This paper describes partial results of a shake table test on a full scale slice of
a reinforced concrete residential building. The objective of this research program
is to verify the seismic response of reinforced concrete wall systems designed for
lateral forces that are significantly smaller than those currently specified in
building codes in United States (UBC 1997). Experience in earthquakes has
shown that structural walls are an excellent lateral force resisting system.
Structural engineers in California have often questioned that the current design
lateral forces for structural wall buildings renders this attractive lateral force
resisting system largely uneconomical and that significant savings could be made
if lateral forces could be reduced, particularly in the foundation and in the lower
levels of the walls where heavy congestion is observed. It is interesting that new
displacement-based design methodologies (Priestley and Kowalsky 1997,
Englekirk 2003) indicate that walls could be designed with significantly less
amounts of longitudinal reinforcement and still perform satisfactorily by properly
controlling interstory drift, and hence displacement sensitive non-structural
damage, and by having sufficient lateral deformation capacity to meet the
demands in a rare but strong earthquake. The test will be conducted in the newly
commissioned Large High-Performance Outdoor Shake Table funded under by
the National Science Foundation under the NEES program. The full-scale
structure will be subjected to historical input ground motions recorded in
Southern California that represent demands of earthquakes with 50% and 10%
probability of exceedance in 50 years.
The test structure is a 7 story wall structure built in full scale (Fig.1). The test consists of
a web wall of 12 ft long and two transverse walls: one is a flange 16 ft long and the other is a
precast segmental. Testing is being performed on the NEES-UCSD Large High-Performance
Outdoor Unidirectional Shake Table. The web wall provides lateral force resistance in the E-W
direction of loading, while the two transverse walls provide transverse and torsional resistance to
the test structure. The total height of the specimen is 63 ft and the total weight 550 kips. A 12 ft
by 26 ft-8 in. slab simple supported on gravity columns exists at every level. The width of the
web wall is 8 in. at the first and seventh floor and reduced to 6 in. in all the other floors. Both
the web and the flange wall are fixed at their base. The thickness of the flange wall is 8 in. at the
1
Graduate Student, Department of Structural Engineering, University of California, San Diego, La Jolla, CA 92093,
USA
2
Associate Professor,Department of Structural Engineering, University of California, San Diego, La Jolla, CA
92093, USA
3
Professor, Department of Structural Engineering, University of California, San Diego,La Jolla, CA 92093, USA
4
Englekirk Partners Consulting Structural Engineers, Inc., 2116 Arlington Avenue, Los Angeles, CA, USA
first floor and 6 in. in all the other floors. The flange wall is connected to the slab with pin pin
connection.
The segmental pier is connected with the slab through pin-pin horizontal steel truss. The
pier at the base is pinned in the E-W direction and fixed in the N-S direction. An elevation as
well as a foundation and floor plan view of the system are shown in Fig. 2. For the gravity
columns high strength steel pin-pin rods grouted in 4 in. pipes were used.
Tunnel steel forms were used for the construction of the walls and slabs. Concrete with
specified compressive strength of 4 ksi and A615 grade 60 steel were used. The construction
sequence included casting a level of the web and flange walls as well as the slab at a time. The
segmental pier was precast in three pieces and assembled afterwards using mortar bed joint and
post tensioning.
The test structure addressed four issues relevant to construction optimization. First the
reduction in the amount of reinforcement in the walls, second the reduction of the width of the
wall from 8 in to 6 in. on levels two to six and the use of a single layer of reinforcement without
any confinement, third the use of capacity design to control the non-linear response of the
structure, and fourth use of electrowelded confinement reinforcement in the first level of the
walls.
The reinforcement of the web wall in the first and seventh floor includes two layers of
vertical reinforcement (8#5) at each end and use of
confinement mesh (#3@4 in. Baugrid). Single layer
of reinforcement (13#4@10 in.) exists between. For
all the other levels, a single layer of vertical
reinforcement with 4#7 at each end and 11#4@10 in.
between exists. No confinement reinforcement is
provided in levels two to six. A single layer of
horizontal reinforcement (#4@8 in.) was also
provided in all the levels. The reinforcement ratio is
=0.66% in levels one and seven and =0.81% in
levels two to six. The reinforcement details of the
web and flange wall are shown on Fig. 3.
In more detail in each of the first two levels eight pairs of displacement transducers were
used at the ends of the wall for obtaining the local strains and curvature profile. The network of
displacement transducers was denser in the first two levels because of expected inelastic
response in this region. In addition to that two diagonal string potentiometers were placed at
every side of the wall for measuring the shear deformations. In levels three to seven two pairs of
strain pots were used at the ends of the web wall for measuring the average strains of these
regions which are expected to remain in the elastic range. On the post tensioned segmental pier
18 additional displacement transducers were placed in different parts. The instrumented parts of
the segmental pier were the connecting regions of the segmental pieces as well as in the
connection regions with the horizontal truss.
A first phase of experimental program investigates the response of the web cantilever
wall configuration to different levels of excitation. The program includes low amplitude white
noise tests, two low intensity earthquakes that are somewhat above the site response spectra for
the period of the building for 50% probability of exceedance event and a large intensity
earthquake whose spectral acceleration in the period range of interest is above the site response
spectra for 10% probability of exceedance in 50 years. For the white noises 3%g, 6%g and 9%g
root mean square (RMS) motions will be used. The white noise motions will be used for system
and damage identification before and after every earthquake test. The two low intensity records
are the vnuy transverse component record from 1971 San Fernando earthquake and the whox
longitudinal component from the Northridge 1994 earthquake. The large intensity record is the
Sylmar Olive View Med 360o component record from the 1994 Northridge earthquake. Different
levels of damage state will be investigated with the two low intensity earthquakes and the one
high intensity earthquake. For the high intensity earthquake significant damage is expected to be
concentrated especially at the base of the structure. The level of this damage will determine the
type and extent of repairability of the wall specimen and the next phases of the experimental
program. The ground acceleration time histories as well the acceleration and displacement
response spectra of the input motions can be seen on Fig. 5.
ag (g)
ag (g)
0 0 0
-1 -1 -1
0 10 20 30 0 10 20 30 0 10 20 30
t (sec) t (sec) t (sec)
1.5
0.4
1
0.2
0.5
0 0
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
T(sec) T (sec)
The first results regarding the system identification are based on a series of ambient
vibration and white noise tests of 2%g and 3%g RMS. The results presented are based on the
Fourier response spectra of the longitudinal and transverse accelerometers of a 3%g RMS 8 min.
long white noise test that was performed on October 12 2005. The layout and labels of these
accelerometers at every level are as shown in Fig. 4.
From the Fourier amplitude spectra of the transverse accelerometers at the east end of the
web wall and close to the flange wall (Ti-1) we observe a peak at f=3.45 Hz. This peak is not
obvious on the other hand on the west end transverse accelerometers. Thus we can conclude that
this mode has a torsional component with rotation in plan view around the west end of the
system. In addition from the transverse accelerometers on the west end of the web wall we
observe especially from the spectra of the lower levels that at f=1.45Hz there is an additional
peak. This peak is not so obvious on the upper levels because it interferes with the peak at
f=1.82 Hz. Also does not appear on the transverse accelerometers on the east end and denotes a
torsional mode with rotation in plan view around the middle of the flange wall (east end of the
system).
We note that the present results were from data taken during the calibration of the
instrumentation and before the refinement of structural details. During the initial white noise
tests for the checking and calibration of the instrumentation as well as for the identification of
the system characteristics, important slackness was observed in the horizontal truss connections.
This slackness was obvious during the test and caused impact between the truss, the wall and
the post tensioned pier. Details like these may are important for interpretate results like the mode
observed at f=1.45 Hz. We speculate that this torsional mode at f=1.45Hz is related with the
slackness in the bracing system and how that is related with the stiffness of the torsional mode.
Future modifications of the bracing system and future tests will prove the validity of that
implication and how similar structural modifications will affect the discussed here results. We
note that during the process of the present results the instrumentation was still under deployment
and calibration and this is the reason that some of the Fourier amplitude spectra are blank.
1000 1000 1000
H7-3 H7-1 H7-2
Ampl.
500 500 500
0 0 0
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
1000 1000 1000
H6-3 H6-1 H6-2
Ampl.
0 0 0
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
1000 1000 1000
H5-3 H5-1 H5-2
Ampl.
0 0 0
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
1000 1000 1000
H4-3 H4-1 H4-2
Ampl.
0 0 0
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
1000 1000 1000
H3-3 H3-1 H3-2
Ampl.
0 0 0
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
1000 1000 1000
H2-3 H2-1 H2-2
Ampl.
0 0 0
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
1000 1000 1000
H1-3 H1-1 H1-2
Ampl.
0 0 0
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
f (Hz) f (Hz) f (Hz)
200 200
T7-2 T7-1
Ampl.
100 100
0 0
0 2 4 6 8 10 0 2 4 6 8 10
200 200
T6-2 T6-1
Ampl.
100 100
0 0
0 2 4 6 8 10 0 2 4 6 8 10
200 200
T5-2 T5-1
Ampl.
100 100
0 0
0 2 4 6 8 10 0 2 4 6 8 10
200 200
T4-2 T4-1
Ampl.
100 100
0 0
0 2 4 6 8 10 0 2 4 6 8 10
200 200
T3-2 T3-1
Ampl.
100 100
0 0
0 2 4 6 8 10 0 2 4 6 8 10
200 200
T2-2 T2-1
Ampl.
100 100
0 0
0 2 4 6 8 10 0 2 4 6 8 10
200 200
T1-2 T1-1
Ampl.
100 100
0 0
0 2 4 6 8 10 0 2 4 6 8 10
f (Hz) f (Hz)
Figure 8. Plan view of the deflected shape of the three first mode shapes (SAP analysis)
Verification of Response
A two-dimensional lumped mass model was developed for the prediction and verification
of the response of the 7-story structure. The model includes a dynamic strut and tie model for the
modeling of the two bottom levels of the wall were inelastic response is expected to concentrate.
The model incorporates capabilities of capturing sliding shear response of the web wall. Levels
three to seven are modeled with beam elements and appropriate modeling of the hysteretic
response. Additional elements were used for the incorporation of characteristics of 3-D response
like the response of the slab. Beam models with appropriate hysteretic modeling are used for the
flange wall as well as for the post tensioned segmental pier. The existing version of the model
includes around 1000 beam and truss elements and about 60 properties describing their
hysteretic behavior. The program Ruaumoko (Carr 2003) is used for the dynamic inelastic
analysis of the system.
The first results with use of the prescribed model are presented below for a white noise
test of two-minute long 2%g RMS that was done on October 10 2005. The first 30 sec of the
floor acceleration measured and predicted response is presented for the fifth, sixth and seventh
floor of the building (Fig. 9). We note that because the middle accelerometer on the seventh
floor gave no data during the specific test the average response of the two horizontal end
accelerometers (H7-2, H7-3) is given as measured response of the seventh floor.
7th floor
0.15
predicted
0.1 average measured of H7-2,H7-3
0.05
a(g)
0
-0.05
-0.1
0 5 10 15 20 25 30
6th floor
0.1 predicted
measured H6-1
0.05
0
a(g)
-0.05
-0.1
0 5 10 15 20 25 30
5th floor
0.1 predicted
measured H5-1
0.05
0
a(g)
-0.05
-0.1
0 5 10 15 20 25 30
t (sec)
Figure 9. Predicted vs measured floor accelerations. First 30 sec of 2%g RMS white noise test.
Conclusions
References
Carr, A., 2003. Ruaumoko User Manual, University of Canterbury, New Zealand.
Englekirk, R. E., 2003. Seismic Design of Reinforced and Precast Concrete Buildings , John Wiley &
Sons, Inc., Hoboken, New Jersey.
Priestley, M.J.N. and Kowalsky, M.J, 2000. " Direct Displacement-Based Seismic Design of Concrete
Buildings. " Bulletin of the New Zealand National Society for Earthquake Engineering, Vol. 33,
No. 4 pp 421-444.