See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/225364881

Probabilistic Fragility Analysis: a Tool for Assessing Design Rules of RC

Buildings,” Earthquake Engineering and Engineering Vibration, 7(1): 45-56

Article  in  Earthquake Engineering and Engineering Vibration · March 2008

DOI: 10.1007/s11803-008-0823-x

CITATIONS READS

36 446

1 author:

Nikos Lagaros
National Technical University of Athens
224 PUBLICATIONS   3,685 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

OCP: Optimization Computing Platform View project

Holistic, High Performance Optimization Computing Platform (HP-OCP) View project

All content following this page was uploaded by Nikos Lagaros on 22 May 2014.

The user has requested enhancement of the downloaded file.

 PROBABILISTIC FRAGILITY ANALYSIS: A TOOL FOR

ASSESSING DESIGNED RULES OF RC BUILDINGS

NIKOS D. LAGAROS

Institute of Structural Analysis & Seismic Research,

National Technical University of Athens,
9, Iroon Polytechniou Str., Zografou Campus, 157 80 Athens, Greece
Tel.: +30 210 772 2625
Fax: +30 210 772 1693
e-mail: nlagaros@central.ntua.gr

Abstract

In this work fragility analysis is performed in order to assess two groups of reinforced concrete structures. The first

group of structures is composed by buildings implementing three, often encountered, design practices namely fully

infilled, weak ground storey and short columns. The three design practices are applied during the design process of a

reinforced concrete building. The structures of the second group vary on the value of the behavioural factor used to

define the seismic forces in prescriptive design procedures. Most of the seismic design codes belong to the class of

prescriptive procedures where if certain constraints are fulfilled the structure is considered safe. Prescriptive design

procedures express the ability of the structure to absorb energy through inelastic deformation using the behaviour

factor. The basic objective of this work is to assess both groups of designs with reference to the limit-state probability

of exceedance. Thus, for both groups of structures four limit state fragility curves are developed on the basis of non

linear static analysis. Moreover, the 95% confidence intervals of the fragility curves are also calculated, taking into

account two types of random variables influencing the structural capacity and the seismic demand.

Keywords: Probabilistic fragility analysis, design practice, behaviour factor, Monte Carlo simulation method

1
INTRODUCTION

Over the last decades the risk management of structural systems have gained the attention of

various economic and technical subjects in modern society. The optimal allocation of the public

resources for a sustainable economy entails the need of rational tools for estimating the

consequences of natural hazardous events on the built environment. The risk management

addresses this claim indicating the way for optimal choices. Thus, the main purpose of the risk

management process is to choose among different options relying on technical and economic

considerations. Risk assessment and decision analysis are the main steps of the risk management

concept. It is therefore essential to establish a reliable procedure for assessing the seismic risk of

the structural systems. Seismic fragility analysis, which provides a measure of the safety margin

of the structural system above specified hazard levels, is considered as the core of the risk

assessment.

A number of methodologies for performing fragility analysis have been proposed in the past

which have been used for assessing the behaviour of structural systems. Kennedy et al. [1]

presented a methodology for determining the probability of earthquake induced radioactive

releases as a result of core melt. Kircher et al. [2] described building damage functions that were

developed for the FEMA/NIBS earthquake loss estimation methodology [3]. Shinozuka et al. [4]

presented a statistical analysis procedure of structural fragility curves. The significance of

inherent randomness and modelling uncertainty in forecasting the building performance was

examined by Ellingwood [5] through fragility assessment of a steel frame. The importance of

fragility analysis in various stages of risk assessment, loss estimation, and decision making in

consequence based engineering to achieve the desirable long-term objective in reduction of loss

and consequence with the most efficient intervention measures was indicated in [6]. A procedure

to account for the uncertainty in the characteristics of future ground motions during seismic

response assessment was presented in the work by Aslani and Mirand [7]. Fragility functions
2
were developed in [8] to identify the method of repair required for older reinforced concrete

beam-column joints damaged due to earthquake loading. A methodology for the risk assessment

of reinforced concrete and unreinforced masonry structures was presented in [9]. Jeong and

Elnashai [10] presented an approach where a set of fragility relationships with known reliability is

derived based on the fundamental response quantities of stiffness, strength and ductility. A set of

procedures for creating fragility functions from various kinds of data was introduced in [11]. In

the work by Shinozuka et al. [12] bridge fragility curves are developed in order to determine the

effect earthquakes have on the performance of transportation network systems.

The majority of the reinforced concrete (RC) buildings are constructed with masonry infill

walls. However, the combination of the masonry infills with the framed structure is most often

neglected during the design procedure, assuming that the contribution on the structural

performance is always positive. Such an assumption may lead to substantial inaccuracy in

predicting the lateral stiffness, strength and ductility of the structure. In a number of works

[13,14] has been studied the effect of weak ground storeys on the seismic performance of RC

frames. On the other hand short columns at the ground storey of the structures are prone to brittle

shear failure which may result in severe damages or even collapse because of the poor ductility

during earthquakes [15,16].

All modern seismic design procedures are based on the principal that a structure will avoid

collapse if it is designed to absorb and dissipate the kinetic energy that is imparted in it during the

seismic excitation. Most of the modern seismic codes express the ability of the structure to absorb

energy through inelastic deformation using the behaviour factor. The capacity of a structure to

resist seismic actions in the nonlinear range generally permits their design for seismic loads

smaller than those corresponding to a linear elastic response. The seismic loads are reduced using

the behaviour factor. The numerical confirmation of the behavior factor became a subject of

research work during the past decade [17,18] in order to check the validity of design theory

3
assumptions and to make structural performance more predictable from engineering point of

view.

The main objective of this study is to assess the way that the design practices are

implemented and more specifically the weak ground storey and short columns design practices

and to examine the influence of the behavioural factor used to define the seismic design forces on

the design obtained. Therefore, two groups of structures are compared with reference to the

limit-state fragilities developed in four drift based limit states. Moreover, statistical analysis is

performed on the fragility curves defining the 95% confidence intervals considering randomness

in both structural capacity and seismic demand.

FRAGILITY ANALYSIS

Earthquake risk assessment of building structures requires the calculation of limit-state

probabilities for a series of limit-states of monotonically increasing severity. The target is to

obtain the limit-state probabilities of exceedance that serve as a hazard curve for structural

damage. The mean annual frequency of the maximum interstorey drift θmax exceeding the value y

is obtained as:

P( θmax ≥ y ) = ∫ P ( θmax ≥ y IM = x ) dλΙΜ ( x ) (1)

where P ( θmax ≥ y IM = x ) is the probability that θmax, exceeds the value y given that IM equals x

and λΙΜ(x) is the mean annual frequency of the chosen intensity measure exceeding x, or in other

words λΙΜ(x) is the hazard curve and dλΙΜ(x) is its slope. The absolute value is used because the

slope has a negative value.

Building fragility curves are lognormal functions that describe the probability of reaching

or exceeding a specific limit state. The conditional probability of being, or exceeding, a particular

4
damage state y given peak ground acceleration PGA (or other seismic demand parameter) is

defined by:

⎡ 1 ⎛ PGA ⎞ ⎤
P ( θmax ≥ y | PGA ) = Φ ⎢ ln ⎜ ⎟⎥ (2)
⎣⎢ β y ⎝ PGA, y ⎠ ⎦⎥

where PGA, y is the median value of peak ground acceleration at which the building reaches the

threshold of damage state y, βy is the standard deviation of the natural logarithm of peak ground

acceleration for the damage state y and Φ is the standard normal cumulative distribution function.

SEISMIC DESIGN PROCEDURES

The majority of the seismic design codes belong to the category of the prescriptive building

design codes, which include: site selection and development of conceptual, preliminary and final

design stages. According to a prescriptive design code the strength of the structure is evaluated at

one limit state between life-safety and near collapse using a response spectrum corresponding to

one design earthquake [19,20]. In addition, serviceability limit state is checked in order to ensure

that the structure will not deflect or vibrate excessively during its functioning. The main principle

of new provisions, EAK 2000 [19] and Eurocode 8 [20] included, is to design structural systems

based on energy dissipation and on ductility in order to control the inelastic seismic response.

Designing a multistory RC building for energy dissipation comprises the following features: (i)

fulfillment of the strong column/weak beam rule, (ii) member verification in terms of forces and

resistances for the ultimate strength limit state under the design earthquake (with return period of

475 years, probability of exceedance 10% in 50 years), with the elastic spectrum reduced by the

behaviour factor, (iii) damage limitation for the serviceability limit state and (iv) capacity design

of beams and columns against shear failure.

According to the Greek national design codes and Eurocodes a number of checks must be

considered in order to ensure that the structure will meet the design requirements. All EKOS 2000
5
[21] or Eurocode 2 [22] checks must be satisfied for the gravity loads using the following load

combination

S d = 1.35∑ j Gkj "+ "1.50∑ i Qki (3)

where “+” implies “to be combined with”, the summation symbol “Σ” implies “the combined

effect of”, Gkj denotes the characteristic value “k” of the permanent action j and Qki refers to the

characteristic value “k” of the variable action i. If the above constraints are satisfied, multi-modal

response spectrum analysis is performed, according to EAK 2000 [19] and Eurocode 8 [20], and

earthquake loading is considered using the following load combination

S d = ∑ j Gkj "+ " Ed "+ "∑ iψ 2i Qki (4)

where Ed is the design value of the seismic action for the two components (longitudinal and

transverse) respectively and ψ2i is the combination coefficient for the quasi-permanent action i,

here taken equal to 0.30.

DESIGN PRACTICES

The behaviour of a building during an earthquake depends critically on its overall shape, size and

geometry. A wide range of structural damages observed during past earthquakes across the world

has been very educative in identifying construction features related to the shape, size and

geometry of the structure that must be avoided. Buildings that have fewer columns or they are

fully infilled in some of the storeys or they have partially infilled storeys tend to be more

vulnerable to earthquake loading. A large number of multi-storey RC buildings collapsed in past

earthquakes, due to construction features like weak stories, short columns, strong beams-weak

columns, large and heavy overhangs and others [23-26].

6
Weak ground storey

RC building structures have become very popular during the last decades in urban Greece. Many

such buildings constructed in recent times have a special construction feature; the ground storey

is left open for the purpose of parking. Such buildings are also called weak ground storey

buildings, while the weak storey is also called soft storey or pilotis. Weak ground storey buildings

have shown poor performance during past earthquakes across the world [23,24]; while a

significant number of them have collapsed. The fully infilled upper storeys are much stiffer than

the open ground storey. Thus, the upper storeys deform almost together, and the maximum

interstorey drift occurs in the weak ground storey. Consequently, the columns in the open ground

storey are severely stressed. If the columns do not have the required strength to resist or do not

have adequate ductility, they might severely damaged which may lead to the collapse of the

building.

Short columns

During past earthquakes, RC buildings having short columns suffered from damages [25,26], due

to the concentration of large shear forces. The short columns are stiffer compared to the regular

size columns attracting larger earthquake forces. If a short column is not adequately designed for

such a large force, it can suffer significant damage during an earthquake. Short columns are

characterized by a small value of the shear span ration α:

Μ sd
α= ≤ 2.5 (5)
Vsd h

where Msd and Vsd are the maximum moment and shear force values obtained from the

combination of Eq. (4) while h is the column depth. Generally speaking the failure modes of short

columns can be classified into two cases: (i) shear failure, occurs when α ≤ 1.50 and (ii) sliding

failure when 1.5 - 2.0 ≤ α ≤ 2.5.

7
CONFIDENCE INTERVALS

In this work, in order to define the 95% confidence intervals for each fragility curve developed,

the procedure suggested by Shinozuka et al. [4] was followed. In particular the Monte Carlo

Simulation (MCS) method was used generating 10,000 simulations on the basis of the non-linear

static analysis in both test examples considered. Two sources of randomness have been

considered: on the ground motion excitation which influences the level of seismic demand, and on

the material properties which affect the structural capacity. The modulus of elasticity of concrete,

steel reinforcement and masonry infill along with the compressive strength, the yield strain and

the ultimate strain for both confined and unconfined concrete were taken as random variables

influencing the structural capacity. More detailed description of the random variables affecting

the structural capacity for both test examples are given in the numerical study section.

The randomness on the seismic excitation is considered through the response spectra. For

this purpose a set of natural records that consist of both longitudinal and transversal components

of the records, are used. The records were selected from the database of Somerville and Collins

[27] corresponding to the hazard level with 10 percent probability of being exceeded in 50 years

(Table 1). The records are scaled to the PGA value obtained from hazard curves originally derived

for Greece by Papazachos et al. [28]. Moreover, based on the assumption that seismic data follow

the lognormal distribution [29], the median spectrum x̂ and the standard deviation δ can be

calculated using the following expressions:

⎡ ∑ n ln( Rd,i (T )) ⎤
x̂ = exp ⎢ i =1 ⎥ (6)
⎢⎣ n ⎥⎦

12
⎡ n ( ln( R (T )) − ln(x)
ˆ ) ⎤
∑
2

δ=⎢ ⎥
i =1 d,i
(7)
⎢ n −1 ⎥
⎣ ⎦

8
where Rd,i(T) is the response spectrum value for period equal to T of the i-th record (i=1,…,n) and

the number of records used and varies for different hazard levels. Therefore, n = 19 for the 10% in

50 years. Both components of the median spectra can be seen in Figure 1.

NUMERICAL STUDY

Two test examples have been considered for performing fragility analysis. For both test examples

the lateral design forces were derived from the design response spectrum i.e. 5%-damped elastic

spectrum divided by the behaviour factor at the fundamental period of the building. Concrete of

class C16/20 (nominal cylindrical strength of 16MPa) and class S500 steel (nominal yield stress

of 500MPa) are assumed. The base shear is obtained from the response spectrum for soil type A

(stiff soil θ = 1, with characteristic periods Τ1 = 0.10s and Τ2 = 0.40s) for the first test example and

soil type B (stiff soil θ = 1.0, with characteristic periods Τ1 = 0.15sec and Τ2 = 0.60sec) for the

second one. The PGA value is equal to 0.31 g, corresponding to zone III for the 10/50 hazard

level. Greece is divided into three zones of equal seismic hazard. Papazachos et al. [28] have

presented a semi-probabilistic approach to the seismic hazard assessment of Greece resulting to

the hazard curves for all zones. The city of Athens which is the location where the two test

examples will be built belong to zone III. Moreover, the importance factor γI was taken equal to

1.0, while damping correction factor is equal to 1.0, since a damping ratio of 5% has been

considered (as it is suggested by EAK 2000 for RC structures).

For both test examples the slab thickness is equal to 15cm and is considered to contribute to

the moment of inertia of the beams with an effective flange width. In addition to the self weight of

the beams and the slab, a distributed dead load of 2 kN/m2 due to floor finishing and partitions is

considered, while live load with nominal value of 1.5 kN/m2 is also applied. In the combination of

gravity loads (“persistent design situation”) nominal dead and live loads are multiplied with load

factors of 1.35 and 1.5, respectively. Following EAK 2000, in the seismic design combination,

dead loads are considered with their nominal value while live loads with 30% of the nominal
9
values. For all test cases, a centreline model was formed based on the OpenSEES [30] simulation

platform. The members are modelled using the force-based fiber beam-column element while the

same material properties are used for all the members of the test cases examined. Soil-structure

interaction was not considered and the base of the columns at the ground floor is assumed to be

fixed.

Test example 1

The first test example is the 3D RC building shown in Figure 2 having four storeys first analyzed

by Lagaros [31]. The lateral design forces were derived from the design response spectrum

described above while a behaviour factor q equal to 3.5 is considered, as it is suggested by EAK

2000 for RC structures. The cross section for all columns is 45×45 cm2 and 30×60 cm2 for all

beams. The infill walls consist of 30cm×20cm×15cm horizontally perforated bricks with

compressive strength equal to 3.0 MPa and modulus of elasticity equal to 2250 MPa. In ELSA

laboratory [14] models, similar to the ones examined in this study, have been tested.

The model employed in this study for simulating the masonry infill panels is based on the

one developed by Perera et al. [32]. According to this model the contribution of the masonry infill

panel to the response of the infilled frame is modeled by a system of two diagonal masonry

compression struts. The two struts are considered ineffective in tension since the tensile strength

of masonry is negligible. The combination of both diagonal struts provides the lateral load

resisting mechanism for the opposite lateral directions of loading. Each strut element is modeled

as a simple longitudinal inelastic spring whose behaviour is described in terms of the axial

force-axial deformation relation of the strut using the notion and principles of continuum damage

mechanics. In the work by Perera et al. [32] the following relation was obtained:

N = K 0 (1- d)δe = K 0 (1- d)(δ - δ p ) (8)

10
where N is the axial force of the strut, δ, δe and δp are the total, elastic and plastic shortenings of

the strut, respectively. K0 is the initial stiffness before cracking and d is the internal damage

variable representing the degradation of the infill. More details about the damage model of the

masonry infill used in this work can be found in the work by Perera et al. [32].

Description of the test cases

For this test example three different cases have been examined: (a) fully infilled model that

corresponds to the design where all circumferential frames in all storeys are considered fully

infilled (see model of Figure 3a), (b) weak ground storey model, where no masonry infill are

present in the ground storey (see model of Figure 3b) and (c) short columns model, where

transverse frames C3-B12-C6-B11-C9 and C1-B8-C4-B7-C7 are fully infilled (see model of

Figure 3a) and longitudinal frames C1-B1-C2-B2-C3 and C7-B5-C8-B6-C9 are partially infilled

in the ground level (see model of Figure 3c). All models were designed to meet the EKOS 2000

[21] and EAK 2000 [19] requirements, implementing all the provisions suggested by the codes in

order to alleviate the effect of the design practices on the structural performance. The weight of

the steel reinforcement and the concrete volume required for the three models are given in Table

2. As it can be seen the concrete volume is the same since the cross section for all columns and

beams is equal to 45×45 cm2 and 30×60 cm2, respectively, all for the three designs. The difference

on the reinforcement of the columns and the beams is due to the implementation of the code

provisions for the various design practices. As it can be seen from Table 2 the fully infilled design

requires less reinforcement than any other design required, while the weak ground storey design

requires the most reinforcement.

Fragility analysis-confidence intervals

Figures 4a to 4d depict the limit state fragility curves for RC buildings for the Slight, Moderate,

Extensive and Complete structural damage states. The limit states are defined with respect to the

11
drift limits given in the work by Ghobarah [33]. For the case of the fully infilled case the drift

limits defining the damage states are equal to 0.1%, 0.4%, 0.7% and 0.8% for Slight, Moderate,

Extensive and Complete structural damage states, respectively. While for the case of the bare

frame the drift limits become 0.2%, 1.0%, 1.8% and 3.0% for four damage states, respectively. In

Figures 4a to 4d, the PGA value for the design earthquake for EKOS 2000 is denoted with a bold

vertical line and the corresponding limit-state probabilities of exceedance are given in Table 3.

Worth mentioning the observation that, although, the probability of exceedance for the fully

infilled design for the Slight damage state is almost the same with that of the weak ground storey

and short columns designs (87% versus 100%), the probability of exceedance for the Moderate

damage state of the fully infilled design is one order of magnitude less than the corresponding

probability of the other two designs (4.5% versus 97.0%) while for the Complete damage state the

probability of exceedance for the fully infilled design is three orders of magnitude less than the

corresponding probability of the other two designs (0.02% versus 42.0% and 13.0%).

In order to define the 95% confidence intervals for all fragility curves two sources of

randomness have been taken into account: (i) those affecting the structural capacity (fc:

compressive strength, εc,y: yield compressive strain, εc,u: ultimate compressive strain, Ec: concrete

modulus of elasticity, fy: steel yield stress, Es: steel modulus of elasticity, finf: perforated brick

compressive strength, Einf: perforated brick modulus of elasticity) and (ii) those affecting the

seismic demand. Details on the two sources of randomness can be found in Table 4. The

confidence intervals for the four sets of fragility curves are also depicted in Figures 4a to 4d,

while the 95% confidence intervals of the probability of exceedance of the four limit states for the

design earthquake can be found in Table 3.

Test example 2

The three storey bare 3D RC building (see Figure 5) first analyzed by Lagaros et al. [34] has been

employed as the second test example of this work. The 3D RC building has been designed to meet
12
the EKOS 2000 and EAK 2000 requirements for different values of the behaviour factor q ranging

from 1.0 to 6.0 with a step size equal to 1.0, while all six cases were designed as bare frames.

Description of the test cases

The parametric study resulted in six different designs depending on the value of the behavioural

factor q. The design process for the Greek design code started with the same initial design having

the same minimum dimensions for the beams and the columns. The minimum dimensions that

have been considered were the following: columns 25x25 cm2 and beams 25x30 cm2. The

dimensions of the columns and the beams were increased until fulfilling the requirements and

provisions of either the EAK 2000 for the current value of the behaviour factor q. Beams and

columns that fail to meet the constraints imposed by EAK and EKOS design codes were increased

in size according to the following procedure: (i) columns: increase the size of the smallest

dimension by 5 cm and if column constraints are not yet satisfied increase the size of the second

dimension by 5 cm, this rule is processed until satisfying all the constraints, (ii) beams: increase

the size of the height of the beam by 5 cm and if constraints are not yet satisfied the width of the

web is increased by 5 cm (the with width of the web is restricted not to exceed 35 cm) until

satisfying all the constraints, similarly with the columns. The weight of the steel and the concrete

quantities required for the six designs obtained with the Greek national design codes are given in

Table 5.

Fragility analysis-confidence intervals

Figures 6a to 6d depict the limit state fragility curves for RC buildings for the Slight, Moderate,

Extensive and Complete structural damage states. In accordance to the previous test example, the

limit states are defined with respect to the drift limits given in [33] for the bare frame. In Figures

6a to 6d, the PGA value for the design earthquake for EKOS 2000 is denoted with a bold vertical

line and the corresponding probabilities of exceedance of the four damage states are given in

13
Table 6. In order to evaluate the performance of the different designs achieved, three

characteristic designs were selected. These designs correspond to the two extreme designs with

respect to the value of the behavioural factor q. The first extreme design, denoted as Dq=1, is the

one achieved for q=1 (permitting linear behaviour only) and the second extreme design is denoted

as Dq=6, corresponding to the largest value of the behaviour factor examined in this study. While

the third design that is employed, denoted as Dq=3, corresponds to the design obtained for q=3.0.

The probability of exceedance of the Dq=1 design for the Slight damage state is almost the

same with the corresponding of the Dq=3 and Dq=6 designs (66% versus 100%), the probability of

exceedance of the Moderate damage state for the Dq=1 design is one order of magnitude less than

the corresponding probability of the other two designs (3.5% versus 17.0% and 78% respectively)

while for the Complete damage state the probability of exceedance for the Dq=1 design is two and

three orders of magnitude less than the corresponding probability of Dq=3 and Dq=6 designs,

respectively. Comparing the values of Table 6 with those of Table 3 worth mentioning that the

behaviour of the Dq=1 design is similar to the fully infilled design of the previous test example in

terms of probabilities of exceedance for the design earthquake for all four limit states.

In order to define the 95% confidence intervals for all fragility curves, similarly to the

previous test example, two sources of randomness have been taken into account: (i) those

affecting the structural capacity and (ii) those affecting the seismic demand. Details on the two

sources of randomness can be found in Table 7. The confidence intervals for the four sets of

fragility curves are also depicted in Figures 6a to 6d, while the 95% confidence intervals of the

probability of exceedance of the four limit states for the design earthquake can be found in Table

6.

CONCLUSIONS

The main purpose of the present study was to examine the effectiveness of the fragility analysis in

order to assess the seismic performance of multi-storey RC buildings designed based on modern
14
codes. For this reason a parametric study has been performed considering two groups of

buildings. In the first one weak ground storey and short columns construction features have been

implemented, while in the second group six different designs have been obtained implementing

different values of the behaviour factor. Fragility analysis has been proved as an efficient tool for

assessing the behaviour of a structural system. Through this parametric study, three were the most

significant findings:

(i) Although, the probability of exceedance of the slight damage state, for the design

earthquake, is of the same order for all three designs. On the other hand, it was found that

the probability of exceedance for the fully infilled design is one and three orders of

magnitude less than that of the other two designs for the moderate and complete damage

states, respectively.

(ii) Similar observations obtained for the structure designed for q=1 compared to those

designed for larger values of the behaviour factor. More specifically the probability of

exceedance of the moderate damage state for the Dq=1 design is one order of magnitude less

than that of the other Dq=3 and Dq=6 designs, while for the complete damage state the

probability of exceedance for the Dq=1 design is two and three orders of magnitude less than

the corresponding probability for Dq=3 and Dq=6 designs, respectively.

(iii) Furthermore, an important observation of this study can be obtained comparing the results

of the two test examples studied. Through this comparison it was found that the behaviour,

in terms of limit-state probability of exceedance for the design earthquake, of the bare

design obtained for q=1 is similar to that of the fully infilled design obtained for q=3.5.

REFERENCES

1. Kennedy, R.P., Cornell, C.A., Campbell, R.D., Kaplan, S., Perla, H.F. (1980), “Probabilistic seismic
safety study of an existing nuclear power plant,” Nuclear Engineering and Design; 59(2):315-338.
2. Kircher, C.A., Nassar, A.A., Kustu, O., Holmes, W.T. (1997), “Development of building damage
functions for earthquake loss estimation,” Earthquake Spectra; 13(4):663-682.

15
3. FEMA-National Institute of Building Sciences. HAZUS-MH MR1, Multi-hazard Loss Estimation
Methodology Earthquake Model, Washington, DC, 2003.
4. Shinozuka, M., Feng, M.Q., Lee, J., Naganuma, T. (2000), “Statistical analysis of fragility curves,”
Journal of Engineering Mechanics; 126(12):1224-1231.
5. Ellingwood, B.R. (2001), “Earthquake risk assessment of building structures,” Reliability
Engineering and System Safety; 74(3):251-262.
6. Wen, Y.K., Ellingwood, B.R. (2005), “The role of fragility assessment in consequence-based
engineering,” Earthquake Spectra; 21(3):861-877.
7. Aslani, H., Miranda, E. (2005), “Probability-based seismic response analysis,” Engineering
Structures; 27(8):1151-1163.
8. Pagni, C.A., Lowes, L.N. (2006), “Fragility functions for older reinforced concrete beam-column
joints,” Earthquake Spectra; 22(1):215-238.
9. Kappos, A.J., Panagopoulos, G., Panagiotopoulos, C., Penelis, G. (2006), “A hybrid method for the
vulnerability assessment of R/C and URM buildings,” Bulletin of Earthquake Engineering;
4(4):391-413.
10. Jeong, S.-H., Elnashai, A.S. (2007), “Probabilistic fragility analysis parameterized by fundamental
response quantities,” Engineering Structures; 29(6):1238-1251.
11. Porter, K., Kennedy, R., Bachman, R. (2007), “Creating fragility functions for performance-based
earthquake engineering,” Earthquake Spectra; 23(2):471-489.
12. Shinozuka, M., Murachi, Y., Dong, X., Zhou, Y., Orlikowski, MJ., (2003), “Effect of seismic
retrofit of bridges on transportation networks,” Earthquake Engineering and Engineering
Vibration; 2(2):169-179.
13. Lee, H.-S., Woo, S.-W. (2002), “Effect of masonry infills on seismic performance of a 3-storey R/C
frame with non-seismic detailing,” Earthquake Engineering and Structural Dynamics, 31(2):
353-378.
14. Negro, P., Verzeletti, G. (1996), “Effect of infills on the global behaviour of R/C frames: Energy
considerations from pseudodynamic tests,” Earthquake Engineering and Structural Dynamics;
25(8): 753-773.
15. Li, Z.-X. (2005), “Theory and technology of split reinforced concrete columns,” Engineering
Mechanics, 22: 127-141.
16. Guevara, L.T., García, L.E. (2005), “The captive- and short-column effects,” Earthquake Spectra;
21(1): 141-160.
17. Fajfar, P. (1998), “Towards nonlinear methods for the future seismic codes,” in Booth, editor,
Seismic Design Practice into the Next Century, Balkema.
18. Mazzolani, F.M., Piluso, V. (1996), The Theory and Design of Seismic Resistant Steel Frames, E &
FN Spon.
19. EAK 2000, National seismic code of Hellas, 2000.
20. EN 1998-1:2003. Eurocode 8: Design of Structures for Earthquake Resistance. Part 1: General rules,
seismic actions and rules for buildings. Commission of the European Communities, European
Committee for Standardization, October 2003.
21. EKOS 2000, National code for concrete building structures of Hellas, 2000.
22. EN 1992-1-1:2002. Eurocode 2: Design of Concrete Structures. Part 1: General rules and rules for
buildings. Commission of the European Communities, European Committee for Standardization,
November 2002.
23. Dolšek, M., Fajfar, P. (2001), “Soft storey effects in uniformly infilled reinforced concrete frames,”
Journal of Earthquake Engineering; 5(1): 1-12.

16
24. Ghobarah, A., Saatcioglu, M., Nistor, I. (2006), “The impact of the 26 December 2004 earthquake
and tsunami on structures and infrastructure,” Engineering Structures; 28(2): 312-326.
25. Chao, H.H., Yungting, A.T., Ruo, Y.H., (2006), “Nonlinear pushover analysis of infilled concrete
frames,” Earthquake Engineering and Engineering Vibration, 5(2):245-255.
26. Mitchell, D., DeVall, R.H., Kobayashi, K., Tinawi, R., Tso, W.K. (1996), “Damage to concrete
structures due to the January 17, 1995, Hyogo-ken Nanbu (Kobe) earthquake,” Canadian Journal of
Civil Engineering; 23(3): 757-770.
27. Somerville, P., Collins, N. (2002), “Ground motion time histories for the Humboldt bay bridge,”
Pasadena, CA, URS Corporation.
28. Papazachos, B.C., Papaioannou, Ch.A., Theodulidis N.P. (1993), “Regionalization of seismic
hazard in Greece based on seismic sources,” Natural Hazards; 8(1): 1-18.
29. Chintanapakdee, C., Chopra, A.K. (2003), “Evaluation of modal pushover analysis using generic
frames,” Earthquake Engineering and Structural Dynamics; 32(3): 417-442.
30. McKenna, F., Fenves, G.L. (2001), The OpenSees Command Language Manual - Version 1.2,
Pacific Earthquake Engineering Research Centre, University of California, Berkeley.
31. Lagaros, N.D. (2007), “Life-cycle cost analysis of design practices for RC framed structures,”
Bulletin of Earthquake Engineering; 5(3):425-442.
32. Perera, R., Gomez, S., Alarcon E. (2004), “Experimental and analytical study of masonry infill
reinforced concrete frames retrofitted with steel braces,” Journal of Structural Engineering;
130(12): 2032-2039.
33. Ghobarah, A., (2004), “On drift limits associated with different damage levels,” International
Workshop on Performance-Based Seismic Design, June 28-July 1.
34. Lagaros, N.D., Fotis, A.D., Krikos, S.A. (2006), “Assessment of seismic design procedures based on
the total cost,” Earthquake Engineering and Structural Dynamics; 35(11):1381-1401.

17
 TABLES

Table1. Natural records [27]

Earthquake Station Distance Site

Tabas (TB) Dayhook 14 rock
16 September 1978 Tabas 1.1 rock
Cape Mendocino (CM) Cape Mendocino 6.9 rock
25 April 1992 Petrolia 8.1 soil
TCU052 1.4 soil
TCU065 5.0 soil
TCU067 2.4 soil
TCU068 0.2 soil
TCU071 2.9 soil
TCU072 5.9 soil
TCU074 12.2 soil
Chi-Chi (CC), Taiwan
TCU075 5.6 soil
20 September 1999
TCU076 5.1 soil
TCU078 6.9 soil
TCU079 9.3 soil
TCU089 7.0 rock
TCU101 4.9 soil
TCU102 3.8 soil
TCU129 3.9 soil

Table 2. Test example 1 - Comparison of steel and concrete quantities

Columns Beams
Model
Steel (kg.) Concrete (m3) Steel (kg.) Concrete (m3)
Fully infilled 5868 23.7 5495 39.5
Weak ground storey 6516 23.7 5794 39.5
Short columns 6210 23.7 5495 39.5

18
 Table 3. Test example 1 – Limit-state probability of exceedance (%)
Weak ground
Limit state Fully infilled Short columns
storey
Mean 87.03 100.00 100.00
Slight + 87.73 100.00 100.00
- 86.32 100.00 100.00

Mean 4.46 97.22 96.60

Moderate + 4.78 97.42 96.84
- 4.16 97.01 96.34

Mean 0.33 81.21 62.47

Extensive + 0.37 82.11 63.70
- 0.30 80.30 61.26

Mean 0.02 42.28 13.44

Complete + 0.02 43.58 14.17
- 0.02 41.01 12.75

Table 4. Test example 1 - Characteristics of the random variables

Parameter Distribution Mean value CoV Correlation

Confined Concrete
fc (MPa) Normal 16.0 0.15 0.3
εc,y Normal 0.002 0.15 0.3
εc,u Normal 0.0035 0.15 0.3
Unconfined Concrete
fc (MPa) Normal 12.8 0.15 0.3
εc,y Normal 0.002 0.15 0.3
εc,u Normal 0.0035 0.15 0.3
Ec (GPa) Normal 28.0 0.15 0.3
Steel
fy (MPa) Normal 500.0 0.05 0.5
Es (GPa) Normal 200.0 0.05 0.5
Masonry Infill
Finf (MPa) Normal 3.0 0.15 0.3
Einf (GPa) Normal 2.25 0.15 0.3
Seismic Action
Seismic Load Lognormal x̂ (Εq. 6) δ (Εq. 7) 0.0

19
 Table 5. Test example 2 - Comparison of steel and concrete quantities
Columns Beams
Design procedure
Steel (kg.) Concrete (m3) Steel (kg.) Concrete (m3)
Dq=1 12700 32 6940 27
Dq=2 7720 21 4180 17
Dq=3 5730 15 3170 13
Dq=4 4600 14 2490 11
Dq=5 4010 12 2140 11
Dq=6 3750 11 1940 10

Table 6. Test example 2 - Limit-state probability of exceedance (%)

Limit state q=1 q=2 q=3 q=4 q=5 q=6

Mean 66.71 92.44 98.69 99.60 99.85 99.95
Slight + 67.94 92.91 98.79 99.63 99.86 99.96
- 65.50 91.96 98.57 99.56 99.83 99.95

Mean 3.44 6.38 16.92 28.68 36.09 78.49

Moderate + 3.70 6.82 17.75 29.81 37.33 79.45
- 3.20 5.98 16.13 27.59 34.88 77.53

Mean 0.27 0.56 3.41 8.12 12.69 42.54

Extensive + 0.30 0.61 3.66 8.61 13.40 43.87
- 0.25 0.51 3.17 7.66 12.01 41.25

Mean 0.01 0.03 0.35 1.57 2.52 8.05

Complete + 0.02 0.04 0.39 1.71 2.72 8.56
- 0.01 0.03 0.32 1.45 2.33 7.58

Table 7. Test example 2 - Characteristics of the random variables

Parameter Distribution Mean value CoV Correlation

Confined Concrete
fc (MPa) Normal 16.0 0.15 0.3
εc,y Normal 0.002 0.15 0.3
εc,u Normal 0.0035 0.15 0.3
Unconfined Concrete
fc (MPa) Normal 12.8 0.15 0.3
εc,y Normal 0.002 0.15 0.3
εc,u Normal 0.0035 0.15 0.3
Ec (GPa) Normal 28.0 0.15 0.3
Steel
fy (MPa) Normal 500.0 0.05 0.5
Es (GPa) Normal 200.0 0.05 0.5
Seismic Action
Seismic Load Lognormal x̂ (Εq. 6) δ (Εq. 7) 0.0

20
 FIGURES

16.0

Median 10in50 X
14.0

12.0

10.0
SA (m/sec )
2

8.0

6.0

4.0

2.0

0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0
T (sec)

(a)

14.0

Median 10in50 Y
12.0

10.0
SA (m/sec )
2

8.0

6.0

4.0

2.0

0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0
T (sec)

(b)
Figure 1. Natural records response spectra and their median a) Longitudinal(x), b) Transverse(y)

21
 (a) (b)
Figure 2. Test example 1 - Geometry of the three storey 3D building (a) layout, (b) front view

(a) (b)

(c)
Figure 3. Test example 1 - The models for the C7-B5-C8-B6-C9 frame (a) fully infilled, (b) weak ground storey, (c) short
columns

22
 1 1

Probability of exceeding the limit state

Probability of exceeding the limit state

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2
fully infilled fully infilled
weak ground storey weak ground storey
short columns short columns
0 0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Design Earthquake PGA=0.31g Design Earthquake PGA=0.31g
PGA (g) PGA (g)

(a) (b)
1 1
Probability of exceeding the limit state

Probability of exceeding the limit state

0.8 0.8

0.6 0.6

0.4 0.4
fully infilled
weak ground storey
short columns
0.2 0.2
fully infilled
weak ground storey
short columns
0 0
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80
Design Earthquake PGA=0.31g Design Earthquake PGA=0.31g
PGA (g) PGA (g)

(c) (d)
Figure 4. Test example 1 - Fragility curves for the four limit states and their confidence intervales, (a) Slight, (b) Moderate, (c) Extensive, (d) Complete

23
 (a) (b)
Figure 5: Geometry of the three storey 3D building (a) layout, (b) front view

24
 1 1
q=1
q=2
View publication stats

q=3
q=4
q=5
Probability of exceeding the limit state

Probability of exceeding the limit state

0.8 0.8 q=6

0.6 0.6

0.4 0.4
q=1
q=2
q=3
q=4
q=5
0.2 q=6 0.2

0 0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Design Earthquake PGA=0.31g Design Earthquake PGA=0.31g
PGA (g) PGA (g)

(a) (b)
1 1
q=1 q=1
q=2 q=2
q=3 q=3
q=4 q=4
q=5 q=5
Probability of exceeding the limit state

Probability of exceeding the limit state

q=6 q=6
0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80
Design Earthquake PGA=0.31g Design Earthquake PGA=0.31g
PGA (g) PGA (g)

(c) (d)
Figure 6. Test example 2 - Fragility curves for the four limit states and their confidence intervales, (a) Slight, (b) Moderate, (c) Extensive, (d) Complete

25