatively
high repair cost and time may render a new construction.

Table 8. Economic loss and casualties for the prototype

Economic loss, Repair
Return period (RP) Casualties
of seismic load
Seismic design billion ₩ time
(people)
(cost ratio: repair/rebuilding) (day)
Design earthquake 0.38
500 yrs. 71 0.067
in KBC 2016 (4%)
(10% exceedance
in 50 yrs.) Earthquake 0.82
150 0.079
with RP of 500 yrs. (9%)
Design earthquake 2.41
2400 yrs. 339 0.725
in KBC 2016 (28%)
(2% exceedance
in 50 yrs.) Earthquake 4.8
360 0.955
with RP of 500 yrs. (55%)

4.2 Example of PBEE in Moderate Seismicity Region: RC High-rise wall structure

A major portion of residential buildings (more than 58 % of total residential units)
29
has been constructed using RC wall structures with the most typical height of 15 stories
as shown in Fig. 35. The most popular plan of residential units is a two-unit plan with
15-story height as given in Fig. 36, which is the prototype of this study. The seismic
design was conducted using KBC2009, and the details are given in Table 9.
70
Year 2010

Ratio of Apartments / Total (%)

35
60 Total No. of housing units: 14,577,419 58.4 Statistics of number of stories
52.5 30 31.8
Total No. of apartment units: 8,576,013
50 47.7
25

Percentage (%)
40 37.5
20

30 15 13.1
22.6
10
20 10
13.4 6.5
4.3 5.1
10 7.9 5 2.4 3 3.3 3.2
0.7
2.2
1.1 1.5 2.3 1.6 2.2 2.1 1.8
0.4 0.4 0.7
0 0

10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
-4
5
6
7
8
9

25~
1980 1985 1990 1995 2000 2005 2010
Number of stories
Year National Census

(a) A bird’s eye view of a (b) Statistics of number of (c) Statistics of number of
district of Seoul residential building unit stories
Fig. 35 High-rise RC residential building structures in Korea

(a) Plan (unit: mm) (b) Elevation

Fig. 36 Prototype building: 15-story RC wall building structure

Table 9. Design seismic load of 15-story prototype building model

according to KBC 2009
Parameter Value
Seismic zone factor S = 0.176 for Seoul
Soil type SC
Design spectral accelerations at 0.2s and 1.0s SDS = 0.352g; and SD1 = 0.191g
Seismic design category C
Response modification factor R=4
Displacement amplification factor Cd = 4
Importance factor IE = 1.2
Fundamental period (empirical equation) Ta,X-dir. = 1.17s; and Ta,Y-dir. = 0.787s

30
 Seismic coefficient (Cs = SD1/(R/IE×1.5T)) Cs,X-dir. = 0.0326; and Cs,Y-dir. = 0.0485
Effective seismic weight, W W = 32,400 kN
Design base shear (V=CsW) VX-dir. = 1,060kN; and VY-dir. = 1,570kN

The seismic fragility of the prototype was obtained using the cloud method as given
in SAC/FEMA approach and the IDA (Incremental Dynamic Analysis). Fragility curves
corresponding to the limit states(LS’s) described in Table 10 with their IDR’s (%) are
shown in Fig. 37.
The probabilities of failure regarding each limit states when the prototype was
subjected to the DE and MCE in Korea are given in Table 11. The prototype has the
probability of 90% for the LS 1, which means the occurrence of major cracks (width >
0.02mm) with that of first yielding of main reinforcement being about 10%. However,
the probability of collapse of the 15 story RC wall building structure appears to be very
low not only for DE but also MCE earthquake.

Table 10. Definition of limit states, LS (Ji et al. 2007)

Level Limit state Description IDR(%)
Minor (including distributed) cracking in the primary load
LS 1 Serviceability 0.20
resisting structural system (crack width > 0.2mm)
Damage First yielding of longitudinal steel reinforcement; or
LS 2 0.58
control presence of first plastic hinge
Collapse Ultimate capacity of main load-resisting structural system; or point
LS 3 1.5
prevention of decreasing capacity in overall load-deformation response

Table 11. Probability of exceeding LS1 to 3 under design earthquake(DE)

and maximum considered earthquake (MCE) in Korea
DE MCE
Limit State
SAC/FEMA IDA SAC/FEMA IDA
LS 1 79% 57% 97% 87%
LS 2 0.3% 1.7% 5.1% 10%
LS 3 0.00000358% 0.0000173% 0.000157% 0.000362%

31
 DE in Korea, Sa(T=1s) = 0.147g
MCE in Korea, Sa(T=1s) = 0.22g
1
Probability of exceeding LS

LS1: 79% LS1: 97%

0.8 LS1: 87% SAC/FEMA LS1
SAC/FEMA LS2
LS1: 57% SAC/FEMA LS3
0.6 IDA LS1
IDA LS2
0.4 IDA LS3
LS2: 1.7%
LS2: 0.3%
0.2 LS2: 10%
LS2: 5.1%
0
0 0.2 0.4 0.6 0.8 1
Sa(1.0s) (g)
Fig. 37 Seismic fragility curves of 15-story RC wall building structure

Seismic loss estimation was conducted using these fragility curves and PACT, the
tool provided by FEMA P58. In this case, Analyses Building Response in the procedure
(Fig. 34) was obtained by using Nonlinear Response Method instead of Simplified
Method. Fig.38 presents the fragility of RC wall panel and loss function regarding repair
cost and time implicit in PACT. The resulting loss estimations are given in Figs. 39 and
40, where the economic loss of one prototype is estimated 0.9 million dollars to 3.7
million dollars (median = 1.8 million US$) while the repair time ranges from 75 days to
300 days (median = 150 days).

(a) Component fragility (b) Repair cost (c) Repair time

Fig. 38 Loss function of wall structure

32
 Repair Cost Repair Time
85%

median repair cost = 1.8 million U.S. $

median rep

15% 15%

Fig. 39 Economic loss by repair cost

Repair Cost Repair Time

85% 85%

epair cost = 1.8 million U.S. $

median repair time = 150 days

15%

Fig. 40 Repair time

33
 4.3 Expected range of force and deformation in a moderate seismicity region:
South Korea
The experimental researches through the earthquake simulation tests to identify
the seismic weakness of reinforced concrete (RC) nonseismic building structures
designed only for gravity loads and to observe the seismic performance of RC
residential building structures designed per the recent Korean seismic code are
presented. Based on all these observations, expected ranges of force and deformation
are summarized for code writers or engineers in moderate seismicity regions.

4.3.1 1:5-scale 3-story RC ordinary moment-resisting frame with nonseismic

detailing
The objectives of the research (Lee and Woo, 2002a) are to investigate the
seismic performance of a 3-story reinforced concrete (RC) ordinary moment-resisting
frame, which has not been engineered to resist earthquake excitations. The prototype
of this test model was adopted from a building structure for the police office, actually
built and in use in Korea. The important characteristics in the Korean detailing practice
are as follows: (1) the splice is located at the bottom of the column, (2) the spacing of
hoops is relatively large, (3) seismic hooks are not used, (4) confinement
reinforcements are not used in beam-column joints, and (5) the special style of
anchorage in the joints.
This model was subjected to the shaking table motions simulating Taft N21E
component earthquake ground motions, whose magnitude of peak ground acceleration
(PGA) was modified to approximately 0.12g, 0.2g, 0.3g, and 0.4g in Table 12. Due to
the limitation in the capacity of the used shaking table, a pushover test was performed
to observe the ultimate capacity of the structure after earthquake simulation tests.
Though the bare frame (BF) model structure in this study (Fig. 41) was designed
only for the gravity loads in zones of low seismicity, the model showed the linear elastic
behavior under the Taft N21E motion with the peak ground acceleration of 0.12g,
representing the design earthquake (DE) in Korea (Fig. 42(a)). The structure could
resist not only the DE, which it would be supposed to resist if it were to be designed
against earthquake but also the higher levels of the earthquake excitations. The main
components of its resistance to the high level of earthquakes are the high over-strength
(Fig. 42(b)). The model structure has the overall displacement ductility ratio of 2.4 and
the over-strength factor of approximately 8.7.

Frame A
420
(Instrumented frame)

840 1680

Frame B

420

50 1260 480 1260 60

3110

(a) Plan (b) Shaking table tests

Fig. 41 1:5-scale 3-story RC moment-resisting frame model (Lee and Woo, 2002a)

34
 Table 12. Test program of BF model
Identification of Test PGA (g) Remarks (Return Period)
TFT_012 0.12 Design earthquake (EQ.) in Korea (500 years)
Earthquake
TFT_02 0.2 Max. EQ. in Korea (1000 years)
Simulation
TFT_03 0.3 Max. considered EQ. in Korea (2000 years)
Test
TFT_04 0.4 Severe EQ. in high seismic regions of the world
Pushover Static Test PUSH - Ultimate capacity of the structure
60
crushing of concrete at column

40
first significant yield
Push-II
24.33 Push-I
Base Shear(kN)

20

0 10.82 y=20.0 u=47.2

-60 -40 -20 0 20 40 60
TFT_012 (Experiment)
-20 TFT_02 (Experiment)
TFT_03 (Experiment)
TFT_04 (Experiment)
Pushover (Experiment)
-40 PUSH-I (Analysis)
PUSH-II (Analysis)
TFT_012(Analysis)
TFT_04(Analysis)
-60
Roof Drift(mm)

(a) Base shear versus roof drift in tests and analyses

Ω = Ωs Ωy = (Cs/Cω)(Cy/Cs)
Ω = 11.1 (with strain aging)
8.7 (without strain aging)

(b) Typical global structural response idealized (c) Development of cracks

as linearly elastic-perfectly plastic curve in pushover test
Fig. 42 Test results of BF model (Lee and Woo 2002a)

4.3.2 1:5-scale 3-story masonry-infilled RC frame with nonseismic detailing

Lee and Woo (2002b) investigated the actual responses of masonry-infilled RC
frame with nonseismic detailing under the simulated earthquake ground motions. After
earthquake simulation tests, the monotonically-increasing lateral load test or the
pushover test was performed to find out the ultimate capacities of the model. By
comparing the results of these tests with those in the case of the bare frame (Lee and
Woo, 2002a), the significance or the effect of masonry infills are evaluated. Two layouts
of masonry infills in Figs. 43(a) and (b) were used for earthquake simulation tests: that is,
fully infilled frame (FIF) and partially infilled frame (PIF). The adopted input ground
accelerogram is the Taft N21E component, and the peak ground acceleration (PGA) was
modified to 0.12g, 0.2g, 0.3g, and 0.4g as shown in Table 12, which is the same as for BF
model. After the series of earthquake simulation tests have been conducted on the FIF
35
model, there appeared to be only minor cracks on the masonry infills with the frame itself
remaining intact. Therefore, a portion of masonry infills was removed as shown in Fig. 43(b)
and then this model, defined as PIF, was again subjected to the same series of earthquake
simulation tests as the FIF.
The masonry infills can be beneficial to the seismic performance of the structure
since the amount of the increase in strength appears to be greater than that in the
induced earthquake inertia forces, while the deformation capacity of the global structure
remains almost same regardless of the presence of the masonry infills. The maximum
base shear of FIF, PIF, and BF under DE in Korea (TFT_012) was 32.0 kN, 37.3 kN,
and 17.6 kN, respectively, in Fig. 44(a). These are 2.5 to 5.3 times the design base
shear, 7.03 kN, according to the Korean seismic code. In Fig. 44(b), maximum
interstory drift indices (IDI) in the FIF and PIF models under the varying peak input
accelerations are shown and compared with those measured in the case of BF. The
drifts of the PIF are greater than those of the FIF under the same level of input ground
motions. However, IDI of neither FIF nor PIF exceeds the maximum value of 1.5%
allowed in the Korean seismic code even under TFT_04.

Frame A: Instrumented frame Frame A: Instrumented frame

Infill wall Infill wall

Frame B Frame B

(a) Shaking table test of FIF (b) Pushover test of PIF

Fig. 43 1:5-scale 3-story masonry-infilled RC frame model (Lee and Woo, 2002b)

2
FIF
PIF 1.68
Interstory drift index(%)

BF
1.6

the maximum allowable under design earthquake: 1.5%

1.2
1.08
TFT_012 (design EQ.) 0.77
0.8
0.51

0.4 0.26 0.28 0.3

0.24
0.111 0.188
0.106
0 0.042
TFT_012 TFT_02 TFT_03 TFT_04

(a) Base shear versus roof drift in tests (b) Change of maximum interstory drift

36
 Fig. 44 1:5-scale 3-story masonry-infilled RC frame model (Lee and Woo, 2002b)
4.3.3 1:5-scale 10-story RC Box-type Wall Building Structure Model
The number of apartment housing units is more than 58% of the total number of
housing units in Korea. These residential apartment buildings generally consist of high-
rise reinforced concrete (RC) wall structures, and should be designed and constructed
to resist the earthquake according to Korea Building Code (AIK 2005), and existing
buildings not satisfying these codes should be evaluated and retrofitted. The seismic
performance of the high-rise residential building model was evaluated based on the
results of earthquake simulation tests (Lee et al., 2012) and nonlinear time history
analyses (Hwang and Lee, 2015).
The prototype for the experiment was chosen to represent the most typical design in
Korea. The prototype was designed according to the old design code of Korea, AIK2000.
The thickness of walls is 180mm or 160mm with that of slabs being 200mm. The
reinforcement of the walls is two-layered, and the steel ratio of the vertical reinforcement
ranges from 0.34% to 0.90%, while the horizontal steel ratio is 0.29%. Considering the
capacity of the available shaking table and the feasibility of model reinforcements, a 1:5
scale 10–story building model was chosen (Fig. 45). To investigate the influence of the
slab, the analytical model without the slabs is also modeled. Model SB has both slabs
and coupling beams, and Model NS has only coupling beams without slabs.
The experimental and analytical models possessed a large overstrength (Fig. 46(a)).
Under the maximum considered earthquake (MCE) in Korea, the maximum base shear
coefficients of the experiment and the analysis are 0.206 and 0.17 in the X direction,
respectively, and 0.272 and 0.30 in the Y direction, respectively, which are 2.5~3.0
times larger than the seismic coefficients, Cs, respectively. In the results of the static
pushover analyses, the overstrength of the model with slabs, Ω, which is defined as the
ratio of the maximum strength of the fully-yielded system to the seismic coefficients, is
3.22 in the X direction and 4.2 in the Y direction. In the capacity curves, the lateral
strength dropped suddenly after the point of the peak resistance due to the shear
failure in the Y-directional outer walls. The overstrength of the model is larger than the
value of the overstrength factor, 2.5, given in KBC 2005 and IBC 2000. In Fig. 46(b),
under the DE in Korea, the maximum interstory drift ratio (IDR) in the analytical results
is 0.331% in the 6th story in the X direction and 0.195% in the 7th story in the Y
direction. It is comparable to that of test results, 0.307% in the 5th to 6th stories in the X
direction and 0.252% in the 9th to 10th stories in the Y direction, which satisfy the
allowable interstory drift ratio of 1.5% imposed by KBC 2005 (IBC 2000).

(a) Plan (b) Elevation (c) Shake-table test setup

37
 Fig. 45 1:5-scale 10-story RC Box-type Wall Building Structure model (Lee et al, 2012)
0.35 0.6SB, flexible-base SB, flexible-base
Base shear / Building weight

Base shear / Building weight

X-dir. (+) Y-dir.
SB, fixed-base (+) SB, fixed-base
0.3 0.5NS, flexible-base NS, flexible-base
NS, fixed-base NS, fixed-base
0.25 Experiment under MCE
0.4
0.2 Ω = 3.22 Analysis under MCE
0.3 Analysis under
0.15 Concepcion EQ.

0.1 Ω = 2.4 0.2

Ω = 4.2 Ω = 3.36 Steel, ε = 0.002m/m
Concrete, ε = 0.002m/m
0.05 Cs, design = 0.072 0.1 Concrete, εc,ult = 0.006m/m
Cs, design = 0.108 Shear stress degradation
0 0 in wall, ε = 0.01m/m
0 0.005 0.01 0.015 0.02 0.025 0 0.0025 0.005 0.0075 0.01 0.0125 0.015
Roof drift (ratio) Roof drift (ratio)
(a) Capacity curves
11
Roof 11
Flexible- Roof DE Flexible-
Model SB, Flexible-base Model SB, Flexible-base
DE Instant: 2.31s (max. roof drift (-X)) Instant: 10.81s (max. roof drift (-X))
in Korea base in Korea base under MCE in Korea under 2010 Concepcion earthquake
9 (X-dir.) Fixed- 9 (Y-dir.) Fixed-
base base
Exp. Exp.
7 7
Floor

Floor

5 5

3 3

1 1
-0.6 -0.3 0 0.3 0.6 -0.6 -0.3 0 0.3 0.6
Drift (%) Drift (%)
(b) Envelope of interstory drift under DE Y1 Y2 Y3 Y4Y5 Y6Y7 Y8 Y9 Y10 Y1 Y2 Y3 Y4Y5 Y6Y7 Y8 Y9 Y10

5 5
Flexible-base Y2 Flexible-base Y2
MCE Y4 Concepcion Y4
in Korea Y7 EQ. Y7
4 (2.31s) Y9 4 (10.81s) Y9
Floor

Floor
3 3

2 2
-0.00119 0.00215 -0.0154 0.0236

1 1
-0.003 -0.0015 0 0.0015 0.003 -0.03-0.02-0.01 0 0.01 0.02 0.03
Axial Strain (m/m) Axial Strain (m/m)
10 10 0.0250
Stress (MPa)
Stress (MPa)

0.0015
0 0

-10 -10
-0.0012 X4Y4
X4Y4 -0.0155 9s to 12s
-20 -20
-0.004-0.002 0 0.002 0.004 -0.03-0.015 0 0.015 0.03
Strain (m/m) Strain (m/m)
(c) Relations of hysteretic curves between (d) Distribution of plastic hinges and axial
base shear and roof drift under DE and strain of inner walls under MCE and
MCE Concepcion EQ.

(e) Crack patterns in exterior walls

Fig. 46 Experimental and analytical results of 1:5-scale 10-story RC box-type wall
building structure model (Lee et al. 2012 and; Hwang and Lee 2015)

In the test results, outer walls have many horizontal cracks at the lower stories
subjected to a large membrane force (Fig. 46(e)). In the analytical model, the axial

38
strains of wall boundaries at various locations are measured. Under the MCE in Korea,
the maximum axial strain demands of the wall boundaries in the lower part of the first
story are within 0.006m/m in tension and 0.0012m/m in compression (Fig. 46(d)). The
tensile strains in the outer walls are larger than the value of steel yield strain, 0.002m/m,
which are consistent with the horizontal cracks in the experiment. The probability of the
damage due to the concrete crushing and rebar buckling is very low under the MCE in
Korea.
During the 2010 Concepcion, Chile earthquake (Mw 8.8), the main observed
damage to slender walls was concrete spalling in unconfined elements and buckling
and fracture of the reinforcement. Under this earthquake, the total dissipated energy is
approximately 10 times larger than that under MCE in Korea. The maximum tensile and
compressive strains, 0.0252m/m and 0.0154m/m, respectively, occurred at the wall
boundaries, which indicates a potential for severe damage due to the concrete spalling
and reinforcement buckling at the walls.
4.3.4 1:15-scale 25-story RC Flat-Plate Core-Wall Building Model
Recently, the number of high-rise buildings (higher than 30 stories) has been
increasing, for the efficient use of available housing site. For the high-rise buildings, a
combined system of core shear walls: a lateral load resistance structural system, and
flat-plates: a gravity load resistance structural system, has been widely used. These
structural types in current seismic provisions, KBC2009 and IBC2006, are classified as
dual frame or building frame system. For the shear walls in the building frame system,
special shear walls, for which special seismic detailing requirements are imposed, or
ordinary shear walls, which have a height restriction, have generally been used. Lee et al.
(2015) investigated the seismic characteristics of this structure through shaking table
tests on 1:15 scale 25-story RC flat-plate core-wall building mode (Fig. 47).
28500 1900
8100

540

10200 680
Prototype building 1:15 scale model
Height : 79.5 m Height : 5.3 m
5400 5400

360 360
5400

Column : 900 × 900mm Column : 60 × 60 mm

28500

27000
11400

1900

1800
2000

760

Slab thickness : 300mm Slab thickness : 20 mm

3400

3575 3575
Wall thickness : 600mm Wall thickness : 40 mm
f’c = 40 MPa f’c = 40 MPa
8100

540

2450 150
fy = 400 MPa fy = 400 MPa
Y Y
750 8700 9600 8700 750 50 580 640 580 50
27000 1800
X X

(a) Prototype (b) Plan of protype buidling and 1:15 scale model

10400 600 3375

600

250
25-D29@400

21-D16@250

37-D16@125

Y
10200
11400

X
600

6-D16@400
2000
460

600

2-D29@400 9-D29@400 2-D29@460 9-D29@400

3975 2450 3975 3975

(c) Overview of the (d) Details of core wall and rebar fabrications of the core wall
39
shaking table test setup in the 1:15 scale model
Fig. 47 1:15-scale 25-story RC flat-plate core-wall building model (Lee et al. 2015)
In Fig. 48(a), under the design earthquake in Korea (DE, 0.187XY), the base shear
coefficients were 0.0361 in the X direction and 0.0518 in the Y direction, which are 1.5-
and 2-fold larger than the design base shear coefficient of 0.0253, respectively. The
strength increased gradually with the significant decrease of stiffness, and a large over-
strength occurred (Fig. 48(b)). Under the DE (0.187XY), the maximum inter-story drift
ratio was 0.31% from the 10th to 13th stories in the X direction and 0.30% from the
18th to 21th stories in the Y direction in Fig. 48(c), which satisfy the allowable inter-
story drift ratio of 1.5% imposed by KBC 2009 (IBC 2006).
Base shear coefficient, Cs

Base shear coefficient, Cs

0.1 0.1
XY Excitation X-dir. XY Excitation Y-dir.
0.08 X Excitation 0.08 Y Excitation
0.4g 0.4g
0.3g
0.06 0.06 0.154g0.187g 0.3g
0.154g 0.187g
0.04 0.04 Ω = 2.05
Ω = 1.43
0.02 0.07g Cs, design = 0.0253 0.02 0.07g Cs, design = 0.0253
0.035g 0.035g
0 0
0 20 40 60 0 20 40 60
Roof displacement (mm) Roof displacement (mm)
(a) Correlation between maximum roof drift and base shear coefficient
100 100 100 100
0.07XY X-dir. 0.187XY X-dir. 0.3XY X-dir. 0.4XY X-dir.
Base shear (kN)

Base shear (kN)

Base shear (kN)

Base shear (kN)

Vmax = - 23.9kN Vmax = 41.9kN Vmax = - 67.9kN Vmax = - 70.5kN
50 50 50 50

0 0 0 0
Table Table Table Table
-50 Excitation -50 Excitation -50 Excitation -50 Excitation
No No No No
k = 4.71 kN/mm Excitation k = 2.36 kN/mm Excitation k = 1.61 kN/mm Excitation k = 0.97 kN/mm Excitation
-100 -100 -100 -100
100 -60 -30 0 30 60 100 -60 -30 0 30 60 100 -60 -30 0 30 60 100-60 -30 0 30 60
0.07XY Y-dir. 0.187XY Y-dir. 0.3XY Y-dir. 0.4XY Y-dir.
Base shear (kN)

Base shear (kN)

Base shear (kN)

Base shear (kN)

VRoof
max
displacement (mm)
= 26.5kN VRoof
max
displacement (mm)
= 60.2kN VRoof
max
displacement (mm)
= 79.9kN VRoof
max = displacement
83.3kN (mm)
50 50 50 50

0 0 0 0
Table Table Table Table
Excitation Excitation Excitation Excitation
-50 -50 -50 -50
No No No No
k = 5.26 kN/mm Excitation k = 3.67 kN/mm Excitation k = 2.55 kN/mm Excitation k = 1.79 kN/mm Excitation
-100 -100 -100 -100
-60 -30 0 30 60 -60 -30 0 30 60 -60 -30 0 30 60 -60 -30 0 30 60
Roof displacement (mm) Roof displacement (mm) Roof displacement (mm) Roof displacement (mm)
(b) Hysteretic relation of the base shear and roof displacement
30 0.187XY 30 30 30
28 No Excitation 28 No Excitation 28 Max. Roof Accel. Table Excitation 28 Max. Roof Accel. Table Excitation
Roof
26 0.3XY Roof
26 Roof
26 Roof
26
24 0.4XY 24 24 X-dir. 24 Y-dir.
2222 22F
22 22
22 22F
22
20 20
1.5%18F 1.5% 20 20
Story
Story

Story

Floor

1818 18 18
18 -34.8 38.5 18F 18 49.3
16 16 16 16 -77.5
1414 14F
14 14
14 14F
14
12 12 12 12
1010 10F
10 10
10 10F
10
8 8 8 8
66 6F
6 0.187XY 66 0.187XY 6F 6 0.187XY
4 4 4 0.3XY 4 0.3XY
2 X-dir. 2 Y-dir. 0.3XY 2 2
0 0 0.4XY 0
0.4XY 0
0.4XY
-0.02 -0.01 0 0.01 0.02 -0.02 -0.01 0 0.01 0.02 -90 -60 -30 0 30 60 90 -90 -60 -30 0 30 60 90
Interstory drift ratio (rad) Interstory drift ratio (rad) Shear force (kN) Shear force (kN)

(c) Envelope of interstory drift ratio (d) Distribution of story shear

Vx = 58.8 kN 60
X-dir. (0.3XY) Vy = 12.8 kN
1.0 DL
DL/3.59
Time = 14.04 sec εy = 0.002
εy
인장변형(+)(+)
Tension 0.00094 0.0078 30 εc = 0.003
εc
Moment (kNm)

εc
εc = 0.006
-0.00018 φx-dir.=0.0085
압축변형(-) 0.0013 0
Compression (-) φx-dir.= 0.034rad/m φy = 0.0104 φcl = 0.019
φu = 0.041
φx-dir. = 0.0085rad/m (tension) -30 φx-dir. = 0.034
Y-dir. (compression)
0.0014
Short wall
-0.00018 -60
0.0058 -0.15 -0.075 0 0.075 0.15
X-dir. 0.00078
Curvature (rad/m)

(e) Strain distribution of the core wall (f) Relation of the moment and curvature (M-φ)
at the bottom of the first story under MCE in core wall (X-dir.)

40
 Fig. 48 Shake-table test results of a 1:15-scale 25-story RC flat-plate core-wall building
model (Lee et al. 2015)
The model displayed behavior in the first mode during free vibration after the
termination of excitation, and the maximum values of the base shear and roof drift in
this duration can be either similar to or larger than the values of the maximum
responses during the table excitation. The higher modes were observed in both the X
and Y directions in the vertical distribution of story shear. When the roof acceleration
reached a maximum, the effect of the second and third modes governed, and the
largest story shear was apparent from the 14th to 21st stories instead of the first story
(Fig. 48(d)).
In accordance with the displacement-based design method proposed in ACI 318-05,
special boundary details were imposed on the short wall in the first story with the
expected plastic rotation of θp = 0.00537 rad (Fig. 48(e)). No significant plastic
deformation was observed under the MCE in Korea. At the bottom 70 mm of the first
story, the measured maximum curvature when the end of the boundary element in the
short wall is in compression is φx-dir. = 0.0085 rad/m, which is approximately 21% of
0.041 rad/m, the ultimate curvature corresponding to the expected compressive strain
of 0.00638 m/m (Fig. 48(f)). This result, together with the findings mentioned above,
implies that the design requirements on the boundary elements of the walls given in
ACI 318-05 may be overly conservative, particularly for the wall design of high-rise RC
building frames or dual-frame structures with more than 20 stories.

5. SUMMARY AND CONCLUSIONS

5.1 Summary
Characteristics of earthquake ground motions in moderate seismicity regions are as
follows:
• The probability of collapse can be very low in moderate seismicity regions with that
of non-structural damage being very high.
• The damages were concentrated to the region within the short epicentral distance.
• The duration is relatively short, so resonance effect can be minor. And, intensity of
high-frequency contents is very high at near field but decays very rapidly as the
epicentral distance increases.
• Spectral accelerations of high frequency are very high and can cause the brittle
failure such as shear failure of short columns and crushing of window glasses.
• Spectral displacement can be significantly small when compared to the spectral
acceleration. Therefore, flexible structures generally have a low probability of large
inelastic excursions.
• The impact of high-frequency ground motions to the lower-frequency structures
can cause non-vibratory unidirectional overload to the shear-critical members such
as short columns.
• Typical building structures in a moderate seismicity region such as South Korea,
which were not designed seismically, have retained a large overstrength, so it is
not reasonable to assume all the non-seismically designed building structures

41
 would collapse as many media rouse the public to the unjustified fear.

The success of PBEE in high-seismicity regions as well as in moderate-seismicity

regions depends on how we can estimate the actual behavior and loss reasonably.
That is, every assumption in the analysis should be verified with experimental
observation of structures. Estimation of earthquake load should be ascertained by the
real records of EQ ground motions and the rationale for extrapolation of the ground
motion prediction equation to the maximum magnitude earthquake should be provided.
The followings are considered to be prerequisites for the success of PBEE in any cases.
• Estimation of actual seismic demands on the structural and nonstructural
responses can be possible only by the provision of seismic hazard curves for all
structures. Also, guides to input ground motions to be used for the linear and
nonlinear analysis should be provided.
• Database of existing structures regarding design, construction, and maintenance
should be established. Moreover, database of the mathematical behavior models,
resistance capacity for all kinds of major structural and non-structural elements
should be set up with their probability distributions.
• The linear and nonlinear behavior models of elements and joints should be verified
through experiments, and the reliability in the prediction of overall structural
behaviors using these models should be confirmed. Also, the user of the nonlinear
software should have a full understanding of the nonlinear analysis and the
limitations of the used software.
• Database should be set up for the derivation of fragility curves of structural and
nonstructural elements and for estimation of economic losses due to the damages.
• To ensure the true realization of PBEE, the inspection process on the quality of
design, construction, and maintenance should be established. For this, a reliable
system of peer review should be provided.

5.2 Conclusions
• PBEE can be used as a tool to evaluate the appropriateness of the existing
seismic code, which was developed mainly for the high-seismicity regions, and to
adapt this code to the moderate-seismicity regions. To do this, first, the design of
structures according to the requirements of the current codes, second, perform
first- and second-generation PBEE on these designed structures. For example,
each building structures (infilled masonry, or masonry structure, RC moment frame,
steel moment frame, wall structure, dual structure, so on) designed exactly per the
current prescriptive seismic codes are evaluated using PBEE procedure. Based on
these results, appropriateness of performance factors such as R, Cd and Ω will be
verified regarding the actual behaviors through PBEE procedure. Also, the
maximum deformations in moderate-seismicity regions are estimated with the
probability distribution and used to determine the appropriate requirements for
seismic details, which will clearly lead to the alleviation of requirements for seismic
details made mainly for the high-seismicity regions.
42
• September 12, 2016, Gyeongju earthquake has provided valuable data of
earthquake ground motions representing the moderate-seismicity region. By
analyzing and utilization of these data, it becomes possible to establish the
seismological model in Korean Peninsula. It is necessary to build up the seismic
hazard map appropriate for Korean Peninsula by simulating the earthquake ground
motion with this developed the seismological model and the probability theory. The
research on the faults in Korean Peninsula should cover not only paleoseismic
geological study on the faults developed over several million years, but also
provide the information on the faults behaviors which occurred within Holocene
period (11,000 years), including the return period of 2,500 years, 500 years, and
much shorter durations because this information only can make a meaningful
contribution to seismic design and retrofit.
• The earthquake tectonics in Korean Peninsula does not belong to the plate
boundary or the plate boundary related intraplate, but belongs to the category of
intraplate or mid-plate regions, whose slip rate is less than 0.1 mm/year. The
earthquake in these regions are called small earthquakes whose maximum
magnitudes will generally be 6.0~6.5, and do not show the surface ruptures with
hidden faults. Because the historical catalog over the past 2,000 years in Korea
cannot be used reliably to predict the maximum magnitude earthquakes in the
return period of 2,500 years, it seems more reasonable to determine the design
earthquake having the return period as short as possible, such as 500 years (10%
probability of exceedance in 50 years).
• Because any building structure retains some minimum level of earthquake
resistant capacity, it is a good approach to evaluate this level of resistance and to
use this information for the seismic strengthening for the target maximum
earthquake. Though there has been almost no severe earthquake disaster over the
past several centuries in Korea, the news of devastated cities around the world
due to the severe earthquakes might cause unjustified fears to Korean people and
lead to over- or unnecessary design and construction, which should be avoided
anyway.
• One example of the over- or unnecessary design and construction may be the use
of dampers to retrofit low-rise school buildings in Korea. As shown on Sept. 12,
2016, Gyeongju earthquake, the characteristics of the near-source earthquake
(maximum magnitude Mw ≤6.5) in moderate-seismicity regions can be described
as an impulsive load. However, the efficacy of damper for this type of load is
questionable and should be reevaluated and, if the response of the structure
appears to be unsatisfactory, redesign and reconstruction should be conducted.
• Also, low-rise and high-frequency structures, subjected to a very high impulsive or
implosive earthquake load due to the near-source earthquake, can lead to brittle
shear failure of the critical beams and columns. Special design requirements to
ensure the safety against this failure should be developed.
• Although the probability of collapse of building structures appears to be very low in
moderate-seismicity regions, the failure of windows, dislocation of ceilings and
falling of roof tiles were shown to be highly probable. Since a major portion of the
economic loss is due to these non-structural failures, it is necessary to develop

43
appropriate design requirements specific to the moderate-seismicity regions.

44
ACKNOWLEDGMENTS

The research presented herein was supported by the National Research Foundation of
Korea (NRF-2009-0078771), the Ministry of Land, Infrastructure and Transport of Korea
(17AUDP-B066083-05), the Ministry of Public Safety and Security of Korea (MPSS-NH-
2013-70), and the Korea University Grant. The authors are grateful for these supports.

REFERENCES

References are listed alphabetically for each part as follows:

Building codes and guidelines:

ACI Committee 318 (2005) Building code requirements for structural concrete and
commentary (ACI 318-05), American Concrete Institute, Detroit.
AIK (2000), AIK 2000, Architectural Institute of Korea (AIK), Seoul, Korea. (in Korean)
AIK (2005), Korean Building Code. KBC 2005, Seoul, Korea. (In Korean)
AIK (2009), Korean Building Code. KBC 2009, Seoul, Korea. (In Korean)
AIK (2016), Korean Building Code. KBC 2016, Seoul, Korea. (In Korean)
ASCE (2010), Minimum design loads for buildings and other structures, ASCE/SEI 7-10,
American Society of Civil Engineers (ASCE), Reston, Virginia, US.
ASCE (2013), Seismic evaluation and retrofit of existing buildings, ASCE 41-13,
American Society of Civil Engineers (ASCE), Reston, Virginia, US.
ATC (1996a), Improved Seismic Design Criteria for California Bridges: Provisional
Recommendations: ATC-32. National Bureau of Standards: Washington DC, 1996.
ATC (1996b), Seismic evaluation and retrofit of existing concrete buildings: ATC-40,
Applied Technology Council (ATC), Redwood City, CA.
FEMA (1996), NEHRP guidelines for the seismic rehabilitation of buildings, FEMA 273
Commentary, Federal Emergency Management Agency (FEMA), Washington DC.
FEMA (2000), Prestandard and commentary for the seismic rehabilitation of buildings.
Report No. FEMA-356, Washington, DC, 2000.
FEMA (2012), Seismic performance assessment of buildings. Report FEMA P-58,
Federal Emergency Management Agency, Washington DC, U.S.A.
ICC (2000) International Building Code, IBC 2000, International Code Council, Country Club Hills, IL.
ICC (2006) International Building Code, IBC 2006, International Code Council, Country Club Hills, IL.
SEAOC (1995), Vision 2000, Performance based seismic engineering of buildings, Vols.
I and II: Conceptual framework, Structural Engineers Association of California
(SEAOC), Sacramento, CA.

Chapter 1:
Par. 2, Page 1 ~ Par. 2, Page 2 excerpted from:
Günay, S., & Mosalam, K. M. (2013) PEER performance-based earthquake
engineering methodology, revisited. Journal of Earthquake Engineering, 17(6), 829-
858.

45
Chapter 2:
Par. 4, Page 2 excerpted from:
Krawinkler, H., & Miranda, E. (2004). Performance-based Earthquake Engineering.
Chapter 9 of Earthquake Engineering: from Engineering Seismology to
Performance-based Engineering. Bozorgnia and VV Bertero, Editors, CRC Pres.
Par. 5, Page 2 ~ Par. 2, Page 3 excerpted from:
Porter, K. A. (2003). “An overview of PEER’s performance-based earthquake
engineering methodology,” Proceedings of ninth international conference on
applications of statistics and probability in civil engineering.
Par. 1, Page 4 ~ Par. 6, Page 6 excerpted from:
Moehle, J., & Deierlein, G. G. (2004). “A framework methodology for performance-
based earthquake engineering”, The 13th world conference on earthquake
engineering, Paper No. 679, pp. 3812-3814, August 1-6, Vancouver, B.C., Canada.

Chapter 3:
Pars. 1~3, Page 7 / Pars. 1~3, Page 8 excerpted from:
Scholz, C. H. (2002). The mechanics of earthquakes and faulting. Cambridge university
press.
Pars. 1~5, Page 9 excerpted from:
Nordenson, G. J., & Bell, G. R. (2000), “Seismic design requirements for regions of
moderate seismicity”, Earthquake spectra, 16(1), 205-225.
Pars. 1~3, Page 10 excerpted from:
Frankel, A. (1995). “Mapping seismic hazard in the central and eastern United
States”. Seismological Research Letters, 66(4), 8-21.
Pars. 1~2, Page 14 excerpted from:
Houng, S.E., & Hong, T.K. (2013). “Probabilistic analysis of the Korean historical
earthquake records,” Bulletin of the Seismological Society of America, 103(5), 2782-
2796.
Pars. 1~3, Page 16 excerpted from:
Hong, T.K., Lee, J.H., Kim, W.H., Hahm, I.K., Woo, N.C., & Park, S.J. (2017). “The 12
September 2016 ML 5.8 midcrustal earthquake in the Korean Peninsula and its
seismic implications,” Geophysical Research Letters. 44(7), 3131-3138.
Others:
Bakun, W. H., and M. G. Hopper (2004). “Magnitudes and locations of the 1811-1812
New Madrid, Missouri and the 1886 Charleston, South Carolina, earthquakes,” Bull.
Seismol. Soc. Am. 94, 64–75.
Boore, D. M., Atkinson, G. M. (2008) “Ground-Motion Prediction Equations for the
Average Horizontal Component of PGA, PGV, and 5%-Damped PSA at Spectral
Periods between 0.01 s and 10.0 s,” Earthquake Spectra. 24(1), pp 99-138.
Boore, D. M., Campbell, K. W., & Atkinson, G. M. (2010). “Determination of stress
parameters for eight well-recorded earthquakes in eastern North America,” Bulletin
of the Seismological Society of America. 100(4), 1632-1645.
Cornell, C. A. (1968) “Engineering seismic risk analysis,” Bulletin of the Seismological
Society of America. 58(5), pp. 1583-1606.
46
Fardis, M.N. (2014) Comments on the seismic design provisions of the Korean Building
Code 2009. (Opinion paper)
Johnston, A. C. (1996). “Seismic moment assessment of earthquakes in stable
continental regions—III. New Madrid 1811-1812, Charleston 1886 and Lisbon 1755,”
Geophys. J. Int., 126, 314–344.
Korean Broadcasting System (KBS), available from: http://news.kbs.co.kr/news/view.do?ncd=3366101,
last accessed 29 June 2017. (in Korean)
Korea Meteorological Administration (KMA), available from: http://www.kma.go.kr/, last
accessed October 2016. (in Korean)
Kramer, S. L. (1996). Geotechnical Earthquake Engineering, Prentice Hall. New York.
Lam, N. (2014) Displacement based assessment of structures for low and moderate
seismic regions (Opinion paper).
Lee, C.H. (2017) “Earthquake Engineering Analysis of Ground Accelerations Measured
in the 912 Gyeong-ju Earthquake,” Journal of the Koran society of Civil Engineers,
65(4), 8-13. (in Korean)
McGuire, R. K. (1976) “Probabilistic seismic hazard analysis and design earthquakes:
closing the loop,” Bulletin of the Seismological Society of America, 85(5), pp. 1275-1284.
NEMA (2012), Active fault map and seismic hazard map. National Emergency
Management Agency, Report No. NEMA-NH-2009-24. (in Korean)
Ohmynews, available from: http://www.ohmynews.com/NWS_Web/View/at_pg.aspx?CNTN_CD=A0002243576,
last accessed 29 June 2017. (in Korean)
USGS, available from: https://earthquake.usgs.gov/earthquakes/eventpage/us20005iis#finite-fault,
last accessed 29 June 2016. (in Korean)
Yonhapnews, available from:
http://www.yonhapnews.co.kr/bulletin/2016/09/19/0200000000AKR20160919070351053.html, last accessed 29 June
2017. (in Korean)
YTN, available from: http://www.ytn.co.kr/_ln/0115_201609131800184070, last accessed
29 June 2017. (in Korean)

Chapter 4:
Chung, K.R., Chung, H.J., Kang, M.S., Kim, S.H., and Park, K.M. (2013) "Eliminating
special seismic boundary of special shear wall system using NLTHA." Korea Concrete
Institute Conference, 2013 Fall. Sokcho, Korea: Korea Concrete Institute. (in Korean)
Ji, J., Elnashai, A.S., & Kuchma, D.A. (2009). Seismic fragility relationships of
reinforced concrete high‐rise buildings. The Structural Design of Tall and Special
Buildings, 18(3), 259-277.
Hwang, K. R., & Lee, H. S. (2015). Seismic performance of a 10-story RC box-type wall
building structure. Earthquake and Structures, 9(6), 1193-1219.
Lee, H. S., and Woo, S. W. (2002a) “Seismic performance of a 3-story RC frame in a
low-seismicity region”, Engineering Structures, 24(6), 719-734.
Lee, H. S., and Woo, S. W. (2002b). “Effect of masonry infills on seismic performance
of a 3‐storey R/C frame with non‐seismic detailing”, Earthquake Engineering &
Structural Dynamics, 31(2), 353-378.
Lee, H. S., Hwang, S. J., Lee, K. B., Kang, C. B., Lee, S. H., and Oh, S. H. (2012).
“Earthquake Simulation Tests on a 1: 5 Scale 10-Story RC Residential Building Model,”
The 15th World Conference on Earthquake Engineering (15WCEE), Lisbon, Portugal.
47
Lee, H. S., Hwang, K. R., & Kim, Y. H. (2015). “Seismic performance of a 1: 15‐scale
25‐story RC flat‐plate core‐wall building model,” Earthquake Engineering &
Structural Dynamics, 44(6), 929-953.

48