Вы находитесь на странице: 1из 22

Engineering Failure Analysis 14 (2007) 1–22

www.elsevier.com/locate/engfailanal

What is to be learned from damage and failure of reinforced


concrete structures during recent earthquakes in Turkey?
1
M.H. Arslan *, H.H. Korkmaz
Selcuk University, Engineering and Architecture Faculty, Department of Civil Engineering, Konya, Turkey

Received 9 January 2006; accepted 11 January 2006


Available online 20 March 2006

Abstract

In Turkey, generally, building stock is formed from reinforced concrete structures and during last earthquakes, a large
number of these buildings in the epicenter regions were collapsed leading to widespread destruction and loss of life. In this
paper, the performance of reinforced concrete buildings during recent earthquakes in Turkey is discussed. The objective of
this paper is to provide a brief overview of damage as observed following the earthquakes. The failure modes consist of
foundation failures, soft stories, strong beams and weak columns, lack of column confinement, poor detailing practice and
non-structural damages. Observations from the earthquake damages are discussed and compared with TEC-98 (Turkish
Earthquake Code) and TBC-500-2000 (Turkish Building Code) requirements. Measurements of some damaged reinforced
concrete member examples are given and important general lessons learned from these earthquakes are formulated.
Finally, a short overview of the emergency management measures taken is also presented.
 2006 Elsevier Ltd. All rights reserved.

Keywords: Damages; Reinforced concrete structures; Failure types; Earthquake; Seismic code

1. Introduction

Turkey is situated on an active earthquake zone with shortest return periods and earthquakes caused loss of
lives in the history. In the last century, over than twelve major earthquakes with minimum magnitudes 7 (Ms)
caused significant casualties and extensive structural damage in Turkey. Earthquakes in Turkey are generally
of in land types that are more destructive than off shore types, even their magnitude could be smaller [1–3].
The earthquakes are concentrated along the North Anatolian Fault (NAF), East Anatolian Fault (EAF),
North East Anatolian Fault (NEAF) and West Anatolian Fault (WAF) as a result of north-ward motion
of the Arabian Plate and African Continent [4]. Most of the population and industry are under the threat
of a possible major earthquake. The most obvious example of this is the Kocaeli Earthquake occurred in

*
Corresponding author. Tel.: +90 332 223 1971; fax: +90 332 241 0635.
E-mail addresses: mharslan@selcuk.edu.tr, musahakanarslan@hotmail.com (M.H. Arslan), hhk73@selcuk.edu.tr (H.H. Korkmaz).
1
Tel.: +90 332 223 1965.

1350-6307/$ - see front matter  2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engfailanal.2006.01.003
2 M.H. Arslan, H.H. Korkmaz / Engineering Failure Analysis 14 (2007) 1–22

Nomenclature

C0 seismic code coefficient


K coefficient related to the type of framing system
I importance factor
S, S(T) spectral coefficient
A0 effective ground acceleration coefficient
R structural behavior factor
Ra ductility factor
TA spectrum characteristic period
T fundamental period of building
Ac gross section area of column
Nd max greater of the factored axial forces calculated under vertical loads only and under simultaneous
action of vertical and seismic loads
fck characteristic compressive cylinder strength of concrete
ql the longitudinal rebar ratio
‘b development length
fyd design yield strength of longitudinal reinforcement
fctd design tensile strength of concrete
/ rebar diameter
W minimum column dimension
D longer column dimension
bw beam width
sk spacing of transverse reinforcement
s1, s2, s3 tie spacing
M ra ; M r€u ultimate moment resistances calculated at the bottom and the top of column
Mri, Mrj ultimate moment at the ends of beam
Ve shear force
Aw effective web area of column cross-section
Vr shear strength of a cross-section column
Di storey drift
Vn column shear strength
Vp max. probable shear force required for the plastic hinge form. at column ends
Mp max. plastic moment capacity of the column
L clear height of the column
ln clear height of column between beams, clear span of beam between column
hi story height
vc shear strength carried by concrete
Asw transverse reinforcement area within a spacing
fywd yield strength of transverse reinforcement

Marmara Region with the magnitude of 7.4 on 17 August 1999. This earthquake caused economic power
decrease around 20 billion US dollars and over than 20,000 people is dead. During the last century, about
500.000 building collapsed and were heavily damaged.
The structural damage in all the recent disasters, considering the magnitude of the event, was much
heavier than one would normally expected in a country better prepared for disasters [5,6]. Leave the
non-engineered buildings aside, engineered structures in Turkey are far from possessing qualities that
would ensure satisfactory seismic performance. Although Turkey has a developed seismic code called
‘‘The 1998 Turkish Earthquake Code’’ [7], [TEC-98] which was prepared to ensure that all structures have
M.H. Arslan, H.H. Korkmaz / Engineering Failure Analysis 14 (2007) 1–22 3

adequate stiffness, strength and ductility, but most of the structures do not have these properties. The
main reason underlying this fact is the lack of inspection and supervision at the design and construction
stages.
Recently, the great earthquake occurred in the Turkey, strongly reminded people of the lessons to be
learned for increased safety in the future. In this paper, recent earthquakes and their damages on especially
reinforced concrete (RC) frame systems are studied. Seismic code requirements are discussed and compared
with observed details. Some analytical investigation on the seismic response of representative buildings help
in explaining observed damages. Also, in this paper construction practice in Turkey and mistakes causing
the damage or failure are discussed.

2. TEC-98 analysis requirements for EQ resistant buildings

After the Erzincan earthquake (1939, M = 7.9), the first Turkish seismic design code for buildings was pub-
lished in 1940. In the code, the base shear, V, was calculated as the product of a lateral force coefficient, C, and
the weight of the building W. The value of C was equal to 0.10 regardless of the location of designed structure
and the base shear force was uniformly distributed over the height of the building.
The earthquake code of Turkey was changed in 1968. The uniform load profile of 1940s code was
replaced with a first mode (inverted triangle) profile similar to the UBC [8]. In 1975, important provisions
were added to the seismic code, including new methods for calculating earthquake loads on building and
ductile detailing requirements for RC structures [9]. The lateral force coefficient C was defined in the follow-
ing equation:
C ¼ C0  K  I  S ð1Þ
where C0 was a seismic code coefficient and varied between 0.10 and 0.03, K was a coefficient related to the
type of framing system, I was an importance factor, S was a spectral coefficient.
After the 1992 Erzincan Earthquake, the Turkish Earthquake Code was substantially changed in 1998 [4].
The important provision in 1998’code is, RC buildings are classified as systems of either high or nominal duc-
tility, based on the detailing of the components. The lateral force coefficient C of the 1975 code was replaced
by
AðT Þ A0  I  SðT Þ
C¼ ¼ ð2Þ
Ra ðT Þ Ra ðT Þ
In Eq. (2), A0 is the effective ground acceleration coefficient and varied between the 0.10 g (for the four degree
zone) and 0.40 g (for the one degree zone). The importance factor, I, varies between 1 and 1.5, and equal to 1.0
for ordinary structures. S(T) spectral coefficient is based on fundamental period of building, T. In the code
S(T), is given by three equations in the short-period, constant acceleration, and constant velocity ranges.
S(T)–T curve vary as a function of soil type and max amplification of S(T) is 2.5.
The earthquake loads are reduced by dividing them into load reduction factor, R, which is similar to the
response modification factor in US codes. The seismic load reduction factor varied between 1,5 and 8. In gen-
eral, R is equal to Ra (Eq. (3)). If the natural structure period is less than spectrum characteristic period TA, Ra
is less than R. The maximum value of R depends on the assumed ductility (high or nominal) of the system and
varies between 3 and 8 [10,11]
1 þ 1:5ðT =T A Þ
If 0 6 T 6 T A then Ra ¼  2:5  or if T P TA then Ra ¼ R ð3Þ
1 þ R1 ðT =T A Þ
For comparing base shear force level of TEC-98 and UBC-97, a typical RC moment-resisting frame system
that is the most common seismic framing in Turkey is chosen (Fig. 1). The illustrative RC building consists of
four storey and three bays. Plan dimension of building is 18.50 m · 12.50 m. The height of ordinary floor is
3.15 m and the height of basement floor is 4.5 m, respectively. The ductility level is nominal and thickness
of the slab is 12 cm. There is no shear wall except for basement. Base shear is calculated according to the
TEC-98 and UBC and compared in Fig. 2 with respect to the local site class. The other parameter is seismic
zone. It can be seen easily from Fig. 2 that, for a common example of RC moment-resisting frames with
4 M.H. Arslan, H.H. Korkmaz / Engineering Failure Analysis 14 (2007) 1–22

Fig. 1. Model of sample building [11].

Fig. 2. The comparison of the base shear of UBC and TEC-98 [11].

nominal ductility (equivalent of OMRF in the US), the ordinates of the lateral-force coefficient of 1997 UBC
exceed those of the TEC-98. Also TEC-98 calculates 2–2.5 times less base shear than UBC. In addition, TEC-
98 do not take into account ‘‘near fault’’ effect.

2.1. Earthquake resistant design requirements and observed damages

This section mentions about components of RC frames, namely columns, beams, beam–column joints.
As addressed above, the important provision in 1998’code is, RC buildings are classified as systems of
either high or nominal ductility, based on the detailing of the components. In the high ductility buildings,
detailing requirements are more stringent. For instance, transverse reinforcement requirements along
the height of a column shown in Fig. 3 and all hoops must have 135 seismic hooks at both ends
(Fig. 4).
M.H. Arslan, H.H. Korkmaz / Engineering Failure Analysis 14 (2007) 1–22 5

Fig. 3. Column details from 1998 Turkish seismic code.

Fig. 4. Special seismic hoops.

2.1.1. Detailing of components


Ductility is the one of the most important parameter for RC structures to resist earthquake effects.
Dynamic energy must be dissipated by forming plastic hinge location at the end of the frame members. Espe-
cially, adequate transverse reinforcement, compression steel bars and low-level axial load is help to form plas-
tic hinge. Based on author’ observations, no additional transverse reinforcement for the purpose of the
confinement are provided in the hinge, joint and splice regions.

2.1.1.1. Columns. The main observation is that the majority of moment-frame component damages were
occurred in columns. These damages were due to unsuitable transverse reinforcement, inadequate cross-sec-
tion dimensions, non-ductile details, unconfined lap splices and inadequate anchorage lengths and excessive
6 M.H. Arslan, H.H. Korkmaz / Engineering Failure Analysis 14 (2007) 1–22

beam strength. According to TEC-98, gross section area of column shall satisfy the condition given in the fol-
lowing equation:
N d max
Ac P ð4Þ
ð0:50  fck Þ
Shorter dimension of columns with rectangular section shall not be less than 250 mm and section area shall
not be less than 75,000 mm2. Although minimum dimension of column has been 250 mm since TEC-1975, in
practice there has been observed too many application that contains column dimensions is less than 250 mm.
In the chapter 21 of ACI 318-95 [12], the minimum dimensions of columns is 300 mm. The reason for this is to
satisfy minimum rigidity, decrease the axial load level, and thus increase in ductility.
The longitudinal rebar ratio (ql) ranges between 1% and 4%. In the code, selection of low steel ratio is
encouraged because, low steel ratio is an amplification of larger cross-section. This situation effects ductility
to increase. The authors have observed that in all seismic regions column’s ql is ranges between 1% and 2%.
Diameter (12–16 mm) smooth rebar are generally used but with respect to TEC-98 minimum bar diameter
must be 14 mm in order to prevent buckling. Fig. 5 shows columns with buckled longitudinal bars which expe-
rienced high axial load.
In TEC-98 with respect to transverse reinforcement spacing and detailing requirements, columns are
divided into central zone and confinement zones [4]. Wide spacing of lateral ties is a common shortcoming
observed in Turkish RC structures. During site observations after earthquakes, it is noted that the stirrup
spacing was much more than the maximum value allowed by the design code [13]. In damaged columns,
author did not observe a tie spacing of less than 100 mm and 135 hooks at the end of column ties as shown
in Fig. 6. Observed ties typically had hooks at their ends that were bent only 90. During the earthquake, the
concrete cover outside the ties spilled or fell off, ties loosened (opened up) and the limited strength and con-
finement afforded by the ties were lost. In all codes including TEC-98, special seismic hooks for high ductility
level are required [4,14].
Most observed defect in damaged members is unconfined lap splices (Fig. 7) and inadequate anchorage
lengths (15–30 cm), lower longitudinal reinforcement ratios than considered in design. Lap splices in
moment-frame columns were typically made immediately above the floor framing or the foundation. That
means the lap splices in column were located in a plastic hinge zone that is the most critical region of RC mem-
bers. In the code, lap splices of column longitudinal reinforcement should be made, as much as possible, within
the column central zone defined. The splice length shall be equal to the development length ‘b given in TBC-
2000 [15] for tension bars. In Eq. (5), development length is given
 
fyd
‘b ¼ 0:12 / P 20/ ð5Þ
fctd

Fig. 5. Columns failure with buckled longitudinal bars.


M.H. Arslan, H.H. Korkmaz / Engineering Failure Analysis 14 (2007) 1–22 7

Fig. 6. Wide spacing of lateral ties as well as low quality of concrete.

Fig. 7. Inadequate lap splice.

2.1.1.2. Beams. In the earthquake, first plastic hinges are desired to occur at the end of the beams. As well as
columns, beams are also divided into a central zone and confinement zones (potential plastic hinge sections) in
the TEC-98 with respect to transverse reinforcement. Fig. 8 shows the transverse reinforcement requirements
according to the TEC-98. In Fig. 9, revealing location of central bent-up longitudinal beam-bar, eccentrically
framing beam into column and infrequent stirrups is seen.
According to TEC-98, in all seismic zones special seismic hoops and special seismic crossties must be
bended 135. In practice beam widths ranged between 200 and 250 mm, and transverse ties are smooth rein-
forcing bars of 6–8 mm diameter with 90 hooks. According to TEC-98, width of the beam web shall be at
least 250 mm.
8 M.H. Arslan, H.H. Korkmaz / Engineering Failure Analysis 14 (2007) 1–22

Fig. 8. Transverse reinforcement for beams.

Fig. 9. Damaged beams.

In Turkey, bent-up longitudinal bars are used in order to help moment carrying capacity of supports and
shear capacity of the beam. On the other hand these bent-up bars cannot resist cyclic shear forces. Also during
observations, it is seen that beam bottom reinforcing bars were inadequately anchored through the beam–col-
umn joint [14].

2.1.1.3. Beam–column joints. Many of the beam–column joints are heavily damaged as result of insufficient lat-
eral ties at the beam–column joints. Views of one of the damaged and undamaged joints are shown in Fig. 10. In
damaged joint confinement reinforcement did not exist and beam reinforcing bars anchorage in the joint is inad-
equate. In Table 1, observed applications about detailing and details dictated by the TEC-98 are given in a form
suitable for comparison. The data of Table 1 coupled with field observations clearly show that the TEC-98 were
not followed in detailing RC columns, especially in potential plastic hinge zones at the end of the columns. The
main reason underlying this fact is the lack of inspection and supervision at the design, construction stages.
M.H. Arslan, H.H. Korkmaz / Engineering Failure Analysis 14 (2007) 1–22 9

Fig. 10. Failure of beam–column joint and undamaged joint.

2.1.2. Strong beam-weak column


The use of frames (structures formed by beams and columns) in seismic design seeks to ensure that the dam-
age from intense earthquakes is produced in beams rather than in columns, due to the greater risk of the build-
ing collapsing from damage to the columns [13,16].
In Turkey, deep beams were used with flexible columns. That means, during the dynamic loading, beams
remain elastic whereas columns suffer from shear failure or compression crushing. This undesirable behavior is
illustrated in Fig. 11 which were taken after the 1999-Marmara Earthquake. In Fig. 11, the column depths are
sizeable along the building alley sides, but they are narrow and thereby less able to resist by flexure the seis-
mically-induced horizontal forces parallel to the front street, particularly when compared to the beam depth
and strength at the story above, not surprisingly, the building collapsed swaying in the direction of weakest
column flexural strength. Fig. 12 was taken from a failed building from Marmara Region. This sample build-
ing is a good example in order to explain strong beam-weak column relation. Columns of the structure are
placed such their long dimensions are parallel to the each other. When earthquake hit that structure in a direc-
tion perpendicular to the strong axis, structure could not show enough rigidity.
To the TEC-98, in structural systems comprised of frames only or of combination of frames and walls, sum
of ultimate moment resistances of columns framing into a beam–column joint shall be at least 20% more than
the sum of ultimate moment resistances of beams framing into the same joint. Eq. (6) is given in TEC-98
6
ðM ra þ M r€u Þ P ðM ri þ M rj Þ ð6Þ
5
In the sample building given in Fig. 12, all joint investigation has been done according to the TEC-98. The
summary of all joint moments and plastic hinge location can be seen in Tables 2 and 3. In the weak axes, TEC-
98 dictation cannot be provided in 13 of 20 joints. This kind of failure type has been very common in the
recent earthquakes in Turkey.

2.1.3. Short columns


Short column is a frequently observed problem in Turkey [14,16]. During earthquake, lateral forces are
transmitted to the columns as moment and shear forces. When the length of the columns is lowered, it
becomes stiffer, more rigid in bending and receives higher moments in earthquake. Since the moment arm
is short, the shear forces become more pronounceable. This phenomenon is known as ‘‘short column prob-
lem’’. Short column problem may arise due to arrangement of infill walls or other non-structural, architectural
members. Sometimes, the infill walls are shorter than the column. The top of the walls may be opened for
lightening purposes in schools, hospitals and dormitories. Also, sometimes windows may be opened at the
top of the periphery shear walls in the basement of the structures. In these cases, the column length in between
the wall is confined by the wall and its stiffness is increased. Due to increase in stiffness, the columns share
more flexural moments and this cause increase in the shear forces (see Fig. 13).
10 M.H. Arslan, H.H. Korkmaz / Engineering Failure Analysis 14 (2007) 1–22

Table 1
Observed and code specified detailing of dimensions
Details on 1998 Turkish seismic code Details observed

Number Specification Turkish 1998 Code Observed


1 Minimum dimension of column Width (W) P250 mm 200 mm
Depth (D) P 300 mm
Section area P 75000 mm2
2 Minimum dimension of beam bw P 250 mm 200 mm
3 Joint region tie spacing s1 6 100 mm (unconfined) Not observed
s1 6 150 mm (confined)
4 Beam end zone-tie spacing hooks s2 6 H/4 150-250 mm all 90
s2 6 8/
s2 6 150 mm 135
5 Column end zone-tie spacing hooks s3 6 W/3 Not observed, same as middle zone
s3 P 50 mm
s3 6 100 mm 135
6 Column longitudinal rebar hook Must be avoided 90
P
7 Strong column-weak beam PM columns P 65 61.0
M beams

8 Bottom reinforcement of adjacent span P‘b <‘b


9 Lap splice length in columns P1.25 · ‘b <‘b
P1.50 · ‘b
D: longer column dimension; W: minimum column dimension, H: beam height; bw: beam width; s1, s2, s3: tie Spacing; ‘b: development
length given in TS-500.

In the TEC-98, where short columns can not be avoided, very stringent requirement for transverse rein-
forcement of short columns are put forward. For this, shear force, Ve, to be taken into account for the design
of transverse reinforcement of columns shall be calculated by the following equation:
M a  M u€ max
Ve ¼ 6 V r ! 0:22  Aw  fcd ð7Þ
‘n
In the equation, the moments shall be calculated at bottom and top ends of the short columns as
Ma @ 1.4Mra and M u€ ffi 1:4M r€u . In shear design, the flexural moment is 40% higher than that used in flexure
calculations because, ductile flexural failure is preferred rather than the brittle shear failure in RC members.
Fig. 14 shows a good example of this kind of damage from Kocaeli–Marmara Region.
M.H. Arslan, H.H. Korkmaz / Engineering Failure Analysis 14 (2007) 1–22 11

Fig. 11. Strong beam-weak column failures.

Fig. 12. Columns are placed such that their long dimensions are parallel to the each other.

2.1.4. Soft and weak storey


The other important commonly observed damage type for earthquakes in Turkey is the soft storey effect
[4–14]. Several types of architectural and structural plans lead to the formation of so-called ‘‘soft’’ stories,
12 M.H. Arslan, H.H. Korkmaz / Engineering Failure Analysis 14 (2007) 1–22

which are stories that are more vulnerable to seismic damage than others due to the fact that they are less stiff,
less resistant, or both. In Turkey, generally the first floor levels are used as shopping or commercial purposes.
This can be seen especially in buildings on the side of the main streets. The walls parallel to the street are omitted
and enclosed with glass partitioning walls for presentation purposes. Consider the building in Fig. 15 in which
the moment frame is both flexible and weak in the first storey by comparisons with the upper stores. Deforma-
tions are concentrated in the first storey of this building because the front of the building was open in the first
storey. The first-storey columns in this building were severely damaged and likely close to failure due to gravity
load instability. On the other hand, it is known that, clay infill walls increases the rigidity of the structure and
limits the storey drift. In Table 4, the effect of non-structural infill walls are presented. The increase in lateral
load carrying capacity and decrease in story drift is distinguishable. From the behavior of systems, it is clearly
understood that, infill walls increase about 2–3 times higher lateral load carrying capacity of structural system.
In the code, for columns and structural walls of the ith storey of a building for each earthquake direction
shall satisfy the unfavorable one of the conditions given in the following equation:
Table 2
The summary of all joint moments and plastic hinge location on the weak axe
P P P P
Number of joints Mrc (kN m) Mrb (kN m) Mrc/ Mrb
1 295.5 343.2 0.86
2 305.3 513.6 0.59
3 313.7 522.4 0.60
4 310.2 513.8 0.61
5 291.3 357.2 0.81
6 514.1 363.1 1.41
7 647.3 525.2 1.23
8 596.3 527.2 1.13
9 573.9 522.3 1.02
10 288.1 363.1 0.79
11 564.1 362.3 1.55
12 697.3 531.2 1.31
13 625.7 542.8 1.15
14 593.9 552.3 1.07
15 293.4 373.9 0.79
16 542.2 365.3 1.48
17 667.3 521.2 1.28
18 685.7 552.8 1.24
19 604.9 591.3 1.02
20 303.4 393.9 0.77

Table 3
The summary of all joint moments and plastic hinge location on the strong axe
P P P P
Number of joints Mrc (kN m) Mrb (kN m) Mrc/ Mrb
1 472.8 349.3 1.35
2 493.5 628.8 1.27
3 475.1 350.3 1.35
4 803.7 351.2 2.28
5 814.3 622.3 1.31
6 921.2 363.2 2.53
7 815.4 357.2 2.28
8 812.8 662.3 1.22
9 927.3 371.8 2.49
10 794.2 356.4 2.23
11 788.3 652.6 1.21
12 921.7 375.2 2.45
M.H. Arslan, H.H. Korkmaz / Engineering Failure Analysis 14 (2007) 1–22 13

Fig. 13. Short-column failure due to in fill wall.

Fig. 14. Short column due to additional beam which is not exist in the design.


ðDi Þmax =hi 6 0:0035
ð8Þ
ðDi Þmax =hi 6 0:02=R
Table 5 shows storey drifts values for a building having this type of weakness.

2.2. Shear failure

Shear failure after inelastic cyclic loading is often observed in RC beam or column whose shear strength is
slightly larger than its flexural strength. In the AIJ Design Guidelines (1999), this kind of failure is attributed
14 M.H. Arslan, H.H. Korkmaz / Engineering Failure Analysis 14 (2007) 1–22

Fig. 15. Failure on ground floor due to soft story (Adapazarı).

Table 4
Story drift values
Di
Story number Di (cm) hi (cm) hi maxðDhii Þ
4 2–5 300 0.006–0.017 0.0035 or 0.02/R
3 2–5 300 0.006–0.017 0.0035 or 0.02/R
2 2–5 300 0.006–0.017 0.0035 or 0.02/R
1 (weak or soft story) 15–30 400 0.0375–0.075 0.0035 or 0.02/R

Table 5
The effect of infill brick wall to system behavior [17]
Infill type
Bare frame Type-1 Type-2 Type-3 Type-4 Type-5 Type-6

Fmax (kN) 80 395 220 290 270 235 320


Dmax (mm) 83 14 28 19.5 15.3 22.1 16.8

to the two reasons, reduction of effective compressive strength of concrete due to intersecting flexural–shear
cracks and reduction of aggregate interlocking due to wide flexural–shear cracks [17–20].
As shear failure proceeds, degradation of the concrete core may lead to loss of axial load carrying capacity
of the column. As the axial capacity diminishes, the gravity loads carried by the column must be transferred to
M.H. Arslan, H.H. Korkmaz / Engineering Failure Analysis 14 (2007) 1–22 15

neighboring elements. A rapid loss of axial capacity will result in the dynamic redistribution of internal actions
within the building frame and may progressively lead to collapse. A simple way to check for shear failure in a
frame system with double-curvature columns is to compare the column shear strength, Vn, with the maximum
probable shear force required for the plastic hinge formation at column ends, Vp,
2M p
Vp ¼ ð9Þ
L
Shear force causes tensile stress in concrete in the diagonal direction of the member axis (Fig. 16). After the
concrete cracks under the tensile stress, the stress must be transferred to the lateral reinforcement. Brittle shear
failure occurs in the diagonal tension mode when the minimum amount of lateral reinforcement is not pro-
vided in the member. Fig. 16 illustrated shear effect on a column after Adana Earthquake.
Transverse reinforcement in concrete columns is used to fulfill three main functions. These functions
include restraining longitudinal reinforcement against buckling, increasing shear resistance, and confining
concrete for improved deformability. Short and stubby columns attract shear stresses that may exceed diag-
onal tension capacity of concrete. The excess shear in these columns is resisted by transverse reinforcement.
Shear reinforcement is usually designed following the 45 truss analogy employed in the ACI-318 [21] design
code.
During the last few decades several shear strength model have been proposed and used for the design and
evaluation of RC columns. A short brief of this model is given in Table 6 for a sample column. In Table 6, only
concrete contribution and transverse reinforcement is given. According to the ASCE-ACI Committee 426 Pro-
posals [22], the most critical mechanism were identified as the shear transfer by the transverse reinforcement

Fig. 16. Shear failure mechanism.


16 M.H. Arslan, H.H. Korkmaz / Engineering Failure Analysis 14 (2007) 1–22

and concrete. Shear transfer by uncracked concrete, interface shear transfer in the cracked concrete, aggregate
interlock, dowel shear carried by the longitudinal reinforcement and arch action in deep members is com-
monly neglected.
In lightly reinforced columns after the shear failure degradation of the core concrete may lead to loss of
gravity-load-carrying capacity. A sudden loss of column axial capacity will lead to transfer of column gravity
loads to neighboring frame members with ensuing dynamic redistribution of forces within those members and
a possible subsequent building collapse.

2.2.1. Shear failure of flat-slab construction (punching failure)


A flat slab without column capitals is popular in some regions because it does not have girders below a slab
level. The critical part of the flat slab system is the vertical shear transfer between the slab and column. The
shear failure at the connection leads to ‘‘the pan-cake collapse’’ of the building, leaving no space between the
adjacent floors after the collapse. Serious failure was observed in the 1999 Marmara–Kocaeli Earthquake
(Fig. 17).
Pure punching failure capacity of the connections is defined using the eccentric shear stress model of ACI
318-99 Section 11-12 [23]. Slab-column connections have suffered distress in numerous earthquakes, and in

Table 6
A short brief of shear strength model
Model for shear failure VC VS Vn
Asw fywd d
1 TBC-500-2000 and TEC-98  · 0.65 · fctd
VC = 0.8 ·b·d
pffiffiffiffiffi
VS ¼ s VC + VS
P 0 Asw fywd d
2 ACI-318-02 V C ¼ 2 1 þ 2000A g
fc bd VS ¼ s VC + VS
 
Asw fywd d
3 ASCE-ACI Committee 426 Proposals V C ¼ vc 1 þ f3P 0 bd VS ¼ VC + VS
c Ag s


V C ¼ vc 1 þ pPffiffiffi0 bd
6 fc Ag
pffiffiffiffiffiqffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Asw fywd d
4 SEAOC V C ¼ 3:5 fc0 1 þ 0:002 APg Ac VS ¼ s VC + VS
 pffiffiffiffiffi
Asw fywd d
5 Ascheim and Moehle V C ¼ a0 1 þ 2000A P
c
fc0 bd VS ¼ s VC + VS
pffiffiffiffiffi Asw fywd d
6 Caltrans V C ¼ F 1 F 2 fc0 ð0:8Ac Þ VS ¼ s VC + VS
pffiffiffiffiffi Asw fyw D0 
7 Priestly V C ¼ k fc0 Ae VS ¼ s  cot 30 VC + VS
pffiffiffiffiffi Asw fyw ðD0 cÞ
8 Kowalski V C ¼ abj fc0 0:8Ag VS ¼ s  cot 30 VC + VS
pffiffiffiffiffirffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
P =A

Asw fywd d
9 Konwinski V C ¼ a fc0 1 þ pgffiffiffi0 0:85Ag V S ¼ 0:85  s VC + VS
12 fc
 pffiffiffiffiffi
P Asw fywd d
10 FEMA-273 V C ¼ 3:5k k þ 2000A g
fc0 bd VS ¼ s VC + VS

Fig. 17. Shear failure of flat-plate construction (punching failure).


M.H. Arslan, H.H. Korkmaz / Engineering Failure Analysis 14 (2007) 1–22 17

several cases have contributed to collapse [24–27]. The presence of heavy vertical loads is believed to have
resulted in excessive shear stress on connections, with resultant decrease in connection moment capacity
and ductility and increase in P-delta moments. Combined with the relatively large flexibility of slab-column
frames, several collapses resulted [25,28]. In the event of punching failure at a connection, bottom slab rein-
forcement anchored through the columns has been observed to be an effective means of preventing or delaying
collapse; lack of such reinforcement has been observed to result in catastrophic failures [29].

2.3. Architectural considerations

Some design configurations have advantages during the earthquake. The structures that have symmetrical
configurations both in vertical and horizontal directions behave better. Otherwise torsional effects come into
picture and extra moments will be developed at the columns. Square or circular plans behave better but struc-
tures that have L, H or T type plans can damaged due to stress concentrations on the inner corners (Fig. 18).
Base dimensions and height of the building must be proportional to the each other. The structure in Fig. 19
has 4 m plan dimensions and 13 m height. Also there are no basement in that structure. Since the center of the
mass of the structure is ascended with increase in height, higher overturning moments were developed and the
structure failed.
Frequently, buildings are generally placed adjacent to each other. In that placement configuration, only two
parallel faces of the structure are used and the other two faces, perpendicular to the previous ones, are in con-
tact with the ones of neighboring structure. During the earthquake, due to differences in dynamic properties of
the adjacent buildings, structures hit to each other. If the floors are not at the same elevation, floor of the one
building damaged to the others columns. This is known as hammering effect and can be seen after earthquake
very frequently in Turkey (see Fig. 20). To prevent this damage, suitable space should be left between
two buildings. According to the TEC-98, Minimum size of gaps shall be 30 mm up to 6 m height. From
thereon a minimum 10 mm shall be added for each 3 m height increment. Unless a larger value is obtained

Fig. 18. Stress concentration on the inner corner of an L type structure [30].
18 M.H. Arslan, H.H. Korkmaz / Engineering Failure Analysis 14 (2007) 1–22

Fig. 19. The height and plan dimension of the structure are not proportional.

Fig. 20. Hammering damage and failure.

in accordance with requirement defined in previous statement, sizes of gaps shall not be less than the sum of
the absolute values of average storey displacements multiplied by the coefficient a. If all floor levels of adjacent
buildings or building block are the same then a is equal to R/4. If any of the floor levels of adjacent buildings
or building blocks are not the same then a = R/2. As mentioned previous section R is structural behavior fac-
tor. On the other hand, since the economic value of the lots are very high in Turkey occupants or homeowners
want to use maximum area possible, it is not possible to see gaps between adjacent buildings.

2.4. Material quality

Poor quality of material may have been one of the main factors that caused the collapse of many structures.
Damage due to poor quality of material was reported many of other country’s earthquakes [1,19,31]. Ready
mix concrete is relatively new for Turkey. Ready mix concrete has become popular in Turkey, after 75 years it
started to be used in the world. Hand made concrete is generally used. Near seasides, sand obtained from sea is
used without washing. This sand causes corrosion of the steels bars due to chloride ions. Corrosions carry two
problems, decrease in area of the reinforcing bars and decrease in anchorage.
Turkish labor bewares of the principles of quality of concrete. General mistakes are: (a) lack of proper con-
crete mix design, lack of sieve analysis (oversized aggregates and poor gradation); (b) lack of vibration after
M.H. Arslan, H.H. Korkmaz / Engineering Failure Analysis 14 (2007) 1–22 19

concrete placement; (c) insufficient curing of concrete after placement, without attention to weather condi-
tions; (d) high water/cement ratio for workability.
In the TEC-98 it is stated that, in all buildings to be built in seismic zones, concrete with strength less than
that of C16 (compression strength of concrete is 16 MPa) shall not be used. However it is mandatory to use
C20 (compression strength of concrete is 20 MPa) quality or higher strength concrete in below defined build-
ings to be built in the first and second seismic zones. As mentioned above, lack of anchorage of beams, and
insufficient splice lengths is secondary effect of low quality level of concrete. In various studies on concrete
strength in different regions of Turkey, it is concluded that the average concrete compressive strength in exist-
ing buildings is around 10 MPa [1,13,32,33]. Especially in the Kocaeli Earthquake, average concrete compres-
sive strength that was taken damaged and undamaged buildings was as low as 1/3 of the design strengths very
low and far from TEC-98 and TBC-2000 requirements. Fig. 21 represents low quality of materials used.
Fig. 22 gives a basic information about material strength. As illustrated in Figure, even though the concrete
strength must be emphasized upper than 14 MPa in the project (in spite of TEC-98 restriction), experimental
study shows that real strength is lower.
Usage of low quality steel reinforcement (round bars instead of deformed bars) is also one of the reasons
of structural damage. Particularly low deformation capacity of steel bars seems to be the most important
problem [32,33].

Fig. 21. Low quality of concrete.

Fig. 22. Comparative experiment of concrete compression strength.


20 M.H. Arslan, H.H. Korkmaz / Engineering Failure Analysis 14 (2007) 1–22

2.5. Foundation failures

Foundation failures were observed for many buildings with large settlements, and in some cases, entire
structures overturned. This effect was most pronounced in city of Adapazari. Foundation failures generally
occurred because of soil liquefaction or bearing pressure failures [34,35]. Note that geotechnical evaluations
of site conditions are apparently often not conducted in Turkey, except for important structures.

2.6. Damage due to earthquake-induced liquefaction

Liquefaction is a frequent occurrence in earthquakes. Ground shaking is widely considered to be the primary
cause of damage to structures, loss of life and injuries due to earthquakes [34]. If the liquefaction sensitivity is
high, it is high probability to occur at low level of ground shakings. The liquefaction may initiate at an ground
acceleration value of 0.2 g. In Kocaeli Earthquake, peak ground accelerations (PGA) recorded is between 0.23
and 0.41 g. The 1999 great earthquakes of Turkey contributed to the knowledge of liquefaction-related move-
ments on structural and foundation performance of buildings [36]. The slender buildings with shallow mat foun-
dations enormously deflected from the verticality, whereas the slender buildings with the single foundations

Fig. 23. Bulging of ground adjacent to settled building, Adapazari, Turkey, 1999.

Fig. 24. Tilted building.


M.H. Arslan, H.H. Korkmaz / Engineering Failure Analysis 14 (2007) 1–22 21

manifested both tilting and settlement. The buildings whose height-to-width ratio is smaller than unity experi-
enced settlement as great as 3 m. One such building’s second floor became the first floor due to the sinking into
the liquefied ground [37–39]. There is no a map showing the probability of liquefaction potential of the regions.
Fig. 23 shows the extent of ground heave all around the building as the soil under the building is pushed out-
ward during settlement. As a result of liquefaction, reduction in bearing capacity the soil was reduced and float-
ing buildings were observed. Examples of overturned buildings are shown in Fig. 24.
Most of the buildings that collapsed or suffered structural damage rested on shallow footings. Those struc-
tures were overturned and although there was no structural damage on the frame system, they became
unusable.

3. Results and conclusions

After the last earthquakes (1992–2004), the total damage and the consequent loss of life in Turkey are
humiliatingly too great as compared to those for similar intensity earthquakes recorded elsewhere in the
world. Main structural design and construction mistakes causing collapse of the reinforced concrete buildings
in Turkey are same with the observations after previous earthquakes. Some important lessons for design and
construction practice can be learned from damage observed during last earthquakes. Ductile detailing and
construction is needed to control damage level and to guarantee survival after earthquakes. The inadequate
structural systems and/or members, low quality materials and inappropriate construction are main reasons
of the damage.
Common mistakes can be listed as follows:

 Use of low quality concrete, especially low compressive strength


– Use of sand obtained from sea that contains salt and other chemicals.
– Producing concrete on the site by workers. Use of high water/cement ratios for workability.
– Lack of sieve analysis and proper gradation of aggregate.
– Insufficient vibration and curing of concrete.
– Not obeying concrete mix designs, use of low cement values.
 Mistakes related to selection of settlement sites. Especially settlement on the sites where alluvial deposits
exist and liquefaction danger is high
 Use of non-ductile details and design
– Absence of 135 seismic hooks.
– Lack of transverse ties.
– Strong beams and weak columns.
 Architectural design mistakes
– Placement of columns with orientation of strong axis on the same direction.
– The soft and weak story phenomenon.
– Short column problem.
– Heavy cantilever overhangs.
– Unsymmetrical buildings.
– Adjacent buildings with no separation.
 Construction mistakes, not obeying project of the structure

3.1. What to do for preventing failure of reinforced concrete structures?

Construction of buildings to withstand earthquakes can be possible if and only if the above mistakes are
stopped to be repeated. This can be achieved if engineers, architects and contractors understand principle
of earthquake resistant design and construction principles. An effective control mechanism is needed to pre-
vent similar catastrophes in the future. Also education is main factor to be considered. Workers, craftsman,
contractors, architects and also engineers must be educated about the principles of earthquake resistant design
concept.
22 M.H. Arslan, H.H. Korkmaz / Engineering Failure Analysis 14 (2007) 1–22

References

[1] Cagatay I. Experimental evaluation of buildings damaged in recent earthquakes in Turkey. Eng Fail Anal 2005;12(3):440–52.
[2] General Directorate of Minister Affairs, Earthquake Research Department; 1999.
[3] Aydan Ö. The seismic characteristic and the occurrence pattern of Turkish earthquakes. Turk Earthquake Found 1997. TDV/TR 97-007.
[4] Doğangün A. Performance of reinforced concrete buildings during the May 1, 2003 Bingöl Earthquake in Turkey. Eng Struct
2004;26:841–56.
[5] USGS. Prelimanry Earthquake Report. US Geological Survey, National Earthquake Information Center, World Data Center for
Seismology. Denver. Available from: http://neic.usgs.gov/neis/world/turkey/turkey_faults.html.
[6] Tankut T. Disaster preparedness in Turkey – an overview. Advances in civil engineering, 1–3 November 2000, vol. 1. Gazimagusa,
Turkish Republic of Northern Cyprus; 2000. p. 87.
[7] Turkish Earthquake Code (TEC-98). Regulations on Structures constructed in Disaster Regions, Ministry of Public Works and
Settlement, Ankara; 1998.
[8] UBC-97 Uniform Building Code, International Conference of Buildings Official, USA; 1997.
[9] TEC-75. Ministry of Public Works and Settlement. Specification for structures to be built in disaster areas. Government of Republic
of Turkey; 1975.
[10] Yılmaz Ç, Keyder E, Balkaya C. Evaluation of response modification factor (R). In: Steel structures in the Turkısh Earthquake Code,
Ugur Ersoy symposium on structural engineering, Ankara; 1999. p. 221–34.
[11] Kaltakcı Y, Arslan MH. Comparison of UBC97 and TEC-98 according to base shear force. Turk Chamber Civil Eng J 2005:13–7.
[12] ACI 318-95. American Concrete Institute. Building code requirements for structural concrete. Farmigton Hills (MI); 1995.
[13] Adalier K, Aydingun O. Structural engineering aspects of the June, 1998 Adana-Ceyhan (Turkey) earthquake. Eng Struct
2001;23:343–55.
[14] Sezen H, Whittaker AS, Elwood KJ, Mosalam KW. Performance of reinforced concrete buildings during the August 17, 1999
Kocaeli, Turkey earthquake, and the seismic design and construction practice in Turkey. Eng Struct 2003;25:103–14.
[15] TS-500-2000 (TBC-2000). Turkish Standart Instıtute. TS-500-2000 Building code requirements for reinforced concrete. Ankara,
Turkey; 2000.
[16] 1 May 2003 Bingol Earthquake Engineering Report, Structural Engineering Research Department, Tubitak, Ankara, Turkey.
[17] Kaltakcı MY, Arslan MH. The investigation of brick wall affects on load reduction factor. 119, Earthquake Symposium, 2005,
Kocaeli, Turkey.
[18] Yoshimura K, Kuroki M. Damage to building structures caused by the 1999 Chi–chi earthquake in Taiwan. Available from: http://
www.arch.oita-u.ac.jp/a-kou/taiwan1.pdf.
[19] Lindeburg M. Seismic design of building structures. Belmont (CA): Professional Publications, Inc.; 1998.
[20] Goel RK. Performance of buildings during the January 26, 2001 Bhuj earthquake. Available from: http://ceenve.ceng.calpoly.edu/
goel/indian_eqk/index.htm.
[21] ACI 318, Building Code requirements for structural concrete, Reported by ACI Committee 318.
[22] ASCE-ACI Committee 426, Shear provisions for building codes. ACI Journal, American Concrete Institute September; 1977. p. 458–
69.
[23] EERI. Chi–Chi, Taiwan earthquake reconnaissance report. Earthquake spectra supplement volume; 2001:17.
[24] Arnold C, Reitherman. Building configuration and seismic design. New York: Wiley; 1982.
[25] ‘‘Architects and Earthquakes’’, AIA Research Corporation Report, Washington, DC; 1975.
[26] Bertero VV. Strength and deformation capacity of buildings under extreme environment. In: Pister KS, editor. Structural engineering
and structural mechanics. Englewood Cliffs (NJ): Prentice Hall, Inc.; 1980. p. 188–237.
[27] Kasai K, Popov EP. On seismic design of eccentrically braced steel frames. In: Proceedings of the 8th world conference on earthquake
engineering, San Francisco, vol. V; 1984. p. 387–94.
[28] Ersoy U, Ozcebe G. Reinforced concrete, METU, Ankara, Turkey; 1996.
[29] http://nisee.berkeley.edu/bertero/html/balanced_stiffness_strength_and_ductility.html.
[30] Anonymous. Earthquake-final determinations. Special Publication, Konya Bayindirlik Mudurlugu, Konya, Turkey.
[31] Naderzadeh A, Khademi MH. A preliminary report on the Ardekul, Iran earthquake of 10 May, 1997, Earthquake Hazard Centre
Newsletter, vol. 1(3). No. 7, ISSN 1174-3646; 1998.
[32] Celep Z, Ozer E. Post-earthquake rehabilitation of moderately damaged reinforced concrete structures. In: Proceedings of the 2nd
Japan–Turkey workshop on earthquake engineering; 1998. p. 61–72.
[33] Igarashi S. Recommandations on minimazing earthquake damage in big cities in Turkey. In: Proceedings of the international
conference on the Kocaeli earthquake; 1999. p. 271–86.
[34] Juliet FB, Julian JB. Earthquake losses due to ground failure. Eng Geol 2004;75:147–79.
[35] Bruneau M. Building damage from the Marmara, Turkey earthquake of August, 1999. J Seismol 2002;6:357–77.
[36] EERI Special Earthquake Report. The Izmit (Kocaeli), Turkey Earthquake of August 17; 1999.
[37] Mollamahmutoglu M, Kayabali K, Beyaz T, Kolay E. Liquefaction-related building damage in Adapazari during the Turkey
earthquake of August 17, 1999. Eng Geol 2003;67:297–307.
[38] Pekdamar F. Usage and applications of the performance method to evaluate the existing concrete buildings, Graduate thesis for the
fulfillment of Master’s Degree. Istanbul Technical University, Istanbul; 2003.
[39] Karakostas C, Lekidis V, Makarios T, Salonikios T, Sous I, Demosthenous M. Seismic response of structures and infrastructure
facilities during the Lefkada, Greece earthquake of 14/8/2003. Eng Struct 2005;27(2):213–27.

Вам также может понравиться