Experimental Assessment of the Seismic Response of a Base-Isolated Building Through a Hybrid Simulation Technique

Effect of viscous damping devices on the response of seismically
isolated structures
E. D. Wolff1,5
, C. Ipek2,5
, M. C. Constantinou3,5,
*,†
and M. Tapan4,5
1
Fyfe Co, San Diego, CA 92126, USA
2
Istanbul Technical University, Faculty of Civil Engineering, Department of Civil Engineering, Maslak, Istanbul, Turkey
3
Department of Civil, Structural and Environmental Engineering, Univ. at Buffalo, SUNY, Buffalo, NY 14260, USA
4
Risk Management and Earthquake Research and Application Center, Yuzuncu Yil University, Van, 65080, Turkey
5
Univ. at Buffalo, SUNY, Buffalo, NY, USA
SUMMARY
Viscous and other damping devices are often used as elements of seismic isolation systems. Despite the
widespread application of nonlinear viscous systems particularly in Japan (with fewer applications in the
USA and Taiwan), the application of viscous damping devices in isolation systems in the USA progressed
intentionally toward the use of supplementary linear viscous devices due to the advantages offered by these
devices. This paper presents experimental results on the behavior of seismically isolated structures with low
damping elastomeric (LDE) and single friction pendulum (SFP) bearings with and without linear and
nonlinear viscous dampers. The isolation systems are tested within a six-story structure configured as
moment frame and then again as braced frame. Emphasis is placed both on the acquisition of data related
to the structural system (drifts, story shear forces, and isolator displacements) and on non-structural systems
(floor accelerations, floor spectral accelerations, and floor velocities). Moreover, the accuracy of analytical
prediction of response is investigated based on the results of a total of 227 experiments, using 14 historic
ground motions of far-fault and near-fault characteristics, on flexible moment frame and stiff braced frame
structures isolated with LDE or SFP bearings and linear or nonlinear viscous dampers. It is concluded that
when damping is needed to reduce displacement demands in the isolation system, linear viscous damping
results in the least detrimental effect on the isolated structure. Moreover, the study concludes that the
analytical prediction of peak floor accelerations and floor response spectra may contain errors that need to
be considered when designing secondary systems. Copyright © 2014 John Wiley & Sons, Ltd.
Received 10 January 2014; Revised 29 June 2014; Accepted 30 June 2014
KEY WORDS: seismic isolation; viscous damping; secondary systems; shake table testing; response history
analysis
1. INTRODUCTION
Seismic isolation is considered an effective earthquake mitigation strategy. There are over 3000
seismically isolated buildings in Japan, over 200 buildings in the USA, and several more in as many
as 30 countries, including many in China, Taiwan, New Zealand, Italy, Russia, and Turkey. The
isolation systems of preference in Japan are elastomeric bearings (lead-rubber and high damping
rubber) with few applications of double friction pendulum bearings and of systems combining
elastomeric bearings and either yielding steel devices or fluid viscous dampers (termed ‘oil dampers’
in Japan). In the USA, the preferred isolators are lead-rubber and friction pendulum bearings – the
latter typically nowadays are of the Triple type. A small number of important isolated structures in
*Correspondence to: Mücip Tapan, Risk Management and Earthquake Research and Application Center, Yuzuncu Yil
University, Zeve Kampusu, Van, 65080, Turkey.
†
E-mail: mucip.tapan@gmail.com
Copyright © 2014 John Wiley & Sons, Ltd.
EARTHQUAKE ENGINEERING & STRUCTURAL DYNAMICS
Earthquake Engng Struct. Dyn. 2015; 44:185–198
Published online 24 July 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/eqe.2464

close proximity to seismic faults in California employ hybrid isolation systems combining elastomeric
bearings or friction pendulum bearings with viscous damping devices.
The viscous damping devices manufactured and used in the USA utilize specially shaped orifices to
achieve linear (force proportional to velocity) or nonlinear (force proportional to a power of velocity –
the power typically being 0.4 to nearly unity) viscous behavior [1, 2]. The devices used in Japan
(oil dampers) are typically fitted with a pressure relief valve to have a pre-determined upper bound
of force so that they have a highly nonlinear behavior. Viscous damping devices manufactured in
the USA and implemented in Japan typically have nonlinear behavior with a power of 0.38 as that
represents an approved behavior for Japanese applications [2].
The use of viscous damping devices in isolation systems in the USA intends to limit the isolation
system displacement demands to acceptable levels. Examples of such applications in chronological
order starting in 1994 and ending in 2014 (under construction) are as follows:
(1) Arrowhead Medical Center, San Bernardino, California, high damping rubber bearings, nonlinear
dampers of power of 0.5, capacity 600 mm.
(2) Los Angeles City Hall, high damping rubber bearings, nonlinear viscous dampers of power of
0.5, capacity 600 mm.
(3) Hayward City Hall, single friction pendulum bearings, nonlinear viscous dampers of power of
0.5, capacity 600 mm.
(4) Hearst Memorial Mining Building, Berkeley, California, high damping rubber bearings, nonlinear
viscous dampers of power of 0.5, capacity 815 mm.
(5) New de Young Museum, San Francisco, high damping rubber bearings, flat sliding bearings,
nonlinear viscous dampers of power of 0.5, capacity 760 mm.
(6) New Caltrans District 8 Center, Fontana, California, elastomeric bearings, nonlinear viscous
dampers of power of 0.5, capacity 600 mm.
(7) Mills Peninsula Hospital, Burlingame, California, triple friction pendulum bearings, linear viscous
dampers, capacity 760 mm.
(8) Washington Hospital, Fremont, California, triple friction pendulum bearings, linear viscous
dampers, capacity 915 mm.
The interest in using nonlinear devices is based on two considerations:
(1) Nonlinear viscous devices are capable of more energy dissipation per cycle for the same
displacement amplitude than linear devices, therefore, presumably more effective.
(2) Nonlinear viscous devices provide a safeguard by limiting the transmission of damping force at
large velocities beyond the design values.
Despite these considerations and the widespread application of nonlinear viscous systems,
particularly in Japan (with fewer applications in the US and Taiwan), the application of viscous
damping devices in isolation systems in the USA progressed toward the use of linear viscous
devices. This is intentional as linear viscous damping devices offer advantages that represent the
focus of this paper.
The utility of damping devices as part of seismic isolation systems was questioned by Kelly [3] who
on the basis of approximate analysis of linear elastic and linear viscous isolation systems concluded
that, while additional damping reduces isolation system displacement demands, it does so at the
expense of increased floor accelerations and interstory drifts. This was disputed by Hall [4] who on
the basis of response history analysis demonstrated that supplemental damping offers advantages.
Politopoulos [5] provided further analytical evidence, again for linear elastic and linear viscous
systems, to corroborate the conclusions of Hall [4]. Furthermore, Politopoulos [5] demonstrated that
added linear viscous damping, when not excessive, has beneficial effects on the floor response
spectra. Hall and Ryan [6] conducted response history analyses of isolated structures with due
considerations for the nonlinear behavior of high damping rubber bearings and with linear viscous
dampers to arrive at similar conclusions – that is, added viscous damping reduces the isolation
system displacement demands and may also reduce drifts.
Studies by Providakis [7, 8] followed the paradigm of Hall and Ryan [6] and studied bilinear
hysteretic isolation systems and single friction pendulum systems with added linear and nonlinear
186 E. D. WOLFF ET AL.
Copyright © 2014 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2015; 44:185–198
DOI: 10.1002/eqe

viscous dampers. The studies concluded that while the addition of the dampers improved the
performance of the isolated structure in near-fault seismic excitations, the performance deteriorated
in far-field seismic excitations. Presumably, this effect is the result of ‘too much damping’ in the
weaker far-field motions, particularly when the damping is nonlinear. This is consistent with the
observations of Hall and Ryan [6] and Politopoulos [7] who concluded that too much of linear
viscous damping may be detrimental.
This paper presents experimental results on the behavior of seismically isolated structures with
supplemental viscous dampers. While the experimental study had a wider focus [9], the paper
concentrates only on isolation systems consisting of low damping elastomeric and single friction
pendulum bearings without and with linear and nonlinear viscous dampers. The isolation systems
are tested within a six-story structure configured as a moment frame and then again as a braced
frame. Only horizontal components of seismic excitation are used. Emphasis is placed both on the
acquisition of data related to the structural system (drifts, story shear forces, and isolator
displacements) and on non-structural systems (floor accelerations, floor spectral accelerations, and
floor velocities). Moreover, the accuracy of analytical prediction of response is investigated.
The experimental program served two purposes: (i) to acquire experimental data for use in
validating analytical models of isolated structures, and (ii) in obtaining data that can be used to
derive conclusions on the utility of viscous damping devices in isolated structures. It should be
noted that the latter implies that the test data can be extrapolated to the prototype scale, which can
never be exact. The reason is that it is impossible to satisfy all principles of similarity. An issue is
that in reduced scale testing, thermodynamic similarity requires that the speed of motion is increased
for proper consideration of the isolator heating effects, whereas dynamic similitude requires that the
speed is decreased. Because dynamic similitude needed to be employed in the shake table testing,
heating effects were not properly accounted for [1, 10].
It is concluded that when damping is needed to reduce displacement demands in the isolation
system, linear viscous damping results in the least detrimental effect on the isolated structure.
Moreover, the study concludes that the analytical prediction of peak floor accelerations and floor
response spectra may contain errors that need to be considered when assessing performance or
designing secondary systems.
2. DESCRIPTION OF TEST STRUCTURE AND EXPERIMENTAL PROGRAM
The six-story model used in the earthquake simulator testing is identical to that used in previous testing
of seismic isolation systems at the University at Buffalo [11]. It represents a section in the weak
direction of a steel moment-resisting frame.
The structure is shown in Figure 1. All column and beam sections are S3 × 5.7, and all out-of-plane
braces are L 1½ × 1½ × ¼ (37 × 37 × 6.25). X-bracing consisting of L 1½ × 1½ × ¼ sections could be
added in the middle bay of the two frames on the East and West sides of the model to convert the
structure to a braced frame. The structure is attached to a rigid base consists of two AISC W14 × 90
sections, 5.2 m long with six transversely connected beams. The model has six stories of 0.915 m
height each, giving a total height of 5.49 m above the base. The model is three bays by one bay in
plan, each bay being 1.22 m wide, for total plan dimensions of 1.22 m × 3.66 m. Concrete blocks
were used to add mass to satisfy similitude requirements, bringing the total weight, including the
base, to 233 kN. The structure was constructed to have a length scale of 4.
Tables I and II present the modal characteristics of the moment frame and braced frame versions of
the tested model when fixed at its base. These characteristics were obtained by processing of records
acquired during shake table testing using banded (0–40 Hz) white noise excitation of 0.05 g peak
acceleration. The properties are based on a model with one horizontal DOF per floor of the model.
Details of the identification are presented in Wolff and Constantinou [9].
The seismic isolation system consisted of either four single friction pendulum bearings or
four elastomeric bearings and without or with added linear or nonlinear viscous dampers for a
total of six different isolation systems. The properties of the isolation system elements were
as follows:
RESPONSE OF VISCOUSLY DAMPED SEISMICALLY ISOLATED STRUCTURES 187
Copyright © 2014 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2015; 44:185–198
DOI: 10.1002/eqe

(1) Four single friction pendulum isolators are each having effective radius of curvature Reff = 750 mm,
displacement capacity of just over 200 mm, and coefficient of friction μ having parameters
fmax = 0.077, fmin = 0.020, and a = 20 s/m (a parameter to describe the variation of friction
with velocity of sliding), per equation 1 after Constantinou et al. [1] where V is the velocity
of sliding:
Figure 1. Tested six-story seismically isolated model structure.
Table I. Characteristics of fixed-base moment frame structure.
Mode
Frequency
(Hz)
Damping
ratio
Mode shape
Floor 1 Floor 2 Floor 3 Floor 4 Floor 5 Floor 6
1 2.34 0.048 0.22 0.43 0.60 0.77 0.94 1.00
2 7.90 0.019 0.52 1.05 0.98 0.41 0.40 1.00
3 13.65 0.011 0.98 1.02 0.27 1.27 0.59 1.00
4 19.79 0.003 2.21 0.48 1.99 0.28 1.67 1.00
5 25.45 0.014 2.51 1.66 0.14 2.40 2.86 1.00
6 29.54 0.018 2.16 4.94 4.96 4.22 2.50 1.00
Table II. Characteristics of braced fixed-base structure.
Mode
Frequency
(Hz)
Damping
ratio
Mode shape
Floor 1 Floor 2 Floor 3 Floor 4 Floor 5 Floor 6
1 4.00 0.040 0.17 0.34 0.49 0.67 0.87 1.00
2 17.90 0.017 0.67 1.16 0.96 0.43 0.38 1.00
3 30.70 0.009 1.00 1.06 0.71 1.38 0.63 1.00
Copyright © 2014 John Wiley Sons, Ltd. Earthquake Engng Struct. Dyn. 2015; 44:185–198
DOI: 10.1002/eqe

μ ¼ f max f max f min
ð Þ exp α V
j j
ð Þ (1)
Values of the friction coefﬁcient were obtained in testing of the friction pendulum isolation system
on the shake table under harmonic motion (four isolators under total gravity load of 233 kN) and
used to calibrate equation 1 (see [9] for details). It may be noted that the friction coefﬁcient values
are typical for a seismically isolated structure. The value of parameter a = 20 s/m is on the low side
of typical sliding interfaces used in seismic isolation where values range from about 20 to 100 s/m
[1]. However, because of the velocity of testing being reduced by factor 2, similarity is violated
particularly at very low velocities where the friction forces are strongly dependent on velocity. This
may have some effects on the measured secondary system response.
(2) Four low damping elastomeric bearings each having a bonded diameter of 177.5 mm, a central
hole of 25 mm, and 18 rubber layers of 3.2 mm thickness each (total rubber thickness of 57 mm).
Force-displacement loops of these bearings may be found in [9]. Each of the bearings could be
either modeled as a linear elastic and linear viscous element with effective stiffness equal to
0.32 kN/mm and effective damping constant equal to 3.92 × 103
kN-s/mm (equivalent to effec-
tive damping ratio of 0.045 when load on the bearing is 58.3 kN or one quarter of the model
weight) or as a hysteretic element with characteristic strength of 1.4 kN, yield displacement of
5.7 mm (or equivalently initial stiffness equal to 0.55 kN/mm and yield force of 3.1 kN) and
post-elastic stiffness of 0.30 kN/mm (this stiffness may be calculated by using the bonded area
of 24254 mm2
, rubber thickness of 57 mm, and rubber shear modulus of 0.7 MPa).
(3) Two linear viscous dampers each having damping constant C = 0.0664 kN-s/mm and installed at an
angle of 40.4° with respect to the direction of motion (see [9] for details). The angle of the dampers
varied between 38° and 43° during motion because of the increase or reduction of their length.
(4) Two nonlinear viscous dampers installed in exactly the same manner as the linear dampers. Each
damper had its peak force F related to the peak damper relative velocity VD by
F ¼ CNVα
D (2)
where CN = 2.226 kN (s/mm)α
and α = 0.397. Note that the linear and the nonlinear damping devices
have been designed so that they produce about the same force at the design velocity of 350 mm/s.
The isolated and non-isolated (isolation system locked) model structure was tested with several historic
earthquake motions. Table III lists the motions utilized in testing and their peak ground motion
characteristics in prototype scale. Some of the earthquakes listed, such El Centro, Miyagiken, and Taft,
are far-fault motion while most of the other motions have near-fault characteristics, such as the motions
recorded in the 1994 Northridge earthquake, the 1999 Chi-Chi, the 1995 Kobe, and the 1971 Pacoima
earthquakes. The 1985 Mexico City and 1968 Hachinohe earthquakes were chosen for their content in
long-period components. The 1999 Kocaeli earthquake was added because of its catastrophic nature
and because data became available during testing. Each record was compressed in time by a factor of
two to conform to similitude requirements. That is, all tests were conducted in length scale of 4 and
time scale of 2. Additional scale factors may be found in [9]. Moreover, each record was run with the
amplitude of acceleration multiplied by factors ranging from 0.5 to as much as 5 for the isolated cases
and factors ranging from 0.15 to 0.5 for the non-isolated cases. This resulted in the strongest possible
earthquake motions without damaging the model.
3. EXPERIMENTAL RESULTS ON RESPONSE OF ISOLATED STRUCTURES WITH
VISCOUS DAMPING DEVICES
A detailed presentation of experimental results, including histories of recorded response parameters for
each test, can be found in [9]. Herein, summary results are presented on the peak recorded response of
the isolated structure only in terms of the isolation system displacements, story drift ratio (normalized
by story height of 915 mm), isolation or base shear force, and story shear force, both normalized by the
model weight of 233 kN. Results are presented in Figures 2–5 that present the recorded peak values of
these quantities versus the peak shake table velocity in each of the conducted tests. This velocity
DOI: 10.1002/eqe

describes well the intensity of the seismic input given that the isolated structural system is highly
ﬂexible (with period that falls in the velocity-controlled domain of the spectrum). The term LD in
these ﬁgures denotes the system with linear dampers and the term NLD denotes the system with
nonlinear dampers. Note that the results presented are in the scale of the testing, so that to
extrapolate to the prototype, displacements need to be multiplied by factor 4 and velocities by factor 2.
Table III. List of earthquake motions and characteristics in prototype scale.
Notation Excitation
Peak ground motion
Displacement
(mm)
Velocity
(mm/s)
Acceleration
(g)
El Centro S00E Imperial Valley, May 18, 1940, Component S00E 109 335 0.34
Taft N21E Kern County, July 21, 1952, Component N21E 67 157 0.16
NR Newhall 90° Northridge-Newhall, LA County Fire Station,
January 17,1994, Component 90°
176 748 0.58
NR Newhall 360° Northridge-Newhall, LA County Fire Station,
January 17, 1994, Component 90°
305 947 0.59
NR Sylmar 90° Northridge-Sylmar, Parking lot, January 17, 1994,
Component 90°
152 769 0.60
Kobe N-S Kobe Station, Japan, January 17, 1995, Component N-S 207 914 0.83
Mexico N90W Mexico City, September 19, 1985, SCT Building
Component N90W
212 605 0.17
Pacoima S74W San Fernando, February 9, 1971, Pacoima Dam
Component S74W
108 568 1.08
Pacoima S16E San Fernando, February 9, 1971, Pacoima Dam,
Component S16E
365 1132 1.17
Chi-Chi (Taiwan) Taiwan, September 21, 1999, Station TCU 129,
Component E-W
502 600 0.98
Miyagiken Oki Miyagi, Japan, June 12, 1978, Ofunato-Bochi
Component E-W
51 141 0.16
Hachinohe N-S Tokachi, Japan, May 16, 1968, Hachinohe,
Component N-S
119 375 0.23
YPT 060 Kocaeli, Turkey, August 17, 1999, Yarımca,
Component E-W
570 657 0.27
YPT 330 Kocaeli, Turkey, August 17, 1999, Yarımca,
Component N-S
510 621 0.35
Figure 2. Comparison of peak response of isolated moment frame structure with friction pendulum bearings
and viscous dampers.
DOI: 10.1002/eqe

The results demonstrate that the addition of the dampers generally results in reduction of the isolator
displacements. This reduction is more evident in the tests with the elastomeric bearings due to the low
damping capacity of the bearings themselves (about 5% of critical). Moreover, in the case of the
elastomeric system, the addition of dampers also results in reduction of drifts and shear forces. The
reduction of these quantities is more distinct in the case of the linear dampers.
In the case of the friction pendulum system, which is highly damped by itself (effective damping of
about 20–30%), the addition of dampers results in a general increase in drifts and shear forces. The
increase is clearly more distinct in the case of the nonlinear dampers.
Figure 3. Comparison of peak response of isolated braced frame structure with friction pendulum bearings
and viscous dampers.
Figure 4. Comparison of peak response of isolated moment frame structure with elastomeric bearings and
viscous dampers.
DOI: 10.1002/eqe

To further demonstrate the effects of added dampers, two cases are investigated in more depth. The
moment frame is selected as it is sufficiently flexible to be excited in higher mode response. Two
earthquakes records are selected: the far-fault El Centro component S00E, scaled by factor of 2, and
the near-fault Sylmar 90° component, scaled by factor 1. Results are presented in Tables IV and V.
For each system, the following peak response quantities are presented: isolator displacement; base
shear force and story shear force, both normalized by weight (233 kN); story drift normalized by
story height of 915 mm; base (level just above isolators), third floor and fifth floor acceleration;
base, third floor and fifth floor total velocity; base, third floor and fifth floor spectral acceleration
(peak value of 5% spectral acceleration calculated from the recorded history of acceleration over a
frequency range of 0–20 Hz); and roof acceleration.
For each of the cases shown in Tables IV and V, the effective period and effective damping of the isolated
structure were calculated based on the recorded amplitude of isolation displacement and are reported in the
tables. The results in these tables provide a clearer picture of the general results shown in Figures 2–5. They
clearly demonstrate the beneficial effects of adding linear viscous dampers to low damped elastomeric
isolation systems in terms of the reduction of isolator displacements, drifts, and shear forces. Moreover,
we can observe the reduction in floor accelerations, peak floor spectral accelerations, and floor total
velocities. We can also observe some benefits in the response of the elastomeric bearing isolated structure
when nonlinear viscous dampers are added, but the benefits do not apply to all response quantities.
The results also demonstrate increases in drift and shear forces when dampers are added to the
highly damped friction pendulum system, but with the observation that linear dampers have a lesser
Figure 5. Comparison of peak response of isolated braced frame structure with elastomeric bearings and
viscous dampers.
Table IV. Comparison of peak response quantities of isolated moment frame structure in El Centro S00E 200%.
Isolation system
Effective
period (s)
Effective
damping
Peak isolator
Displacement
(mm)
Peak base
shear/ weight
Peak story
shear/ weight
Peak drift/
height (%)
Low damping elastomeric 0.87 0.05 60 0.32 0.29 0.67
Low damping elastomeric
linear dampers
0.87 0.29 36 0.26 0.24 0.49
Low damping elastomeric
nonlinear dampers
0.87 0.27 25 0.26 0.25 0.56
FPS 1.1 0.31 40 0.14 0.2 0.42
FPS-linear dampers 0.99 0.6 29 0.12 0.25 0.53
FPS-nonlinear dampers 0.92 0.67 23 0.26 0.31 0.66
FPS, friction pendulum system.
DOI: 10.1002/eqe

unfavorable impact than nonlinear viscous dampers. Additional results for the braced frame may be
found in [9] – they result in similar observations as the results for the moment frame.
The information in Tables IV and V on the floor spectra of the tested model is limited to a single
value – the peak value over the frequency range of 0–20 Hz Frequency is in the scale of the
experiments so that in prototype scale, the range is 0–10 Hz. The limit on frequency is necessitated
for accuracy as electronic noise and filtering alter the fidelity of high frequency recordings. The
entire floor spectra provide much more information. Floor response spectra for all conducted tests
are presented in [9]. Samples of such spectra are presented together with analytical predictions of
response in the next section of this paper.
Note that the information in Tables IV and V is limited to the isolated moment frame, which
featured a flexible superstructure with a ratio of effective isolation system period to superstructure
fundamental period in the range of about 2.0–2.6 (depending on the amplitude of displacement of
the isolation system and for strong motions). This ratio is low but not uncommon in modern
seismically isolated structures. For example, the recently constructed (2013) ten-story steel moment
frame San Francisco General Hospital has a ratio of effective isolation system period in the
Maximum Earthquake to superstructure fundamental period equal to 2.0. For the tested stiffer braced
frame, the ratio of effective isolation system period to superstructure fundamental period is in the
range of about 3.4–4.4, which is still low but more representative of typical seismically isolated
structures; for example, the six-story concrete shear wall Erzurum Hospital in Turkey has a ratio of
effective isolation system period in the Maximum Earthquake to superstructure fundamental period
equal to 5.5.
Results on floor spectra presented in [9] for the moment and the braced frame show that, in general,
the addition of viscous damping reduces the amplitude of the peak of the first mode resonant but has a
mixed effect on the second mode resonant, in some cases causing amplification and in others reduction
depending on the floor location and the earthquake excitation. This likely is the result of the fact that
both tested structures had a low ratio of effective isolation system period to superstructure fundamental
period. It would have been expected that stiff isolated structures (with high value of the period ratio)
experience amplification of the higher mode response when damping is added [5]. Yet evidence of
this behavior may be seen in the floor spectra of the braced frame in the case of the El Centro
motion (Figures 4–8 in [9]) where added linear damping in the elastomeric isolation system causes
marked amplification in the second mode of the base and the 5th floor spectra.
4. ACCURACY OF ANALYTICAL PREDICTION OF RESPONSE
The acquisition of test data involves errors due to instrument calibration errors, electronic noise,
improper filtering of data, improper placement of sensors, effects of environmental conditions, and
so on. Care has been taken so that the data presented in this paper and in [9] have been checked for
Table IV. (Continued)
Peak base response Peak 3rd floor response Peak 5th floor response
Peak
roof
acceleration
(g)
Acceleration
(g)
Spectral
acceleration
(g)
Velocity
(mm/s)
Acceleration
(g)
Spectral
acceleration
(g)
Velocity
(mm/s)
Acceleration
(g)
Spectral
acceleration
(g)
Velocity
(mm/s)
0.54 1.88 409 0.45 1.85 506 0.41 2.12 858 0.54
0.45 1.64 312 0.35 1.65 406 0.4 1.33 480 0.52
0.66 2.36 296 0.45 1.64 368 0.63 1.57 453 0.78
0.55 1.77 237 0.35 1.9 227 0.51 1.52 302 0.8
0.56 1.88 241 0.54 1.95 286 0.64 1.81 313 0.81
0.56 2.03 240 0.64 2.21 337 0.79 2.07 0.95 0.95
DOI: 10.1002/eqe

accuracy using redundant measurements and reduction of data by independent means. Specifically,
absolute displacement records were numerically differentiated to obtain acceleration records, which
were then compared to direct records of acceleration. In doing so, accurate histories of absolute
velocity were also obtained. Histories of story shear forces were obtained by processing of records
of acceleration (removal of any torsional effects, multiplication by known mass, and adding up) and
then compared with direct measurements of force by load cells. Samples of such comparisons may
be found in [9].
Analytical models of the tested structure have been developed in programs SAP2000 [12] and 3D-
BASIS-ME [13]. Details of the developed analytical models, program input files, and comparisons of
experimental and analytical results for each conducted test are available in [9]. Herein, only selected
comparisons of experimental and analytical results obtained by the computer program SAP2000 are
presented. The analytical model for the elastomeric isolators was based on the hysteretic
representation described earlier (characteristic strength of 1.4 kN, yield displacement of 5.7 mm, and
post-elastic stiffness of 0.3 kN/mm). The model of the dampers was based on C = 0.0664 kN-s/mm
for the linear dampers and CN = 2.226 kN (s/mm)α
and α = 0.397 (per equation 2) for the nonlinear
dampers. The model for the friction pendulum bearings was based on an effective radius of 750 mm
and friction per equation 1 with fmax = 0.077, fmin = 0.020, and a = 20 s/m. Note that the parameters
used in modeling the isolation system components are based on test data.
The model of the superstructure was sufficiently detailed to be capable of predicting the identified
modal properties of the structure when fixed at the base with acceptable accuracy. Structural damping
was modeled so that the identified modal damping ratio values were approximated: damping ratio of
0.04 in the first mode and 0.02 in all higher modes of the superstructure. In program 3D-BASIS-ME
[13], the superstructure damping matrix is formed and used in the analysis. However, in program
SAP2000, a global damping matrix is formed that includes the isolation system DOFs. Accordingly,
the specification of structural damping requires special attention to avoid introducing parasitic viscous
damping in the isolation system that results in the reduction of isolation system displacement
demands. This issue has been addressed in Sarlis and Constantinou [14] who provided
recommendations on how to avoid this problem in program SAP2000. In the analysis presented in
this paper, this problem has been alleviated by specifying a reduced value for the first mode damping
ratio and by anchoring this damping ratio on the post-elastic stiffness of the isolators (instead of the
effective or the elastic stiffness). The model in program SAP2000 featured a 3D representation of the
tested structure in which all modes (or Ritz vectors) were accounted for in the nonlinear analysis.
However, sample analyses have shown that far fewer modes could provide sufficiently accurate
global response quantities such as isolator displacements, story drifts, and floor shear forces.
Comparisons of experimental and analytical results for the elastomeric system and the friction
pendulum system, both with linear viscous dampers, in tests in the flexible moment frame
configuration in the El Centro S00E motion scaled by factor of 2 are presented in Figures 6, 7.
Histories of isolation system displacement, isolation system or base shear force-displacement loops,
and floor response spectra (5% damped over frequency range of 0–20 Hz) are compared. Additional
response quantity comparisons may be found in [9].
Table V. Comparison of peak response quantities of isolated moment frame structure in Sylmar 90° 100%.
Isolation system
Effective
period (s)
Effective
damping
Peak isolator
displacement
(mm)
Peak
base
shear/
weight
Peak
story
shear/
weight
Peak
drift/
height
(%)
Low damping elastomeric 0.87 0.05 65 0.36 0.33 1.03
Low damping elastomeric linear dampers 0.87 0.29 42 0.29 0.27 0.55
Low damping elastomeric nonlinear dampers 0.87 0.24 32 0.31 0.3 0.61
FPS 1.24 0.22 61 0.17 0.24 0.48
FPS-linear dampers 1.11 0.49 41 0.26 0.26 0.55
FPS-nonlinear dampers 1.07 0.49 37 0.26 0.29 0.68
DOI: 10.1002/eqe

In general, the results presented in this paper and many more presented in [9] show that the analytical
models predict well global response quantities such as the isolator displacements, story and base shear
forces, axial isolator forces, and story drifts. Moreover, comparisons of analytical and experimental
histories of floor total velocities (may be found in [9]) were in very good agreement. The maximum
differences observed between peak experimental and analytical response quantities were about 15% or
Figure 6. Comparison of experimental and analytical results for the elastomeric bearing/linear viscous
damper isolated moment frame in the El Centro S00E 200% excitation.
Table V. (Continued)
Peak base response Peak 3rd floor response Peak 5th floor response
Peak
roof
acceleration
(g)
Acceleration
(g)
Spectral
acceleration
(g)
Velocity
(mm/s)
Acceleration
(g)
Spectral
acceleration
(g)
Velocity
(mm/s)
Acceleration
(g)
Spectral
acceleration
(g)
Velocity
(mm/s)
0.48 2.07 473 0.39 1.77 581 0.52 2.02 656 0.54
0.45 1.26 354 0.33 1.06 417 0.46 1.2 472 0.54
0.61 1.49 315 0.46 1.36 373 0.6 1.4 436 0.76
0.65 1.95 296 0.39 1.69 314 0.59 1.81 382 0.84
0.6 1.79 286 0.41 1.64 303 0.63 1.67 399 0.87
0.65 1.71 287 0.48 1.63 296 0.65 1.98 433 0.86
DOI: 10.1002/eqe

less. Such differences are small for practical purposes, but they should be viewed in the light of the fact
that the development of the analytical models was based on extensive experimental data and that the
tested structures lacked the complexity of real structures. For real building structures, in which
knowledge of properties is incomplete, errors in the analytical prediction of response will likely much
exceed 15%. Also, note that the errors in the analytical prediction could result in either over-prediction
or under-prediction of the exact response.
An example of the analytical prediction of axial isolator forces is presented in Figure 8 for the test of
the moment frame with friction pendulum bearings and nonlinear viscous dampers in the Sylmar 90°
motion. The bearings undergo uplift that is correctly captured in the analysis. It is interesting to note
that the same structure for the same seismic motion but with linear dampers or without dampers did
not experience uplift. Also, it should be noted that the dampers were inclined in space and provided
a component of damping force in the vertical direction at the mid-point between the isolators in the
direction of motion. Moreover, the same structure for the same seismic motion but with the
elastomeric bearings did not experience uplift (tension) when linear dampers were used but did so
when tested without dampers or with nonlinear dampers.
The analytical prediction of more complex response parameters such as peak floor acceleration values
and floor response spectral values may contain larger errors. For example, the floor response spectral
Figure 7. Comparison of experimental and analytical results for the friction pendulum/linear viscous damper
isolated moment frame in the El Centro S00E 200% excitation.
DOI: 10.1002/eqe

values in Figure 6 are under-estimated in the analysis by as much as 50%. Other examples presented in
[9] show over-estimation of peak floor accelerations and floor spectral values by more than 30% when
compared to exact values. It is expected that errors in estimating peak values of acceleration and floor
response spectra of real structures to be higher than those observed in the comparisons of this study.
SUMMARY AND CONCLUSIONS
Based on a total of 227 experiments on flexible moment frame and stiff braced frame isolated structures
with elastomeric bearings or friction pendulum bearings and linear or nonlinear viscous dampers with 14
historic ground motions of far-fault and near-fault characteristics, the following observations were made:
(1) Added damping, whether linear or nonlinear viscous, results in reductions of isolator displace-
ments, drifts and story shear forces in structures with low damping elastomeric isolation
systems. The level of added effective damping was about 20% in the experiments. The elasto-
meric bearing effective damping was 5%. The amount of reduction of isolator displacements
was greatest in the nonlinear viscous dampers. However, the amount of reduction of drift and
shear forces was greatest in the linear viscous dampers.
(2) Added damping, whether linear or nonlinear viscous, results in reductions of isolator dis-
placements in structures with highly nonlinear and highly damped isolation systems, such
as the tested friction pendulum system. The level of added effective damping was about
25% in the experiments. The effective damping contributed by the friction pendulum isola-
tors was in the range of 20–30% depending on the amplitude of displacement. However,
the addition of dampers generally increased drifts and story shear forces, and the amount
of increase was noticeably more when nonlinear dampers were used.
Figure 8. Comparison of experimental and analytical histories of bearing axial loads for the friction
pendulum/nonlinear fluid viscous damper isolated moment frame in the Sylmar 90° 100% excitation
where uplift occurred.
DOI: 10.1002/eqe

(3) Added damping had mixed effects on floor accelerations and floor total velocities. Nevertheless,
the addition of nonlinear dampers always resulted in higher floor accelerations than when linear
dampers were added.
Comparisons of analytical (using the commercial program SAP2000) and experimental results
demonstrated good capability to predict structural system response parameters such as isolator
displacements, interstory drifts, story shear forces, base shear forces, axial bearing forces, and floor
total velocities. In general, errors in the predictions of these quantities were less than 15% of the
exact values. However, peak floor accelerations and floor spectral values were under-predicted by as
much as 50% or over-predicted by as much as 30% in the well-identified and simple structural
systems that were tested. It is expected that errors in the analytical prediction will be larger in actual
complex and less known in properties structural systems.
The presented results provide information to justify the use of linear viscous dampers, as compared
to nonlinear viscous dampers, in seismic isolation systems as providing reduction in isolation system
displacements and in having beneficial or the least detrimental effects on drifts, forces, velocities,
and accelerations. After all, the observations of this study are consistent with those of Kelly [3],
Hall [4], Hall and Ryan [6]. and Politopoulos [5]. That is, too much damping is detrimental. This
study made clear that when supplemental damping is needed, it is best to add linear viscous damping.
Moreover, the presented results caution on the blind use of response history analysis results on peak
values of accelerations and floor response spectra for assessing performance and for the design of
secondary systems.
ACKNOWLEDGEMENTS
Partial support for the work presented in this paper has been provided by the Multidisciplinary Center for
Earthquake Engineering Research, University at Buffalo, Buffalo, NY, USA. The friction pendulum isola-
tors were manufactured by Earthquake Protection Systems of Vallejo, California. The viscous dampers were
manufactured by Taylor Devices of North Tonawanda, NY, USA. The elastomeric bearings were
manufactured by the Dynamic Isolation Systems of Sparks, Nevada, USA.
REFERENCES
1. Constantinou MC, Whittaker AS, Kalpakidis Y, Fenz DM, Warn GP. Performance of seismic isolation hardware un-
der service and seismic loading. Report No. MCEER-07-0012, Multidisciplinary Center for Earthquake Engineering
Research, Buffalo, NY, 2007.
2. Taylor DP. Private communication with MC Constantinou, October 2013.
3. Kelly JM. The role of damping in seismic isolation. Earthquake Engineering and Structural Dynamics 1999; 28:3–20.
4. Hall J. Discussion. The role of damping in seismic isolation. Earthquake Engineering and Structural Dynamics
1999; 28:1717–1720.
5. Politopoulos I. A review of adverse effects of damping in seismic isolation. Earthquake Engineering and Structural
Dynamics 2008; 37:447–465.
6. Hall JF, Ryan KL. Isolated buildings and the 1997 UBC near-source factors. Earthquake Spectra 2000; 16(2):393–411.
7. Providakis CP. Effect of LRB isolators and supplemented viscous dampers on seismic isolated buildings under near-
fault excitations. Engineering Structures 2008; 32:1187–1198.
8. Providakis CP. Effect of supplemental damping on LRB and FPS seismic isolators under near-fault ground motions.
Soil Dynamics and Earthquake Engineering 2009; 29:80–90.
9. Wolff ED, Constantinou MC. Experimental study of seismic isolation systems with emphasis on secondary system
response and verification of accuracy of dynamic response history analysis methods. Report No. MCEER-04-0001,
Multidisciplinary Center for Earthquake Engineering Research, Buffalo, NY, 2004.
10. Kalpakidis IV, Constantinou MC. Principles of scaling and similarity for testing of lead-rubber bearings. Earthquake
Engineering and Structural Dynamics 2010; 39(13):1551–1568.
11. Mokha AS, Constantinou MC, Reinhorn AM. Experimental study and analytical prediction of earthquake response
of a sliding isolation system with spherical surface. Report No. NCEER-90-0020, National Center for Earthquake
Engineering Research, State University of New York, Buffalo, New York, 1990.
12. Computers and Structures Inc. SAP2000. Integrated finite element analysis and design of structures. Version 7.44,
Berkeley, California, 1998.
13. Tsopelas PC, Constantinou MC, Reinhorn AM. 3D-BASIS-ME: computer program for nonlinear dynamic analysis
of seismically isolated single and multiple structures and liquid storage tanks. Report No. MCEER-94-0010,
Multidisciplinary Center for Earthquake Engineering Research, Buffalo, New York, 1994.
14. Sarlis AA, Constantinou MC. Modeling triple friction pendulum isolators in program SAP2000. Document
distributed to the engineering community together with example files, University at Buffalo, 2010.
DOI: 10.1002/eqe

Experimental Assessment of the Seismic Response of a Base-Isolated Building Through a Hybrid Simulation Technique

More Related Content

Similar to Experimental Assessment of the Seismic Response of a Base-Isolated Building Through a Hybrid Simulation Technique (20)

Recently uploaded (20)

Experimental Assessment of the Seismic Response of a Base-Isolated Building Through a Hybrid Simulation Technique