Proceeding the 6th Civil Engineering Conference in Asia Region: Embracing the Future through

Sustainability
ISBN 978-602-8605-08-3

HYBRID ESTIMATION OF EARTHQUAKE RESPONSES OF

AN R/C BUILDING DAMAGED BY THE 2009 WEST SUMATRA,
INDONESIA EARTHQUAKE
Yasushi Sanada1, Swe Zin Win2, and Nguyen Khanh2
1
Osaka University, 2-1 Yamadaoka Suita Osaka Japan, [email protected]
2
Toyohashi University of Technology, 1-1 Hibarigaoka Tempakucho Toyohashi Aichi Japan

ABSTRACT
A magnitude 7.6 earthquake occurred on September 30/2009 damaged a large number of R/C buildings at
Padang which is the capital city of West Sumatra, Indonesia. However, no strong motion record was
observed in the most severely damaged area, which causes serious difficulties in analyzing earthquake
damage to buildings in detail. Therefore, the major objective of this study is to estimate seismic responses
of a typical earthquake-damaged R/C building based on the structural damage observed after the
earthquake. This paper focuses on a 7-story R/C building damaged by the earthquake. Damage to the
building is summarized and investigated through several seismic performance evaluations.
To experimentally clarify the seismic performance of seriously damaged columns in the building, a half-
scale R/C column with lap splices was prepared. Quasi-static cyclic loading test of the specimen was
carried out to examine the relationship between structural damage and lateral drift. Moreover, a pushover
analysis was conducted to estimate seismic responses of the entire building based on the damage level-
lateral drift relationship of R/C columns evaluated experimentally.

INTRODUCTION
A disaster investigation team of Architectural Institute of Japan was organized after the 2009 West
Sumatra, Indonesia earthquake to conduct a post-earthquake investigation in Padang which is the capital
city of West Sumatra (Sanada et al., 2009). The objective of the investigation was to collect data on
typical structural details and damage. Padang is situated 60 km far from the epicenter, as shown in Fig. 1
(USGS). A large number of R/C buildings were severely damaged at the downtown area of Padang,
however, no strong motion record was observed at the area, which causes difficulties to analyze
earthquake damage to buildings. Therefore the main objective of this study is to estimate seismic forces
acted on buildings based on the earthquake damage to engineering buildings observed after the
earthquake.
N

Y-dir.
X-dir.

Direction to
epicenter

Fig. 1: Location of earthquake epicenter Fig. 2: Satellite image of Padang

One of the earthquake-damaged buildings focused on by the investigation team was a 7-story R/C
building which was located at the downtown of Padang, as shown in Fig. 2. According to the
 Y. Sanada, S. Z. Win, and N. Khanh

investigation, it was found that one of the major causes of the building collapse was improper
design/construction of columns. Most of the columns in the collapsed story were severely damaged to the
lap splice zone. Therefore, in this study, a structural test of half-scale R/C column with lap splices was
conducted to evaluate the relationship between structural damage and failure behavior. Moreover,
a pushover analysis was performed to estimate seismic forces acted on the entire building. Experimental
and analytical results were combined to propose a procedure to estimate seismic forces acted on damaged
buildings based on their damage levels.

OUTLINE OF INVESTIGATED BUILDING

Structural data
The investigated building was a 7-story R/C building newly constructed in 2007 which was designed
according to the Indonesian seismic provisions published in 2002 (see National Standardization Agency,
2002). Total floor area was 5,815 m2 and building area was 1,707 m2. It was used for office with a car
showroom in the first story. Fig. 3 shows the panoramic view of the earthquake-damaged building before
and after the earthquake attacked. Non-structural brick walls were used for exterior and interior walls and
severely damaged, as shown in the photograph. The plans of the ground floor and 3rd floor are shown in
Fig. 4. The lower floors with a mezzanine have larger floor areas than the 3rd and upper floors. Five types
of design details of the third-story columns are shown in Tab. 1.

7th floor

6th floor

5th floor
Before earthquake
4th floor

3rd floor

Southwest face above the 3rd floor

Collapsed column on the 3rd floor

After earthquake (bottom: close-up)
Fig. 3: Overview of investigated building before and after the earthquake

Earthquake Damage
According to Fig. 3, the third-story columns on the southwest face of the building suffered serious
damage. Concrete cover spalling, remarkable crushing of concrete and exposure of reinforcing bars were
observed in a number of columns. Damage to each structural member was inspected and classified into
one of the damage classes 0 through V summarized in Tab. 2 according to Nakano et al. 2004.
The damage level of each column on the 1st and 3rd floors was also shown in Fig. 4.
The main reasons for severe damage to columns seemed to be as follows:
1) 90 degree hook end of transverse reinforcements might be less effective for confinement against
buckling of main reinforcements and core concrete collapsing during the earthquake.
2) Moreover, these hook ends were located at the same corner, as shown in Fig. 1 (close-up). Therefore,
confinement of core concrete might be decreased.
3) Because of a large amount of main reinforcements in the columns, the lap joint was congested.

TS13-2
 Y. Sanada, S. Z. Win, and N. Khanh

The reasons above were found out based on visual investigations. Therefore, to verify the negative effects
on the seismic performance, a structural test was conducted using a half-scale RC column whose details
are explained in the next chapter. Moreover to quantitatively evaluate damage to each story, residual
seismic capacity index (R) was evaluated by using Eq. 1 presented by Nakano et al. 2004. The damage
criteria based on R is shown in Tab. 3. The value of R and resultant damage level at each floor in both
directions are presented in Tab. 4. As a result, the third floor was evaluated as the most severely damaged
floor in the building.

A
j 0
j

R 100 (%) (1)

Aorg

where, Aj: total number of brittle and ductile columns having damage class 0 through V, Aorg: total
number of investigated columns.

Z
12
600
0
A Tab. 1: Design details of
600 B
0
the third-story columns
0
600

600 C
11 00 0
00 600 D
I0 0 C1 C2 C3 C4 C5
I0 600 E
0

0
600

I0 F
N 10 I0 400
0
0I
I0 II
I0
Column type
0

I0
600

I0
II
9 I0 I0
I0
0

II
350

8 II 00 0I
II BxD 550 x 550 300 x 750 O 650 300 x 850 500 x 500
II I
0

II
430

7 I0 main bar 18-D25 12-D25 24-D25 16-D25 16-D25

II
I I0 hook
I0
I II
D10 @ 150
II
0
450

6 II I0
I0
0 III
I0
0 II 0 II 00 V IV
5 00 00 III IV IV V C4 V C2 IV
4 III IV IV V
2000
3679

I C1 C1 C1 C1
I III IV
I 00 00 00 00 II II 3 III IV
4 II C1
2000 C3 III
I0
3 III
6000

II C1
I
6000

00 00 00 IV III IV I
2
00 00 2
III
III C3 IV III IV I
C3 C5 C1 C1
I
6000

I
6000

00 C1
0I 00 00 I
II 0
1 I
III III II I 0
0 1 III III I I 0
5900

0 I III N C3 C5 C5 C1 C1
00 II II
0I 00 0 I
I 1900 1900
0
II III
X-dir 1900 6000 6000 II III
6000 6000
Y-dir 5250 6000 6000 6000 3200 1900 6000
Z A B C D E F A B C D 3200 E F

Fig. 4: Plans of the 1st (left) and 3rd (right) floors with column damage levels

Tab. 2: Damage class definition of RC columns

Damage class Description of damage
I Visible narrow cracks on concrete surface (crack width is less than 0.2 mm).
II Visible clear cracks on concrete surface (crack width is about 0.2 - 1.0 mm).
Local crush of concrete cover.
III
Remarkable wide cracks (crack width is about 1.0 - 2.0 mm).
Remarkable crush of concrete with exposed reinforcing bars.
IV
Spalling off of concrete cover (crack width is more than 2.0 mm).
Buckling of reinforcing bars.
Cracks in core concrete.
V
Visible vertical and/or lateral deformation in columns and/or walls.
Visible settlement and/or leaning of the building.

TS13-3
 Y. Sanada, S. Z. Win, and N. Khanh

Tab. 3: Damage rating criteria

Damage level R
Slight 95 (%)  R
Light 80 (%)  R  95 (%)
Moderate 60 (%)  R  80 (%)
Heavy R  60 (%)
Collapse R  0 (%)
Tab. 4: Residual seismic capacity index (R) at each floor
R and damage level
Story
X-direction Y-direction
6F 98.7 Slight 98.7 Slight
5F 91.3 Light 91.8 Light
4F 90.5 Light 87.5 Light
3F 49.1 Heavy 48.2 Heavy
2F 97.4 Slight 97.4 Slight
2F (mezzanine) 95.0 Slight 90.2 Light
1F 90.9 Light 95.1 Slight

EXPERIMENTAL PROGRAM

Details of Test Specimen

To experimentally clarify the relationship between damage level and failure behavior of the column,
a half-scale RC column specimen was prepared representing seriously damaged columns in the building.
The dimensions of the column were 275 mm square section with 1700 mm height. The detailed
reinforcements were 20-D13 for main bars with standard lapping and  6@100mm for transverse
reinforcements with a 90 degree hook. The standard lapping has 40 db lap length, where db is the diameter
of the main bar and 1:6 ratio slope at the end of lapped zone. Figure 5 shows the configuration and bar
arrangements of the specimen. Tables 5 and 6 describe the material properties of concrete and
reinforcements.
950
Negative Positive

Q : Lateral Load
450

N : Axial Compression
450

Horizontal jack
( 2000 kN )
1700

Q
N N
W E
6
450
1700

1
Unit:mm
Close-up ( Lapping)
Vertical jack ( 2000 kN )
78

20-D 13
A 6 @ 100 mm Fig. 6: Schematic view of test set-up
A
40
Drift Angle (rad.)
520

20
27

5
27
5

0
450

Section A-A
-20
Unit : mm -3
-40x10

Fig. 5: Details of specimen Fig. 7: Loading history

TS13-4
 Y. Sanada, S. Z. Win, and N. Khanh

Tab. 5: Material properties of concrete

Ec (N/mm2) fc (N/mm2) fcr (N/mm2)
4
2.5 x 10 22.7 2.1
Ec: Young’s modulus, fc: uniaxial compressive strength, fcr: cracking stress in tension.
Tab. 6: Material properties of reinforcements
Bar No. Type Es (N/mm2) fy (N/mm2) c ft (N/mm2)
D13 Deformed 1.77 x 105 431 2833 584
6 Plain 1.60 x 105 264 1724 299
Es: Young’s modulus, fy: yield stress, c: yield strain, ft: tensile strength.

Test Set-Up and Procedure

A schematic diagram of the loading apparatus is shown in Fig. 6. Reversed cyclic horizontal loads were
applied to the top of the column using a 2000 kN capacity actuator which was mounted on a reaction wall.
The specimen was tested under a constant vertical load of 272.3 kN ( 0.18 x b x d x fc, where b and d are
width and length of the column, respectively). The axial load was maintained by manually adjusting the
vertical actuators after each load step. According to the investigation results along the X and Y directions,
no significant difference was observed between damage levels in both directions, as shown in Fig. 4 and
Tab. 4. Therefore, in this test, a 45 degree angle relative to the X and Y directions was adopted as the
direction of horizontal loading. The applied loading history in the lateral direction is shown in Fig. 7.
Relative horizontal displacements were measured by displacement transducers and strains of longitudinal
bars and transverse reinforcements were measured by electrical resistance gauges.

TEST RESULTS

Failure Process
Initiated cracks, crack propagation, and major crack widths were observed at every peak and residual
drifts, to identify the failure mechanism of specimen. The initial flexural crack occurred at the top (east
side) and bottom (west side) around a drift ratio (R) of 1/400 (0.25%). But a shear crack initially appeared
at R of 1/200 (0.5%). Longitudinal and transverse reinforcements exhibited initial yielding near the peak
during the cycle to R=1/100 (1%). Afterward bond cracks could be seen along lap joints at the bottom at
R of 1/133 (0.75%). Compression failure was observed at R of 1.5% during the cycle to 1/67. Figure 8
shows the envelope curve of lateral force-drift ratio relationship of the specimen under positive loads. The
maximum strength of 128.3 kN was observed at R of 1/100 (1.0%). After that, the strength of specimen
started to decline. The specimen finally lost its axial resistance during the cycle to R=1/33. The damage
states of the specimen up to failure are shown in Fig. 10, which were taken after the indicated loading
cycles.
Damage Level
I II III IV V 2.0
Shear crack
Crack width ( mm )

120
Lateral Force (kN)

1.5 Flexural crack

IV & V
80 1.0
III
0.5
R =1/133
R =1/100

40
R =1/200

R =1/33

II

0.0 I
0
-1/67

-1/100
-1/133
-1/200
-1/400
0
1/400
1/200
1/133
1/100

1/67

0 1 2 3 4
Drift ratio ( % ) Drift Angle

Fig. 8: Envelope of lateral force-drift ratio Fig. 9: Shear and flexural crack widths at
relationship with damage level residual drifts
classification

TS13-5
 Y. Sanada, S. Z. Win, and N. Khanh

a) R = -1/200(2) b) R = -1/133(2) c) R = -1/100(2) d) R = -1/67(2) e) R = -1/33(2)

(Damage level I) (Damage level II) (Damage level III) (Damage level IV) (Damage level V)
Fig. 10: Failure process and damage levels of test specimen at residual drifts
The design shear strength of the column Qsu (mean) was 113.4 kN which was calculated by using the
following Eq. 2 for practical design in Japan. A good agreement was obtained even though the strength
was evaluated along the principal axes, because it is not usually applied to evaluate shear strengths in
skewed directions. This means that the design equation can be applied to evaluate shear strengths under
lateral loads along the 45 degree direction adopted in this test.

 0.068 pt0.23 18  f c  

Qsu ( mean)    0.85 pw wy  0.1 0 bj (2)
 M Q d   0.12 

where, pt: tensile reinforcement ratio, M/Q: shear span length, d: effective depth, pw: shear reinforcement
ratio, wy: yield stress of shear reinforcement, 0: axial stress, j: distance between tension and
compression forces.

Modeling of Damage Level

The damage level of column was determined based on crack width and failure behavior. It was
categorized into five damage levels through I to V summarized in Tab. 2 by Nakano et al. 2004. Figure 9
shows the damage level classification in accordance with shear and flexural crack widths observed at
residual drifts. In this figure, however, each damage level is determined based on half values of crack
widths described in Tab. 2, considering the half scale of specimen. As a result, the column damage level
was classified into I until the loading cycle to R=1/200, because crack widths of the specimen were less
than 0.1 mm. After that, crack widths from 0.1 mm to 0.5 mm corresponding to damage level II were
observed during the loading cycle to R=1/133. Although shear crack widths observed at the bottom
during the cycle to R=1/100 were larger than 1.0 mm, the damage under this cycle shown in Fig. 10
indicates less than level IV without remarkable concrete crush and exposure of reinforcements. Therefore,
the damage level was classified into level III until this loading cycle. After the cycle to R=1/67, the
damage level was classified as level IV because reinforcing bars began to be exposed with spalling of
concrete cover. Finally, the specimen was determined as damage level V after the cycle to R=1/33
because reinforcing bars were buckled and large cracks were occurred even in core concrete. Figure 8
shows the classification of damage level with the envelope curve of the lateral force vs. drift ratio
relationship.

TS13-6
 Y. Sanada, S. Z. Win, and N. Khanh

PUSHOVER ANALYSIS OF ENTIRE BUILDING

Three-Dimensional Analytical Modeling

Pushover analysis was conducted to estimate seismic forces acted on the entire building during the
earthquake. Beams were replaced by line elements considering tri-linear flexural and bilinear shear
behavior. Columns consisted of tri-linear flexural springs in the X and Y directions and a multi-shear
spring (MSS). The MSS model can take into account nonlinear shear characteristics in multiple
directions. Material properties were determined based on the design drawing. Compressive strength of
concrete, fc was assumed to be 20 N/mm2, while tensile strength of reinforcement, fy was 400 N/mm2.
Young’s modulus of concrete and reinforcement were 2.13 x 104 N/mm2 and 2.05 x 105 N/mm2,
respectively. Moreover the following assumptions were applied when evaluating flexural strengths of
beams and shear strength of columns. Slabs were considered for flexural strengths of beams. Effective
widths were assumed to be half of span lengths perpendicular to beams. On the other hand, the design
equation, Eq. 2 could evaluate the shear strength of specimen under a lateral load along the axis with a 45
degree angle relative to the principal axes. Therefore, shear strengths of columns at a 45 degree angle
were given by the same equation. With rectangular sections, however, they were calculated as an average
of strengths in the X and Y directions.
Lateral loads were applied to the building from the eastern and western directions, which means that the
loading directions have a 45 degree angle relative to the X and Y axes (refer to Figs. 2 and 4), because of
the same reason as mentioned in the section of “Test set-up and procedure.”

Analysis Results and Discussion

Figure 11 shows the relationship between story shear and inter-story drift for each floor under a lateral
load from the east. The inter-story drift in the third story was found to be larger compared to other stories,
which agreed with the actual damage shown in Tab. 4. Moreover, Figure 12 represents the relationships
between the base/third-story shear coefficient and the third-story drift under lateral loads from the east,
comparing to a damage reproduction ratio at each step, which was defined as the following Eq. 3.

Number of columnswhose damagelevel is consistentto observation

Damagereproduction raito(%)  100 (3)
Total number of columnsin the third story

The maximum value of the damage reproduction ratio was 59% (52%) at a drift angle of 6.2 x 10-3 (6.3 x
10-3) (rad.) under the eastern (western) loading. Figure 12 indicates that the base shear coefficient attained
to about 0.20 at the highest reproduction ratio. Figure 13 shows damage level of each third-story column
at the highest reproduction ratio. Comparing to the observation results in Fig. 4, most of the columns
showed good agreements with the observed damage levels. However, the maximum damage level from
the analysis was level IV whereas the real damage level of some columns reached level V.

3
10x10 1F 4F
Shear Force (kN)

2F (attic) 5F
8
2F 6F
6 3F
4
2
0 -3
0 10 20 30 40 50x10
Drift Angle (rad.)
Fig. 11: Shear force and inter-story drift relationships

TS13-7
 Y. Sanada, S. Z. Win, and N. Khanh

Reproduction ratio Shear Coefficient 0.4

III IV
III IV IV III III IV
0.3 4 III IV IV III

2000
III IV
3 III IV
0.2 III
1F III

6000
0.1 3F
III IV III IV II
0.00 2 III IV III IV II
I
20 I

6000
40 III II III III III
1 III II III III III
60 N
80 X-dir
1900
II
1900
II
Y-dir 6000 6000 6000 II II
100 6000
-3 A B C D 3200 E
0 5 10 15 20x10 F

Drift Angle (rad.)

Fig. 12: Estimation of seismic input based on Fig. 13: Analyzed damage levels of
damage reproduction ratio the third- story columns at the highest
reproduction ratio

CONCLUSIONS
Experimental and analytical studies were performed focusing on a mid-rise R/C building damaged by
2009 Sumatra, Indonesia earthquake. Our study targeted to estimate seismic forces acted on the entire
building based on the damage. Major findings are summarized as follows:
1) Structural test was conducted using a half-scale R/C column specimen which represented typical
damaged columns, especially considering lap splice zone. The column specimen was subjected to
lateral loading along the axis with a 45 degree angle to the principal axes based on the actual column
damage. The specimen failed in shear at the lapping zone prior to flexural yielding. Applying the
Japanese conventional equation for shear design to the specimen, the shear strength of specimen was
found to be evaluated well.
2) Column damage was categorized into 5 grades according to the Japanese guideline for post-
earthquake building damage inspection. The damage level-drift angle relationship was evaluated and
applied to the following analytical investigation.
3) Three-dimensional pushover analysis of the entire building was carried out to evaluate the seismic
performance. The analytical results simulated the drift concentration in the third story, which was
consistent to the observed results from the post-earthquake damage survey.
4) A damage reproduction ratio based on the column damage levels was defined to estimate seismic
forces acted on the entire building during the earthquake. As a result, the base shear coefficient was
attained to 0.20 at the maximum reproduction ratio of 59%.

REFERENCES
Nakano, Y., Maeda, M., and Kuramoto H. (2004), Guideline for Post Earthquake Damage Evaluation
and Rehabilitation of RC Buildings in Japan, 13th World Conference on Earthquake Engineering,
Vancouver, B.C. Canada.
National Standardization Agency: Indonesian National Standard (SNI) (2002), Design Methods
Earthquake Withstand for Building Structures. SNI 03-1726-2002, p. 121.
Sanada, Y., Kishimoto, I., Kuroki, M., Sakashita, M., Choi, H., Tani, M., Hosono, Y., Fauzan,
Musalamah, S., and Farida, F. (2009). Preliminary Report on Damage to Buildings due to the September
2 and 30, 2009 Earthquakes in Indonesia. Proceedings of the Eleventh Taiwan-Korea-Japan Joint
Seminar on Earthquake Engineering for Building Structures, pp. 297-306.
U.S. Geological Survey (USGS), URL: http://www.usgs.gov/.

TS13-8