SEISMIC FRAGILITY STUDIES ON LOW-RISE AND MID-RISE REINFORCED CONCRETE

FRAME BUILDINGS USING DUZCE DATABASE

M.ALTUG ERBERIK
Department of Civil Engineering, Middle East Technical University, 06531, Ankara, Turkey

SERDAR CULLU
Department of Civil Engineering, Middle East Technical University, 06531, Ankara, Turkey

In the last decade, severe earthquakes near urban areas have caused substantial physical losses and casualties
in Turkey. Since the majority of the population is living in earthquake-prone areas, it is probable that such
terrible events will take place again in the near future. Moreover, it is not easy to cope with the substantial
direct and indirect economical losses after each devastating earthquake for a developing country like Turkey.
Therefore it is vital to quantify the earthquake risk and to develop strategies for disaster mitigation. In order
to achieve this goal, an extensive and inter-disciplinary study is required. Such a study is composed of two
parts: hazard determination and vulnerability assessment. The part of the study related to structural
engineering is the determination of the vulnerability of existing engineering structures and building stock.
This paper focuses on low-rise and mid-rise reinforced concrete frame buildings, which constitute
approximately 75% of the total building stock in Turkey and which are generally occupied for residential
purposes. The observations related to the reinforced concrete buildings which suffered heavy damage or even
collapsed after recent earthquakes are as follows:
• Most of the buildings are not designed according to the current code regulations,
• Seismic behavior is not taken into consideration in the architectural design and during selection of the
structural system.
• Supervision in the construction phase is not adequate which in turn induces deficiencies like poor
concrete quality, inadequate detailing of reinforcement, etc.
The tool that is employed to assess the seismic performance of low-rise and mid-rise reinforced concrete
frame buildings is the fragility curve in this study. By definition, fragility curves provide estimates for the
probabilities of reaching or exceeding various limit states at given levels of ground shaking intensity for an
individual structure or population of structures [1]. A limit state; which is in the same terms as the response,
usually represents a damage condition or a limitation of usage.
During the construction of the fragility curves for building structures, it is necessary to consider the
country-specific characteristics of the building stock. The reason is that the construction practice may differ
substantially in different countries and since the differences in the countryspecific characteristics of building
structures are directly reflected in the fragility curves, this may lead to erroneous estimates in terms of
earthquake damage and loss. However, this fact has usually been ignored and the results obtained from
fragility analysis of building structures in other countries have been adopted to the earthquake loss estimation
and master-plan studies in Turkey. Hence the main goal of the study is to determine the fragility of low-rise
and mid-rise reinforced concrete frame buildings by considering the country-specific structural characteristics
in order to overcome this deficiency.
The inventory used in this study is composed of 28 reinforced concrete buildings extracted from a
building database of approximately 500 buildings in Duzce city, which has been affected by two major
earthquakes in 1999. Post-earthquake damage assessments of the buildings are also available. This valuable
database has been used before by other researchers [2,3]. In this particular study, it is treated as the field data
which is employed to conduct a very detailed sensitivity analysis based on fragility curves. By using this field
data, the effects of various input and response parameters and methods on the fragility functions are
investigated.
To estimate the seismic vulnerability of a specific building type, two different approaches can be
considered. In the first approach, each building in the stock is examined individually and the vulnerability of
the building stock is obtained by combining the fragility information associated with each building. Very
detailed models and analysis procedures are employed, hence the results will be highly accurate. On the other
hand, this approach is practically and economically unfeasible. The second approach is to conduct the
fragility studies by using the statistical properties of the building population. Simple models and methods are
employed in this approach. The advantage of this method is that it is simple and economically feasible. In
addition, the nontechnical decision makers prefer such simple and rapid estimates of anticipated losses to
develop the proper judgment to execute their mitigation plans. However the obtained results will be crude and
the limitations of the models or the methods should be clearly understood.
In this study, fragility functions of low-rise and mid-rise reinforced concrete structures are constructed by
considering the population characteristics as explained above. The building database is divided into two sub-
groups: buildings with two and three stories are regarded as lowrise (LR) and buildings with four to six
stories are considered as mid-rise (MR). The buildings are further classified as “bare frame” and “infilled
frame”. Each building is represented by an equivalent Single-Degree-of-Freedom (SDOF) system with three
structural parameters: Period or the stiffness (T), strength ratio ( ) and the post-elastic stiffness ( ). These
η

three parameters are obtained from the push-over analysis of the buildings in two orthogonal directions. The
spectral capacity curves of the buildings were constructed by using the software SAP2000 [4] and the results
can be found elsewhere [2]. Bilinearization of the spectral capacity curve is conducted by using a procedure
similar to the one given in FEMA356 [5]. As a result, the building population is characterized by the
structural parameters T, and . (Figure 1).
η

Figure 1. a) A sample taken from the obtained spectral capacity curves for MR/infilled frames, b) Bilinearization of the spectral
capacity curve and the structural parameters to be used in the fragility studies.

Ground motion parameters used in the study are peak ground acceleration (PGA), peak ground velocity
(PGV) and spectral displacement at the fundamental period of the building (Sd). However, the reference
ground motion parameter is considered as PGV and selection of the ground motion set is carried out
accordingly. There are approximately 150 ground motions covering a wide range of ground motion
parameters. Peculiar ground motions records due to extreme near-field and soil site effects are not included in
the dataset.
 The sensitivity analysis on fragility curves is carried out in several steps as outlined below: Three
different approaches are used to quantify the structural inventory that represents the building stock. The
random variables are selected as the period (T), the strength ratio ( ) and the post-elastic stiffness ( ).
η α

a) Use the actual building data in terms of (T, , ), η α

b) Use uniform distribution for the structural parameters (T, , ), η α

c) Use normal statistical distribution for the structural parameters (T, , ) with sampling techniques.
η α

The effect of parameter on the final fragility functions is further investigated, since the random
α

variables can be reduced to two if its effect can be neglected. In part (c), the effect of sample size is also
checked by using 30, 75 and 150 samples.
Attainment of the limit states is of paramount importance in the fragility curve generation. In this study,
the influence of employing deterministic and probabilistic limit states on the final fragility functions is also
examined.
Examples of final fragility curves for mid-rise infilled frames in terms of PGA and PGV are presented in
Figure 2. As a summary, this study investigates the effect of different parameters or approaches on the
fragility functions and uses Duzce database as the test-bed. Final outcome is a set of fragility curves for
typical low-rise and mid-rise reinforced concrete frame buildings in Turkey. Results obtained from the
fragility analysis are also compared with the observed field data.

Figure 2. Sample fragility curves for mid-rise infilled frames in terms of a) PGV, b) PGA

References

[1] Erberik, M.A. and Elnashai, A.S., 2003. “Seismic Vulnerability of Flat-Slab Structures”, Mid-America Earthquake (MAE) Center,
Report No. 03-06, University of Illinois at Urbana-Champaign, IL.
[2] Yakut A., Yilmaz, N. And Bayili, S., 2003. “Analytical Assessment of Seismic Capacity of RC Frame Buildings”, International
Conference on Skopje Earthquake - 40 Years of European Earthquake Engineering SE40EEE, Skopje-Ohrid.
[3] Akkar, S., Sucuoglu, H. And Yakut A., 2004. “Displacement-Based Fragility Functions for Low and Mid-rise Ordinary Concrete
Buildings”, Earthquake Spectra, in review.
[4] Computers and Structures, Inc., 2000. SAP2000 Nonlinear, Version 7.21, Structural Analysis Program, Berkeley CA.
[5] Federal Emergency Management Agency, 2000. FEMA 356 Prestandard and Commentary for the Seismic Rehabilitation of
Buildings. American Society of Civil Engineers, Washington DC.

Keywords: fragility; reinforced concrete buildings; sampling techniques; capacity curve; limit states