Content uploaded by Masaki Maeda
Author content
All content in this area was uploaded by Masaki Maeda on Apr 19, 2015
Content may be subject to copyright.
1
On the Seismic Response of the Building of the
Department of Civil Engineering and
Architecture at Tohoku University
Ying Wang,
a)
and Santiago Pujol,
a)
Hamood Al-Washali,
b)
Kazuki Suzuki,
b)
Masaki Maeda,
b)
Susumu Takahashi,
c)
Toshikatsu Ichinose.
c)
The building of the Faculty of Architecture and Engineering at Tohoku University
survived two strong ground motions. Its survival is not surprising because the
structure was stiff and strong. What is more surprising is the fact that the first ground
motion did not cause severe structural damage while the second motion caused so
much structural damage that the building had to be evacuated and demolished. The
damage occurred despite two key facts: 1) the intensities of the mentioned ground
motions are inferred to have been similar, and 2) the building was strengthened after
the first motion (and before the second) following stringent standards.
INTRODUCTION
The building of The Faculty of Architecture and Engineering at Tohoku University is an
ideal study case. It had a fairly regular structural system, its blueprints were clear and well
preserved, it was instrumented and its instruments were well maintained, and it experienced two
strong ground motions with similar intensities (one in 1978 and another in 2011). Between these
two ground motions, the building was repaired and strengthened and this work was documented
carefully. To have conceived and executed a full-scale experiment to produce the information
produced by this building and the dedicated researchers who studied and monitored it through
decades would have taken not only much time and effort but also a prohibitive amount of
money. This paper reviews the properties and history of the building and focuses on the damage
caused by the Tohoku Earthquake of 2011 and plausible explanations for it.
a)
Purdue University, 550 Stadium Mall Dr., W. Lafayette, IN 47907, USA
b)
Tohoku University, 6-6-11, Aobayama, Aoba-ku, Sendai, JAPAN
c)
Nagoya Institute of Technology, Gokiso, Showa, Nagoya, JAPAN
2
BRIEF DESCRIPTION OF THE STRUCTURE
The building was a nine-story composite building built in 1969 (Figures 1 and 2). It had a
two-story podium with a seven-story tower above it. The structural system consisted of a
combination of structural walls and three-dimensional frames. The building was instrumented
with accelerometers installed in the first and ninth stories (Shiga et al., 1981).
STRUCTURAL LAYOUT
In the lower two stories, the building had eight bays and two overhangs in the East-West (E-
W) direction, and four bays in the North-South (N-S) direction. The floor plan of these two
stories had the shape of the letter H, with outer dimensions of 72 m (E-W) by 36.6 m (N-S). The
floor plans of the upper seven stories were rectangular, with five bays in the EW direction, two
bays and two overhangs in the NS direction, and outer dimensions of 40 m (E-W) by 17.2 m (N-
S). The total floor area was 9200 m
2
. Story heights measured from top-of-slab to top-of-slab
were 5 m for the first story, 4.3 m for the second story, 3.8 m for the third story, and 3.3 m for
the rest. Figure 2 shows floor plans and elevations of the building.
The lateral-load resisting system was a combination of frames and structural walls. Frames
were not discontinued where structural walls were present. Instead, frame elements (both
columns and beams) with the same dimensions as frame elements elsewhere were cast integrally
with the walls. This arrangement resulted in wall boundary elements with the same dimensions
of columns located away from walls. The layout of walls and columns is shown in Figure 2.
Details about the structure are given in Appendix A and
nees.org/warehouse/experiment/3641/project/1122.
In each plan direction, there were two parallel walls. They were continuous from foundation
to roof. In the E-W direction both of these walls were located along the mid column line (axis C).
In the N-S direction these walls were located along the exterior column lines of the upper seven
floors. In addition, C-shaped walls were located next to the northernmost column line (axis D).
The floor system consisted of flat slabs supported by frame girders and intermediate beams
framing into these girders at their midspans. The foundations of the building were spread
footings connected by grade beams.
3
STRUCTURAL DETAILS
Beam, column, and wall dimensions and reinforcement are listed in Appendix A. Beams and
columns were reinforced with steel angles, plates, and reinforcing bars in both their transverse
and longitudinal directions. Essentially, the angles and plates formed a lattice frame that was cast
within a concrete frame.
The specified compressive strength of the concrete was 210 kgf/cm
2
(21 MPa, 3000 psi). The
mean compressive strength of samples extracted from the building in April 2011 was 180
kgf/cm
2
(18 MPa, 2600 psi) (Kuji, 2011). The mean compressive strength of cores extracted
from the third story was 150 kgf/cm
2
(15 MPa, 2100 psi).
The longitudinal reinforcing steel bars were specified to meet Japanese Standard SD35
(nominal yield stress of 3500 kgf/cm
2
-345 MPa, 50 ksi-, expected yield stress of 4000 kgf/cm
2
-
390 MPa, 57 ksi-). Steel angles were specified to meet Standard SS40 (nominal yield stress of
2400 kgf/cm
2
-235 MPa, 34 ksi-, expected yield stress of 3000 kgf/cm
2
-295 MPa, 43 ksi-).
Transverse reinforcing bars were specified to meet SR24 (nominal yield stress of 2400 kgf/cm
2
-
235 MPa, 34 ksi-).
MEASURED PERIOD AND ESTIMATED BASE SHEAR STRENGTH
The fundamental or first-mode period of the building has changed over the years. The
stiffness of the structure has changed because of cracking caused mainly by earthquakes and
because of strengthening done in 2001. Numerical analyses of a linear-elastic model of the as-
built building made ignoring the flexibility of the foundation soil indicate that, in the longitudinal
direction (E-W), its initial period was approximately 0.5 seconds while in the transverse
direction (N-S) it was 0.4 seconds. Motosaka et al. (2004) report initial periods approximately
equal to these values for displacement amplitudes not exceeding approximately 1/10
5
times the
building height.
Published limit analysis results (Suzuki et al., 2013) and (Kimura et al., 2012) show that the
base shear strength of the building was 1) likely to have been between 0.3 and 0.5 times its
weight
a
and 2) controlled by a flexural failure mechanism with hinges in columns and walls at
the base of the third floor.
a
The variation being related mainly to differences in the assumed distribution of lateral forces.
4
RESPONSE TO THE EARTHQUAKE OF 1978
One of the major earthquakes that affected this building was the 1978 Miyagi-Ken-Oki.
Figure 17 shows acceleration records obtained by Tsamba and Motosaka (2011) at the first story.
The peak ground acceleration was 0.26 g in the N-S direction, 0.21 g in the E-W direction, and
0.16 g in the vertical direction. The peak ground velocity was approximately 0.35 m/s in the N-S
direction, and 0.25 m/s in E-W direction.
The earthquake caused shear and flexural cracks (with thicknesses not exceeding 1mm) in
the exterior shear walls (Figure 4), short beams, and a few columns of the third and fourth stories
(Shiga et al., 1981). Approximately 2.5% of the windows broke and furniture overturned.
THE RETROFIT OF 2001
In 2001 the building was retrofitted to reduce torsion and increase the shear strength of the
exterior walls in the N-S direction. The retrofit was limited to the upper seven stories. The
concrete of the webs of the exterior walls in the N-S direction (Axes 2 and 7) was replaced with
thicker cast-in-place webs made with concrete with a cylinder compressive strength of 300
kgf/cm
2
(29 MPa, 4.3 ksi). Steel jackets were mounted on short beams in the interior frames
(Axes 3 to 6 between Axes C and D) in the N-S direction. These beams were expected to be
vulnerable to shear. Steel braces were fitted into two bays of the southernmost frame (Axis B
between Axes 3 and 4 and Axes 5 and 6) in the E-W direction to reduce torsion, and portions of
the floor slabs (between Axes 2 and 3 and Axes 6 and 7) were thickened and reinforced with
additional steel welded wire. Figure 5 shows photographs taken during the retrofit.
Figure 6 shows details of the reinforcement used in the replaced wall panels. As the existing
concrete of the webs of exterior walls (oriented in the N-S direction) was removed, the existing
reinforcement was cut 0.20 m away from beams and columns. New web longitudinal and
transverse reinforcement was provided in two layers (see Table 1 for details). Post-installed
anchors (deformed headed studs with 13-mm shafts and spaced at 0.10 m) were glued 110 mm
into beams and columns, and were embedded 260 mm in the new webs. These lengths were
determined using Japanese design provisions for shear transfer. Spiral reinforcement, with a bar
diameter of 6 mm, a pitch of 50 mm, and an outer diameter of 120 mm, was provided around the
5
perimeter of the new webs. The spirals were used to prevent splitting and sliding failure at
anchor bolts.
Table 1: Retrofit details for the walls on Axes 2 and 7
Story
Web
Thickness
(mm)
Web reinforcement Anchors
Layers Bar diameter (mm) and
spacing (mm) Layers Bar diameter (mm) and spacing
(mm)
9 180 2 Longitudinal D10 @200,
Transverse D13 @200 1 D13 @ 100
8 180 2 Longitudinal D10 @200,
Transverse D13 @200 1 D13 @ 100
7 180 2 Longitudinal D10 @200,
Transverse D13 @200 1 D13 @ 100
6 180 2 Longitudinal D10 @200,
Transverse D13 @200 1 D13 @ 100
5 200 2 D13 @ 200 1 D13 @ 100
4 200 2 D13 @ 200 1 D13 @ 100
3 250 2 D13 @ 150 2 D13 @ 150
RESPONSE TO THE EARTHQUAKE OF 2011
The building withstood the March 11th 2011 earthquake but not without heavy structural
damage. It was demolished because the cost of repair was deemed too high for a structure
nearing the end of its expected life span. Figure17b shows acceleration records obtained at the
base of the first story. Peak ground acceleration was approximately 0.34 g in both the N-S and E-
W directions, and 0.26 g in the vertical direction. Peak ground velocity was approximately 0.45
m/s in the N-S direction, and 0.50 m/s in the E-W direction. At the ninth-story the maximum
acceleration recorded in the transverse direction was 0.93 g (Figure 18b).
The damage concentrated at the base of the third floor in the exterior walls oriented in the N-
S direction (Figure 7). The concrete in the boundary elements of these walls disintegrated along
heights of up to 0.8 m. Longitudinal reinforcement buckled and/or fractured. The joint between
the original beams and the web of the walls cast in the retrofit of 2001 was damaged: the
concrete at the top of the beams spalled and top beam reinforcement was exposed. The interior
C-shaped walls in the third floor had shear and flexural cracks, and spalling at the level of the 4th
floor.
6
Surveys of the building were done in October 2011, March 2012, and June 2012. Figure 8
shows crack maps obtained in those surveys for Axes 2 and 3.
KEY OBSERVATIONS AND INFERENCES
The structure was shored snug during the retrofit of 2001 to avoid large increases in the axial
loads in columns along the retrofitted axes (2 and 7). The top portions of the replaced web
“panels” were cast using a concrete mix designed to reduce shrinkage, which could also have
caused increases in column axial load. No clear signs of shrinkage in the replaced webs were
observed in the inspections made after the 2011 earthquake.
Figure 9 shows one of the columns that disintegrated during the 2011 earthquake. All
deformed bars buckled. Some steel angles also buckled and remained buckled but, interestingly,
other angles did not, and instead they remained straight and had fractures.The angles that did not
have buckled shapes after the earthquake are likely to have fractured at welds or rivet holes along
the clear height of the column. The angles that remained buckled must have either 1) fractured at
rivet holes in beam-column joints or 2) accumulated large plastic tensile deformations followed
by compression. Studies of the steel samples extracted by Takenaka Corporation from damaged
columns indicate that, away from rivet holes and welds, the inspected reinforcing angle segments
did not reach large plastic deformations. The observed buckling may have resulted from a
sequence of events starting with fracture in the beam-column joint followed by pullout and
ending with buckling (Figure 11). The buckling of vertical reinforcing angles may have triggered
the observed spalling of the column concrete shell. Wall sliding is unlikely to help explain the
combination of buckling and fractures observed. The following observations provide more
insight:
a) the shape of the cross sections of interior (C-shape) and exterior walls would indicate that,
as can be confirmed by analysis, the webs of interior walls were at least as vulnerable in
compression (caused primarily by flexure) as the boundaries of the exterior walls under
southward inertial forces. The webs of interior walls did show cover spalling at their connections
with 4
th
-floor beams. Nevertheless, the level of damage was not comparable to the damage
observed in the boundary elements of exterior walls.
b) Figure 7b shows the damage caused by the 2011 earthquake to the intermediate column (at
intersection 2C) in the exterior structural wall on the east elevation of the building. The deformed
7
reinforcing bars buckled despite the fact this column was unlikely to experience large
compression forces during ground motion. Vertical splitting cracks in the column may have been
the result of reinforcement buckling. Such cracks were not visible in the web, where damage
tended to concentrate around the bottom ends of the anchor bolts installed in 2001.
Observations a and b above support the idea that spalling was caused or accelerated by the
buckling of reinforcement with large permanent tensile deformations or pullout failure following
fracture in beam-column joints. A plausible failure process includes these steps:
1) angles fractured in tension in beam-column joints or near the lower ends of columns,
2) deformed bars reached large tensile strain near fractures in angles,
3) deformed bars buckled as the load reversed forcing them to work in compression,
4) angles that had fractured in joints buckled as the load reversed (Fig. 11) while those that had
fractured in the columns are likely to have remained straight after fracture.
Figure 12 shows the top of a 3
rd
–floor beam after the 2011 earthquake. The figure shows:
1) Anchor bolts that pulled out of the beam forming “conical” failure surfaces in the concrete
2) Buckled plain vertical bars which were part of the original web reinforcement
3) Beam stirrups
4) Beam top longitudinal reinforcement
The buckled plain bars were embedded in the webs cast in 2001 approximately 20 bar
diameters. They are unlikely to have developed their strength in such a short length. Again, the
observed buckling may be the result of pullout failures followed by compression. Figure 13
shows a cross section of the web-beam joint as modified in the retrofit. Observe that the dotted
line does not cross any reinforcement anchored effectively to resist large tensile forces. At this
location, the exterior walls were essentially unreinforced. A tensile failure at this location is
likely to have occurred at a small wall drift and may have altered drastically the response of the
wall as discussed later. Pullout of the anchor bolts explains the concentration of damage seen in
the web. This type of failure could have been avoided had the concept of “capacity design” been
followed closely (Sullivan, 2010). The anchorage of reinforcement needs to be stronger than the
reinforcement itself.
8
The cracking away from the 3rd floor beams is also revealing and shows that shear
deformations (and cracks) were larger in 1978 than in 2011, when the deformations seem to have
concentrated at horizontal wall-beam joints.
THE CONUNDRUM
The key question is why was there more damage in 2011 despite the strengthening done in
2001? We see two plausible explanations 1) the demand was higher, and/or 2) the retrofit made
the building more vulnerable.
WAS THE DEMAND AT THE SITE HIGHER?
It is clear that the earthquake of 2011 (The Great East Japan Earthquake, Mw 9.0) was a
larger event than the 1978 earthquake (The Miyagi-Oki Earthquake, Ms 7.7). The critical
question is whether the intensity at the site was larger.
Within 0.5 km, at least four other buildings that survived the 1978 event were evacuated and
demolished after the 2011 event. We do not know all the details about their state after the 1978
earthquake so we cannot make direct comparisons on this basis. Nevertheless, we do know that
1) 4 out of 13 buildings inspected in the area in 1978 had 1.5-mm cracks in structural walls and
were deemed to require structural repairs, and 2) in two of the buildings surveyed in 1978, the
damage in 2011 took place mostly in coupling beams, which were not inspected in 1978.
In terms of PGA, the ground motion of 2011 was 30% more demanding. In terms of PGV,
the increase in demand was also 30%. Nevertheless, PGA and PGV are far from being infallible
intensity indices. A better index is linear spectral displacement. Figure 14 shows that, for
practical purposes, the linear spectra from 1978 and 2011 were essentially equal. This
coincidence is remarkable and makes this case extraordinary. Because linear spectra are not the
only vehicle available to estimate displacement demand, we also considered nonlinear spectra.
Nonlinear spectra are not as crisp as linear ones in that they depend on many parameters in
addition to the ratio of mass to initial stiffness (i.e. post-cracking stiffness, strain-hardening
stiffness, unloading stiffness, reloading stiffness, etc.). Nearly 700 dynamic analyses of nonlinear
SDOF systems were conducted to try to understand to what extent nonlinear oscillators may have
been more sensitive to the 2011 ground motion. The oscillators considered are described in Table
2 and Figure 15. The analysis results are summarized in Figure 16.
9
Table 2:
Oscillator Parameter
Bilinear Trilinear Simplified Takeda
(Otani, 1974)
k2/k1
1 0.5 1
k3/k1
0.05 to 0.1 0 to 0.1 0 to 0.1
k4/k1
1 1 (
∆
y/
∆
max)
0.5
Fy/Weight
a
0.3 to 0.4 0.3 to 0.4 0.3 to 0.4
Fcr/Weight
0.5 x Fy/Weight 0.5 x Fy/Weight 0.5 x Fy/Weight
Viscous Damp. Coeff.
0.02 0.02 0.02
On the basis of the nonlinear-analysis results obtained, it seems reasonable to conclude that
the 2011 record did not produce consistently larger displacement demands in structures with the
required toughness.
From the evidence presented we cannot conclude that the demand in 2011 was drastically
larger than in 1978. We now turn to the recorded response to try to infer the displacement
demand at which the failure of 2011 may have started.
THE NS RECORDS
Figures 17 and 18 show segments of acceleration records obtained at the site and generously
provided to us by Prof. Motosaka of Tohoku University (2011). A quick inspection reveals a
large difference in duration. The 2011 motion lasted more than 4 times the duration of the 1978
motion. We modified the acceleration records by removing signals with periods exceeding 6s
(for records from 1978) and 16s (for records from 2011). The velocity records obtained by
integrating the resulting acceleration records were also modified by removing the mean velocity.
The modified velocities were integrated to obtain estimated displacements. Figure 19 shows
segments of relative displacement histories estimated from the 1978 and 2011 records. They
were computed subtracting computed base displacements from computed 9
th
-floor
displacements. We concentrate on the NS direction because this was the direction of the walls
that failed.
a
If one increases the base- shear coefficient beyond 0.4, the results are expected to get closer to the linear results.
10
The acceleration and relative-displacement records are rich with information. Among other
things they show that the peak relative displacement in 1978 was approximately 21 cm. They
also show that the effective period of the building was approximately 1 sec. in the same ground
motion. The same is true for the initial part of the 2011 record
a
. Nevertheless, 82.2 sec. into the
2011 motion something occurred to the structure. The 2011 acceleration record reached a plateau
(Figure 20) indicating that yielding may have occurred
b
as the displacement reached the
maximum displacement reached in 1978 (21cm). As the building swayed in the other direction a
radical event took place. The acceleration reached another plateau at approximately 82.7 sec. and
soon after it decreased abruptly. The acceleration plateau was reached at an estimated
displacement of approximately 21 cm. The abrupt acceleration drop occurred at an estimated
displacement of nearly 23 cm. We do not claim to have the accuracy to estimate these
displacements within 1 cm. We display the unwarranted number of significant figures simply to
stress that the observed drop in acceleration occurred at a displacement comparable to the
maximum displacement reached in 1978 confirming what the linear displacement spectra
suggested (i.e. that the displacement demands were similar in 1978 and 2011). Additional abrupt
drops in acceleration took place at 83.2 sec. and 83.7 sec. From that instant on, the structure had
an increased effective period of approximately 1.2 sec.
The fluctuations in acceleration could be attributed to 1) the observed fractures in the
boundary wall reinforcement, 2) the effects of higher modes of vibration, or both. We find option
1) more likely because:
a) the change in period that took place after the acceleration drops indicates a large and
abrupt change in stiffness that cannot be explained by referring to higher modes, and
b) analyses of MDOFs and the shape of the computed response spectra (which show high
amplifications for periods close to 1 sec.) tell us that the response of the structure was
dominated by its first mode.
a
Keep in mind that between 1978 and 2011 the building 1) was strengthened and 2) experienced several smaller
earthquakes.
b Analyses of MDOFs and the shape of the response spectra (which show higher amplifications for periods close to
1 sec.) tell us that the response of the structure was dominated by its first mode, which allows us to infer from
acceleration-displacement plots how the stiffness and the strength of the structure varied over time.
11
It is also reasonable to expect the observed fractures to have taken place when peak
accelerations were reached (i.e. when the lateral forces peaked).
Notice that before 82 sec. the relative displacement computed from the 2011 record did not
exceed 10 cm more than 3 or 4 times, and that it never reached 15 cm. This observation suggests
that the increase in duration is unlikely to have been the cause of the failure as most of the cycles
seem to have taken place well within the linear range of response.
Because the response of the structure was dominated by its first vibration mode, we examine
next the relationship between absolute acceleration measured at the 9th floor and the relative
displacement estimated for the same level. We do so expecting this relationship to provide us
with information about how the stiffness and the strength of the structure varied over time.
ABSOLUTE ACCELERATION – RELATIVE DISPLACEMENT RECORDS
Figures 21 and 22 show absolute acceleration – relative displacement curves estimated based
on the records obtained on the 9th and 1st floors in the NS (transverse) direction. Interpreting
these plots is not simple because they are sensitive to 1) the effects of higher modes and 2) the
modifications made to the records to obtain sensible displacement estimates. With these
limitations in mind we do notice in them the following consistent trends:
1) Wide hysteretic loops were observed only in the first 15 sec. of the 1978 motion (Fig.
21a)
2) After that instant, the structure responded nearly as an SDOF showing no clear evidence
of stiffness decay or yielding until 82.2 sec. into the 2011 motion (Figs. 21b-d)
3) Yielding was first reached at a displacement of 20 to 21 cm (points A, A’, Figs. 20, 22a)
4) At 23 cm the first large drop in acceleration (or strength) took place (point B, Figs. 20,
22a)
5) Two consecutive acceleration drops took place between 83 and 84 sec. (points C, D, Figs.
20, 22b)
6) After these drops, the structure was softer and retained the resulting stiffness for the rest
of the motion
12
7) The peak acceleration in 1978 was higher than the peak acceleration in 2011 indicating
that the strength of the structure may have decreased because the failure of anchor bolts
occurred.
SYNTHESIS
The bulk of the evidence presented points in a single direction: the column failures of 2011
are likely to have taken place at a displacement similar to the maximum displacement reached in
1978. Despite this inferred similarity, and despite the strengthening done in 2001, the damage
caused by the 2011 motion was dramatically different. If the demands were not higher, it is
reasonable to conclude that the structure was more vulnerable in 2011. The observed damage
hints that the source of the vulnerability was the web-beam connections modified in the
strengthening of 2001. It is plausible that the discontinuity in the vertical web reinforcement
created during the retrofit of the exterior walls led to a weak (essentially unreinforced) plane at
the base of the third story exterior wall webs. Wall deformations seem to have concentrated at
this level during the earthquake of 2011. The concentration of deformations must have in turn led
to larger unit tensile strains in the reinforcement (Wang, 2014). This plausible increase in strain
could have resulted in the fractures observed and higher probability of buckling of longitudinal
column reinforcement. If “low-cycle fatigue” problems commenced in 1978, the increased
strains must have accelerated them. Had the connection between the reinforcement replaced in
2001 and the rest of the structure not failed, it is likely that the response of the structure in 2011
would have been more similar to the response observed in 1978. Boundary element confinement
could also have helped reduce the damage seen in 2011 by restraining buckling. Confinement,
nevertheless, would not have helped in preventing the observed tensile fractures.
CONCLUSIONS
The available evidence suggests that discontinuities in reinforcement introduced during
retrofit work done in 2001 caused concentration of deformations that led to the failures of wall
boundary elements in the Building of the Faculty of Architecture and Engineering at Tohoku
University during the Tohoku Earthquake of 2011. Reinforcement discontinuities ought to be
avoided at critical sections of elements expected to resist lateral forces induced by earthquakes.
13
ACKNOWLEDGMENTS
We are grateful for the support received from the Japan Science Foundation and the National
Science Foundation of the U.S.A. We also would like to express our gratitude to Takenaka
Corporation, in particular to Dr. H. Kimura, and to all the engineers who helped inspect the
building described here and provided us with feedback about our interpretations. Special thanks
are due to Enrique Villalobos for his help in obtaining and organizing field data. The insight and
guidance provided by Professors A. Shibata and Mete Sozen are invaluable to us. Our work
would not have been possible without the generosity of Professor M. Motosaka.
14
REFERENCES
Shiga, T., Shibata, A., Shibuya, J. and Takahashi, J., 1981. Observations of Strong Earthquake Motions
and Nonlinear Response Analysis of the Building of Architectural and Civil Engineering Department,
Tohoku University, Transactions of the Architectural Institute of Japan (301), 119-129 (in Japanese)
Motosaka, M., Sato, T. and Yamamoto, Y., 2004. Amplitude Dependent Dynamic Characteristics of an
Existing Building Before and After Seismic Retrofit, Proceedings of 13th World Conference on
Earthquake Engineering, Vancouver, B.C., Canada, Paper No. 1023
Kuji Architecture Studio, 2011. Aobayama Campus of Tohoku University (053) The Building of the
Faculty of Architecture and Engineering, Seismic Evaluation Report (Seismic Evaluation considering
the Damage Caused by the 2011 Tohoku Earthquake),
Tsamba, T. and Motosaka, M., 2011. Observational records’ analyses for dynamic characteristics of a
damaged building during the 1978 Miyagi-ken Oki and the 2011 Tohoku Earthquakes, 30th Meeting
of Japan Society of Natural Disaster Science
Suzuki, K., Al-Washali, H., Maeda, M., Wang, Y., Pujol, S. and Ichinose, T., 2013. Performance of the
Building of the Faculty of Engineering at Tohoku University During the Great East Japan Earthquake
of 2011, Proceedings of 10th International Conference on Urban Earthquake Engineering, Tokyo
Institute of Technology, Japan
Kimura, H., Hirabayashi, M., Ishikawa, Y., Tanabe, Y., Maeda, M., and Ichinose, T., 2012. Investigation
on Buildings in Tohoku University Damaged by the 2011Great East Japan Earthquake, Part3 and
Part4 Study on Building of Civil Engineering and Architecture by Earthquake Response Analysis,
Proceedings of Architecture institute of Japan (AIJ) annual meeting, Nagoya, Japan (in Japanese)
Francisco, E. and Motosaka, M., 2006. Dynamic Response Analysis Considering Structural Deterioration
of a Reinforced Concrete Building Using Earthquake Observation Records, Master Thesis submitted
to Department of Architecture and Building Science, Graduate School of Engineering of Tohoku
University
Fujihashi, K. and Inoue, N., 1997. Investigation of Vibration characteristics and seismic evaluation of
Civil Engineering Building of Tohoku University. Master Thesis submitted to Department of
Architecture and Building Science , Graduate School of Engineering of Tohoku University, Japan (
in Japanese)
Shiga, T., Shibata, A., Shibuya, J. and Sato, N., 1981. Earthquake Damage, wall ratio and column ratio of
middle-rise reinforced buildings (part 1), Transactions of the Architectural Institute of Japan for
Tohoku region. Page 5-8(in Japanese)
15
Otani, S., 1974. Inelastic analysis of R/C frame structures, ASCE Journal of Structural Division, V.100,
p1433-1449
Sullivan, T.J., 2010. "Capacity Design Considerations for RC Frame-Wall Structures." Earthquakes and
Structures, Vol. 1, No. 4, pp.391-410.
Wang, Y., 2014. Effects of Web Reinforcement Discontinuities on the Seismic Response of Structural
Walls, Ph.D. Thesis, Purdue University, West Lafayette, IN.
16
Figure 1: Building of the Faculty of Architecture and Engineering at Tohoku University
17
(a)
Figure 2: Plan views: (a) stories 1 and 2, (b) stories 3 to 9 (dimensions in mm)
N
9
8
7
6
5
4
3
2
1
8000
Typ.
10050
Typ.
E
D
C
B
A
10050
6750
6750
18
(b)
Figure 2: Plan views: (a) stories 1 and 2, (b) stories 3 to 9 (dimensions in mm)
N
accelerometer
D
C
B
6750
6750
7 6 5 4 3 2
8000 Typ.
19
Figure 3: Wall layout, and beam and column reinforcement details. See Appendix A for details.
N
20
(a) (b)
Figure 4: Damaged observed in 1978: (a) crack map of the exterior wall along axis 2 (Shiga et al., 1981),
(b) photo taken after the earthquake of 1978 showing cracks in second, third, and fourth stories.
21
(a) (b)
(c) (d)
(e)
Figure 5: Retrofit of 2011: (a) reinforcement at wall web-beam connection, (b) replaced wall web
reinforcement, (c) exterior steel braces in South facade, (d) steel jacket on short beams, and (e) slab
thickening (Courtesy of Tohoku University)
22
Figure 6: Details of the wall web reinforcement installed in 2001 (Appendix A)
23
(a)
(b)
Figure 7: Damage to exterior walls after the earthquake of 2011: (a) disintegrated boundary element, (b)
damage to beam-web joint and intermediate column
24
Figure 8: Crack maps obtained after the earthquake of 2011
Blocked Area
Max. Crack width
away from areas
with spalling=
0.5mm
Max. Crack width
away from areas
with spalling=
1.7mm
25
Figure 9: Base of boundary element at third floor
Figure 10: Fracture at weld
26
Figure 11: Plausible failure sequence (dimensions in mm)
Pullout following fracture
Buckling
27
Figure 12: Close up of web-beam joint
Figure 13: Cross section of web-beam joint
P
ull
out
failure
surface
28
Figure 14: Linear displacement spectra
Figure 15: Load-displacement curve for nonlinear oscilators
0.00
0.10
0.20
0.30
0.40
0.50
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Displacement Response (m)
Period (sec)
1978NS-2%Damping
2011NS-2%Damping
k
1
k
2
k
3
k
4
Force
Displacement
Fcr
Fy
∆y
∆max
29
Figure 16: Nonlinear displacement spectra
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2
Relative Displacement [m]
Period [s]
Nonlinear Displacement Spectra - Spread from 714 Runs
1978
2011
(a) 1978
(b) 2011
Figure 17: Ground acceleration records
40 60 80 100
1−
0
1
Time [s]
Ground Acceleration [g]
0 20 40 60
1−
0
1
Time [s]
Ground Acceleration [g]
(a) 1978
(b) 2011
Figure 18: Ninth-floor acceleration records
40 60 80 100
1−
0
1
Time [s]
9th Floor Acceleration [g]
0 20 40 60
1−
0
1
Time [s]
9th Floor Acceleration [g]
(a) 1978
(b) 2011
Figure 19: Ninth-floor relative displacement
0 20 40 60
40−
20−
0
20
40
Time [s]
9th Floor Displacement [cm]
40 60 80 100
40−
20−
0
20
40
Time [s]
9th Floor Displacement [cm]
(a) acceleration
(b) displacement
Figure 20: Close-up views of ninth-floor acceleration and relative-displacement histories (2011)
81 81.2 81.4 81.6 81.8 82 82.2 82.4 82.6 82.8 83 83.2 83.4 83.6 83.8 84 84.2 84.4 84.6 84.8 85
1−
0
1
Time [s]
9th Floor Acceleration [g]
81 81.2 81.4 81.6 81.8 82 82.2 82.4 82.6 82.8 83 83.2 83.4 83.6 83.8 84 84.2 84.4 84.6 84.8 85
40−
20−
0
20
40
Time [s]
9th Floor Displacement [cm]
A
A’ B
C
D
A
A’
B
C
D
40−20−0 20 40
1−
0
1
9th Floor Relative Displacement [cm]
- Absolute Acceleration [g]
40−20−0 20 40
1−
0
1
9th Floor Relative Displacement [cm]
- Absolute Acceleration [g]
40−20−0 20 40
1−
0
1
9th Floor Relative Displacement [cm]
- Absolute Acceleration [g]
40−20−0 20 40
1−
0
1
9th Floor Relative Displacement [cm]
- Absolute Acceleration [g]
a) 1978 Record - 0 to 15 sec. b) 1978 Record - 15 to 20 sec.
c) 2011 Record – 30 to 45 sec. d) 2011 Record – 45 to 82 sec.
Figure 21: Initial 9
th
-floor acceleration-displacement response
a) 2011 Record - 82 to 83 sec. b) 2011 Record - 83 to 84 sec.
c) 2011 Record - 84 to 85 sec. d) 2011 Record - 85 to 100 sec.
Figure 22: Final 9
th
-floor acceleration-displacement response
40−20−0 20 40
1−
0
1
9th Floor Relative Displacement [cm]
- Absolute Acceleration [g]
40−20−0 20 40
1−
0
1
9th Floor Relative Displacement [cm]
- Absolute Acceleration [g]
40−20−0 20 40
1−
0
1
9th Floor Relative Displacement [cm]
- Absolute Acceleration [g]
40−20−0 20 40
1−
0
1
9th Floor Relative Displacement [cm]
- Absolute Acceleration [g]
A
A’ B D
C
APPENDIX A
Table A1: Size and reinforcement of beams, columns, and walls (Shiga et al., 1981)
Table A1.1: Wall reinforcement and thickness (all dimensions in mm)
Story
Wall type / Thickness Reinforcement
Wa Wb Wc Wd We Wf Wg Thickness Reinforcement
9 150
150
200
150
200
150
9 @ 200, one layer
8 150
150
200
150
200
200
9 @ 200,two layers
7 150
200
200
150
200
250
9 @ 200,two layers
6 150
200
200
150
200
300
13 @ 200, two layers
5 200
300
200
200
200
400
13 @ 200, two layers
4 200
300
200
200
200
500
13 @ 200, two layers
3 250
300
250
250
200
2 300
400
300
300
200
150
1 400
500
400
400
200
250
150
Table A1.2: Column reinforcement (all dimensions in mm)
Story Axes B, C, D
between Axes 2
and 7
Axes A and E
between Axes 1
and 9, Axes B and
D with Axis 1
Axes B and D
with Axes 9
and 9
9 BxD: 800x850
a
8Ls-65x65x6
b
12-D19
c
8
7
6 BxD: 800x850
8Ls-75x75x6 12-
D19
5
4
3 BxD: 800x850
8Ls-75x75x9 12-
D22
2 BxD: 800x850
8Ls-75x75x9 12-
D2
2
BxD: 800x800
8Ls-75x75x9 12-
D22
BxD: 850x850
8Ls-75x75x9
12
-
D22
1 BxD: 850x850
8Ls-75x75x12
12
-
D25
BxD: 800x800
8Ls-75x75x9 12-
D22
BxD: 850x850
8Ls-75x75x12
12
-
D25
a
First line, BxD: North-South dimension x East-West dimension
b
Second line, Steel angles
c
Third line, reinforcing bars
Table A1.3: Beam reinforcement
Table A1.3.1: Beam reinforcement part 1 (all dimensions in mm)
Story Width x Depth,
Steel angles
Axis 2 through 7, Between Axes B and D
Axes B and D Mid-span
Axis C
R
400x800
4 Ls-65x65x6
Top 2-D22 2-D22 2-D22
Bottom 2-D22 2-D22 2-D22
9 Top 4-D22 2-D22 4-D22
Bottom 2-D22 2-D22 2-D22
7, 8 Top 2-D25 and 2-D22 2-D25 2-D25
Bottom 2-D22 2-D22 2-D22
6 Top 4-D25 2-D25 2-D25 and 2-D22
Bottom 2-D25 2-D25 2-D25
5
400x800
4 Ls-75x75x6 Top 4-D25 2-D25 4-D25
Bottom 2-D25 and 2-D16 2-D25 2-D25
4
400x900
4 Ls-75x75x6 Top 4-D25 2-D25 2-D25 and 2-D22
Bottom 2-D25 and 2-D16 2-D25 2-D25
2, 3 400x900 4
Ls-75x75x12
Top 4-D25 2-D25 2-D25
Bottom 2-D25 2-D25 2-D25
Stirrup
s
13 @300
Foun
datio
n 450x1500
Top 5-D22 5-D22 5-D22
Bottom 5-D22 5-D22 5-D22
Stirrup
s
13 @300
Table A1.3.2: Beam reinforcement part 2 (all dimensions in mm)
Story
Axis B and D, between Axes 1 through 9, Axis C between Axes 2
through 7
Width x
Depth, Steel
angles At the ends Mid-span
R
400x800
4 Ls-65x65x6
Top 4-D22 2-D22
Bottom 2-D22 2-D22
6, 7, 8, 9 Top 4-D25 2-D25
Bottom 2-D25 2-D25
5
400x800
4 Ls-75x75x6 Top 4-D25 2-D25
Bottom 2-D25 2-D25
4
400x900
4 Ls-75x75x6 Top 4-D25 2-D25
Bottom 2-D25 2-D25
2, 3
400x1100
4 Ls-75x75x9 Top 4-D25 2-D25
Bottom 2-D25 2-D25
Foundation
450x1500
Top 6-D22 5-D22
Bottom 6-D22 5-D22
Stirrups
13 @ 300
Table A1.3.3: Beam reinforcement part 3 (all dimensions in mm)
Story
Axes 1 through 9, between Axes A and B and Axes D and E
Width x Depth,
Steel angles
Axes A and D Mid-span Axes B and D
3 450x1000
4 Ls-75x75x9
Top 4-D25 2-D25 4-D25 and 2-
D22
Bottom 2-D25
2
-D25 and 2-D16 2-D25 and 2-
D16
Stirrups
13 @ 300
9 @ 300
13 @ 300
2 450x1100
4 Ls-75x75x9
Top 4-D25 2-D25 4-D25
Bottom 2-D25 2-D25 2-D25
Stirrups
13 @ 300
9 @ 300
13 @ 300
Foundation
450x1500
Top 6-D22 5-D22 6-D22
Bottom 6-D22 5-D22 6-D22
Stirrups
13 @300
Table A1.3.4: Beam reinforcement part 4 (all dimensions in mm)
Story
Width x
Depth,
Steel
angles
Axes A and E, between Axes 1 and 9
At the ends Mid-span
3
450x1000
4 Ls-
75x75x6
Top 4-D22 2-D22
Bottom 2-D22 2-D22
Stirrups
9 @ 300
2
450x1100
4 Ls-
75x75x9
Top 4-D25 2-D25
Bottom 2-D25 2-D25
Stirrups
9 @ 300
Foundation
450x1500
Top 4-D22 4-D22
Bottom 4-D22 4-D22
Stirrups
13 @ 300
Figure 23: Elevation of Axis 4 (all dimensions in mm)
Figure 24: Elevation of Axis D (all dimensions in mm)
