You are on page 1of 7

Arab J Sci Eng (2015) 40:23012307

DOI 10.1007/s13369-015-1758-y

TECHNICAL NOTE - CIVIL ENGINEERING

Seismic Fragility Curves for Steel and Reinforced Concrete


Frames Based on Near-Field and Far-Field Ground Motion
Records
Fadzli Mohamed Nazri1 Siti Nur Aqilah Saruddin1

Received: 1 August 2014 / Accepted: 8 June 2015 / Published online: 27 June 2015
King Fahd University of Petroleum & Minerals 2015

Abstract This study presents the fragility curve for steel in identifying the performance level [1]. It can also be ex-
and concrete frames based on near-field (NF) and far-field pressed as the maximum desired extent of damage to a struc-
(FF) ground motion records. The fragility curve is calibrated ture under a specific earthquake design level.
to the incremental dynamic analysis (IDA) curve based on The drift performance level varies depending on the guide-
building materials and frame heights. The frames are de- lines that result from intensive research. For example, FEMA-
signed according to Eurocodes. This study uses seven ground 273 [2] classifies the performance level into four categories,
motion records for NF and another seven records for FF. The namely operational phase (OP) with 0.5 % drift, immediate
five performance levels prescribed by FEMA-273 are used occupancy (IO) with 1.0 % drift, life safety (LS) with less
as structure performance benchmarks in generating the IDA than 2.5 % drift, and collapse prevention (CP) with more than
curve. Results based on the NF and FF records indicate that 2.5 % drift.
a three-storey steel frame is stiffer than a concrete frame. By The PBSD is incorporated with incremental dynamic
contrast, the results for a six-storey structure show that a steel analysis (IDA) to determine the structure drift and develop the
frame is more suitable than low-rise and mid-rise frames. The structure fragility curve. The fragility curve is a log-normal
fragility curve results show that the probability of reaching function that expresses the probability of reaching or exceed-
or exceeding a specified damage state can be determined. ing a specific damage state. The fragility curve is a highly
useful method in predicting the extent of probable damage.
Keywords Fragility curve Drift IDA Performance It can describe the probability of a structure being damaged
level Ground motion records beyond a specific state when subjected to different levels of
ground shaking [3,4].
The fragility curve was introduced to provide accurate
1 Introduction predictions of damage to structures and non-structures. This
method is unique because every developed curve depends
Building damages can be related to structure displacement on ground motions. Numerous researchers have developed
or drift. Drift is normally used as a damage parameter to es- fragility curves for different structure types [311].
timate the performance level of a building structure. Drift Malaysia is categorized as having low seismicity, but this
can be measured using the performance damage level in issue should be taken seriously. Thus, the current study was
performance-based seismic design (PBSD), which is a new conducted to develop the fragility curve for low-rise and
approach to earthquake-resistant design. Drift can be de- mid-rise steel and reinforced concrete frames. In this study,
scribed by considering the maximum allowable damage state we present the development of the fragility curve based on
a simplified fragility curve equation that is based on near-
B Fadzli Mohamed Nazri and far-field ground motion records. A central goal was to
cefmn7@gmail.com (1) determine drift analysis based on different materials and
1 heights that affect dynamic loading and (2) develop the seis-
Disaster Research Nexus, School of Civil Engineering,
Engineering Campus, Universiti Sains Malaysia, 14300, mic fragility curve for steel and concrete frame structures. For
Nibong Tebal, Penang, Malaysia a country like Malaysia, the effects of the near- and far-field

123
2302 Arab J Sci Eng (2015) 40:23012307

earthquake to the structure will be a good input for designers, Table 1 Beam and column sizes of the steel frame
engineers, and practitioners as a reference in seismic design. No. of storey Beam size (mm) Column size (mm)
The impacts of the structural damage can be clearly observed
from the drift results. 3 533 210 109 305 305 198
6 533 210 82 305 305 198

2 Methodology
Table 2 Beam cross section for the concrete frame
The general flowchart for this study is shown in Fig. 1. No. of storey Beam size Reinforcement Shear link
Moment-resisting frames (MRFs) were designed using two (mm) (mm)
different materials and two structural heights.
3 300 700 6T25 8 mm link at
150 mm c/c
Moment-Resisting Frame Design 6 300 700 4T32 8 mm link at
150 mm c/c

Table 3 Column cross section for the concrete frame


Eurocodes
(EC) No. of storey Column size (mm) Reinforcement (mm)

3 500 500 6T32


6 500 500 6T32

Steel Frame Concrete Frame


Design Design
Repeat the analysis with a different storey height

These frames abide by the Eurocodes (EC). Each MRF


had two bays measuring 6 m each. In addition, every frame
had a storey height of 3 m for three- and six-storey concrete
and steel frames. The frame designs were according to EC3
(3-, and 6-storey)

SAP 2000
[12], EC2 [13], and EC8 standards. The steel grade, f y , was
assumed as S275 for the steel frame. The natural periods for
three-storey and six-storey structures were 0.44 and 0.74 s,
Ground Motion respectively. The frames were designed with beam and col-
Repeat the analysis with different

Records
umn dimensions listed in Table 1. Several assumptions were
ground motion records

made during the design of concrete frames as follows:

Drift Limit State Bar diameter: 20 mm


(FEMA-347)
Link diameter: 10 mm
Cover to reinforcement: 25 mm
Concrete strength: 30 N/mm2
Incremental
Dynamic
Steel yield strength: 460 N/mm2
Analysis Beam width: 300 mm
(IDA) Beam depth: 700 mm
Column size: 500 mm 500 mm
IDA Curve
The natural periods of three-storey and six-storey structures
were 0.39 and 0.66 s, respectively. Tables 2 and 3 show the
Drift Limit State cross section, reinforcement beam, and column model used.
Both concrete and steel structures used soil type A as their
Mean,
Standard ground type. We set the peak ground acceleration on soil
Deviation, type A, agr , as 0.5 g or 5 m/s2 to calculate horizontal seismic
action. The value of the importance factor used was 1.0, and
the behaviour factor, q, of the ductility class medium MRFs
Fragility curve was set at 4 as prescribed in EC8.
SAP 2000 software was used in this study to develop the
Fig. 1 General flowchart of the methodology IDA curve. Tables 4 and 5 show the seven ground motions

123
Arab J Sci Eng (2015) 40:23012307 2303

Table 4 Selective ground motion records for NF and Pang [15] stated that drift is a mean drift calculated by
Name record Earthquake location Year dividing the displacement with the total building height.
After the IDA curve in peak ground acceleration (PGA)
NGA 0146 Coyote Lake 1979 drift was developed in format form, the fragility curve was
NGA 0235 Mammoth Lake 1980 plotted using the suggested limit state [16]. The mean and
NGA 0318 Westmorland 1981 standard deviations for every limit state were then calculated.
NGA 0545 Chalfant Valley-02 1986 Many equations are used to generate the fragility curve. How-
NGA 0680 Whittier Narrows 1987 ever, Eq. (1) was used because it has already been simplified
NGA 1646 Sierra Madre 1991 by Ibrahim and El-Shami [3].
NGA 1740 Litte Skull Mountain 1992

P[D/PGA] = ((ln(PGA) )/ ) (1)

Table 5 Selective ground motion records for FF


where D = damage, PGA = peak ground acceleration, =
Name record Earthquake location Year standard normal cumulative distribution, = mean, =
NGA 0101 Northern California-07 1975 standard deviation of the PGA natural logarithm.
NGA 0106 Oroville-01 1975
NGA 0206 Imperial Valley-07 1979
NGA 0221 Livermore-02 1980 3 Results and Discussion
NGA 0225 Anza (Horse Canyon)-01 1980
NGA 0389 Coalinga-02 1983 3.1 Incremental Dynamic Analysis
NGA 1990 Gulf of California 2001
Every structure that has different characteristics and is sub-
jected to different ground motions records different IDA
curve patterns. These IDA curves demonstrate the relation-
for NF and FF used. The criteria used to select ground mo- ship between drift and PGA. This analysis focused only on
tion records were as follows: (a) the JoynerBoore distance the drift until it reached 3 % because the fragility curve to
for NF is less than 20 km, whereas that for FF records is be developed was based on the maximum drift limit for a
more than 20 km [14], and (b) the earthquake magnitude is particular performance level.
within the range of 56. These ground motions were scaled Five performance levels were used to assess the structural
to match the elastic response spectra. The elastic response performance. The vertical gridlines at drifts 0.5, 1.0, 1.5, 2.0,
spectra generated ranged from 0.05 to 0.6 g in increments of and 2.5 % were represented as OP, IO, damage control (DC),
0.05 g. LS, and CP, respectively.
Figure 1 also shows the steps involved in developing the As previously mentioned, the IDA curve was developed
fragility curve. The IDA curve and structure displacement based on the NF and FF records. Figures 2 and 3 show ex-
or drift were obtained to develop the fragility curve. Nazri amples of IDA curves obtained from the analysis of the NF

Fig. 2 Examples of IDA curves based on NF records for a concrete frame structure. a Three-storey concrete frame. b Six-storey concrete frame

123
2304 Arab J Sci Eng (2015) 40:23012307

Fig. 3 Examples of IDA curves based on FF records for a concrete frame structure. a Three-storey concrete frame. b Six-storey concrete frame

Fig. 4 Drift over height: a, b


under NF records; c, d under FF
records. a Three-storey frame. b
Six-storey frame. c Three-storey
frame. d Six-storey frame

and FF records for the three- and six-storey concrete frame discussion. Figure 4 shows the drift based on the structural
structures. material and height for the NF and FF records.
The drift analysis based on structural material and height The graphs show that the storey drift profiles for the three-
were determined using the IDA curve data. The drift was storey NF and FF are similar in appearance, which is similar
based on the IDA curve at a PGA of 0.6 g. The structures to the storey drift profiles for the six-storey NF and FF. How-
were already unstable at such PGA; thus, 0.6 g was consid- ever, every storey of the NF and FF has a different drift
ered as the critical PGA for the frame structures. Considering percentage. The distribution of the storey drift over height
these facts, we selected the drift results at 0.6 g for further for a three-storey frame is almost uniform, which shows that

123
Arab J Sci Eng (2015) 40:23012307 2305

Table 6 Parameters of
No. of storeys OP IO DC LS CP
log-normal distributions for
steel and concrete frames based
on NF
Steel
3 1.455 0.885 1.292 0.772 1.529 0.412 1.230 0.408 1.009 0.409
6 2.532 1.183 2.087 1.081 1.704 1.063 1.431 1.058 1.591 1.104
Concrete
3 1.938 0.575 1.486 0.512 1.177 0.340 0.900 0.373 0.957 0.103
6 2.579 0.449 2.027 0.454 1.447 0.462 1.056 0.447 1.022 0.389

Table 7 Parameters of
No. of storeys OP IO DC LS CP
log-normal distributions for
steel and concrete frames based
on FF
Steel
3 1.536 0.716 1.427 0.378 1.013 0.380 0.993 0.372 0.772 0.372
6 2.754 0.573 2.041 0.454 1.543 0.422 1.205 0.415 1.128 0.407
Concrete
3 2.515 0.740 1.876 0.639 1.518 0.517 1.249 0.395 1.079 0.323
6 3.245 0.407 2.489 0.321 1.991 0.342 1.574 0.401 1.333 0.396

the deflected structure shape is an almost straight line. The Seismic fragility is represented in this study by a damage
distribution of storey height for the six-storey frame is non- probability curve (i.e. fragility curve). The fragility curves
uniform in shape because it is influenced by ground motions. are shown in Figs. 5 and 6 based on the tabulated results and
Figure 4a shows that the maximum storey drifts of the Eq. 1.
three-storey steel and concrete frames under the NF records Four figures of the fragility curves were developed, namely
are 2.57 and 2.08 %, respectively, which occur on the second three-storey steel, three-storey concrete, six-storey steel, and
storey. Figure 4b shows that the maximum storey drift of the six-storey concrete based on a comparison between NF and
six-storey concrete frame under the FF records is 11.40 %, FF. The critical result for the three-storey frame was con-
which occurs on the top storeys. The maximum storey drift crete, whereas that for the six-storey frame was steel. The
of the six-storey steel frame is 6.32 %, which occurs on the following observations were noted. The PGA at 0.2 g was
third storey. The drift exceeded the CP level based on the considered as weak ground motions. The OP performance
limit state. levels for concrete NF and FF had probability values of ap-
Figure 4c shows that the maximum storey drift of the proximately 71 and 88 %, respectively. The probability at the
steel frame under the FF records is 2.03 %, which occurs CP level was 0 % for NF and approximately 5 % for FF. The
on the second storey. The maximum storey drift of the con- probability of reaching or exceeding the OP level was 100 %
crete frame under the FF records is 5.47 %, which occurs on for NF and FF when strong ground motions were exposed
the first storey. The critical drift of the steel frame for the six- at 0.5 g. The probability values were approximately 99 and
storey structures under the FF records (Fig. 4d) is 10.89 %, 88 % for NF and FF, respectively, at the CP level.
which occurs on the fourth storey, whereas the critical drift For the six-storey frame, the probability of reaching and
of the concrete frame is 18.15 %, which occurs on the sixth exceeding the OP level was 99 % for NF records and 100 %
storey. Thus, we conclude that the maximum storey drift ex- for FF records, whereas the CP performance levels were 7
ceeded the LS and CP levels compared with the limit state. and 24 % for NF and FF, respectively, when the PGA was
0.2 g. When the PGA was 0.5 g, the probability for both NF
3.2 Fragility Curve and FF was 100 % for the OP performance level. The prob-
ability values were approximately 80 and 94 % for NF and
Two parameters (i.e. mean and standard deviation) are re- FF, respectively, for the CP performance level.
quired to develop the fragility curve. The mean and standard FF records generally provided a higher percentage of prob-
deviation of PGA were calculated at every point across the ability of reaching or exceeding the OP level when exposed
limit state vertical gridlines at drifts of 0.5, 1.0, 1.5, 2.0, and to weak or strong ground motions. However, which records
2.5 %. Calculated parameters are listed in Tables 6 and 7. can provide a higher percentage of probability for CP level

123
2306 Arab J Sci Eng (2015) 40:23012307

Fig. 5 Fragility curve for a three-storey frame. a Fragility curve for a steel frame. b Fragility curve for a concrete frame

Fig. 6 Fragility curve for a six-storey structure. a Fragility curve for a steel frame. b Fragility curve for a concrete frame

cannot be determined because the result does not have a clear records. The drift profiles indicate that the steel frame
pattern. is stiffer than the concrete frame for the three-storey
structures under the NF and FF records. However, the
4 Conclusions concrete frame is stiffer than the steel frames for six-
storey structures under the NF and FF records.
This study attempts to develop the fragility curve for steel (ii) The fragility curves are developed based on the type
and concrete frame structures. The structure frames, which of material used, as well as the height and under dif-
are three and six storeys in height, are designed based on the ferent ground motion records. The fragility curve or
EC. SAP 2000 software is used as the main tool to analyse damage probability curve is based on the limit states
the dynamic loading of the structures. Seven ground motions of five performance levels, which are OP, IO, DC,
classified as NF and FF are used in the analysis. The drift LS, and CP. The three-storey steel frame demonstrates
parameter is considered as one of the key elements in devel- the highest probability of reaching or exceeding the
oping the fragility curve. The following conclusions can be OP (72 %) and CP (7 %) levels based on the ground
drawn from this study: motion records for NF. The concrete frame has the
highest probability of reaching or exceeding the OP
(i) The drift profiles of the three- and six-storey struc- level at 98 % for the six-storey structures. The steel
tures under the NF records are almost similar. How- frame has the highest probability of damage at the
ever, they have different drift percentages. Similar CP level (50 %). The concrete frame has the high-
findings are obtained from the structures under the FF est probability of reaching or exceeding the OP level

123
Arab J Sci Eng (2015) 40:23012307 2307

based on the FF records, which registers 89 % for the 3. Ibrahim, Y.E.; El-Shami, M.M.: Seismic fragility curves for mid-
three-storey structures and 100 % for the six-storey rise reinforced concrete frames in Kingdom of Saudi Arabia. IES
J. Part A Civil Struct. Eng. 4(4), 213223 (2011)
structures. The concrete frame also records the high- 4. Silva, V. et al.: Evaluation of analytical methodologies used to de-
est probability for both storeys at the CP level, which rive vulnerability functions. Earthq. Eng. Struct. Dyn. 43(2), 181
registers 5 % for the three-storey structures and 24 % 204 (2014)
for the six-storey structures. 5. Farsangi, E.N. et al.: Seismic risk analysis of steel-MRFs by
means of fragility curves in high seismic zones. Adv. Struct.
(iii) The higher probability of reaching the OP level oc- Eng. 17(9), 12271240 (2014)
curred at the FF records as shown in a comparison 6. Moharrami, H.; Amini, M.A.: Seismic vulnerability assessment
of the fragility curves between NF and FF records. of process towers using fragility curves. Struct. Des. Tall Spec.
Concrete material exhibited a higher probability of Build. 23(8), 593603 (2014)
7. Siqueira, G.H. et al.: Fragility curves for isolated bridges in eastern
OP for both NF and FF records. However, NF records
Canada using experimental results. Eng. Struct. 74, 311324 (2014)
produced the maximum probability of damage for the 8. Negulescu, C. et al.: Fragility curves for masonry structures sub-
steel frame structure at the CP level stage. Meanwhile, mitted to permanent ground displacements and earthquakes. Nat.
a concrete frame has the highest probability of dam- Hazards 74(3), 114 (2014)
9. Bakhshi, A.; Asadi, P.: Probabilistic evaluation of seismic de-
age for the frame structure exposed to the FF records. sign parameters of RC frames based on fragility curves. Sci.
Iran. 20(2), 231241 (2013)
The study results show that the structural vulnerabilities re- 10. Manafpour, A.R.; Moghaddam, P.K.: Probabilistic approach to
flect the assumptions made by referring to EC8 (i.e. behav- performance-based seismic design of RC frames. In: Vulnerability,
Uncertainty, and Risk@ sQuantification, Mitigation, and Manage-
iour factor and soil class). Although making an assumption ment. ASCE (2014)
based on the ground motion records is impossible, using rec- 11. Hancilar, U. et al.: Earthquake vulnerability of school build-
ommendations from the EC as guidelines for design purposes ings: probabilistic structural fragility analyses. Soil Dyn. Earthq.
is the fundamental approach. Thus, further investigations Eng. 67, 169178 (2014)
12. BSI: Eurocode 3: design of steel structuresPart 1-1: general rules
should be conducted by focusing on the behaviour factor and rules for building. In: BS EN: 1993-1-1: 2005, British Stan-
recommended in the EC, soilstructure interaction based on dards Institution, London, 2005
different soil types, and design connections. 13. BSI: Eurocode 2: design of concrete structurePart 1-1: general
rules and rules for buildings, In: BS EN: 1992-1-1: 2004, London,
Acknowledgments This research was supported by Universiti Sains British Standards Institution, 2004
Malaysia under Research University (Individual) Grant (814223) and 14. Li, Y.; Fan, F.; Hong, H.: Influence of number of records and scaling
MyBrain15. We would like to thank the anonymous reviewers for sug- on the statistics of seismic demand for lattice structure. Thin-
gestions and comments. Walled Struct. 87, 115126 (2015)
15. Nazri, F.M.; Pang, Y.K.: Seismic performance of moment resisting
steel frame subjected to earthquake excitations. Front. Struct. Civil
Eng. 8(1), 1925 (2014)
References 16. Xue, Q. et al.: The draft code for performance-based seismic design
of buildings in Taiwan. Eng. Struct. 30(6), 15351547 (2008)
1. Hokmabadi, A.; Fatahi, B.; Samali, B.: Recording inter-storey drifts
of structures in time-history approach for seismic design of building
frames. Aust. J. Struct. Eng. 13(2), 175 (2012)
2. FEMA-273: 73: NEHRP guidelines for the seismic rehabilitation
of buildings. Federal Emergency Management Agency (1997)

123

You might also like