The Professional Journal of the Earthquake Engineering Research Institute

PREPRINT
This preprint is a PDF of a manuscript that has been accepted for
publication in Earthquake Spectra. It is the final version that was
uploaded and approved by the author(s). While the paper has been
through the usual rigorous peer review process for the Journal, it has
not been copyedited, nor have the figures and tables been modified for
final publication. Please also note that the paper may refer to online
Appendices that are not yet available.

We have posted this preliminary version of the manuscript online in

the interest of making the scientific findings available for distribution
and citation as quickly as possible following acceptance. However,
readers should be aware that the final, published version will look
different from this version and may also have some differences in
content.

The DOI for this manuscript and the correct format for citing the paper are
given at the top of the online (html) abstract.

Once the final, published version of this paper is posted online, it will
replace the preliminary version at the specified DOI.
Application of isolation to high-rise buildings:
A Japanese design case study through a US
design code lens

Tracy C Becker,a) M.EERI, Shunji Yamamoto,b) Hiroki Hamaguchi,c)

Masahiko Higashino,c) M.EERI, and Masayoshi Nakashima,d) M.EERI

Base isolation of high-rise buildings has growing popularity in Japan, yet it is

uncommon in most of the world. While tall buildings already have long periods
and thus lower input accelerations, the addition of isolation can decrease
interstory drifts and greatly decrease floor acceleration, protecting building
content. By protecting building content, high-rises can be kept fully operational
and occupiable after earthquakes. The Japanese design code has clearly outlined
procedures for designing isolated high-rises, facilitating the implementation of
isolation; however, other design codes, and specifically the US code, make the
adoption of isolation difficult. Using a design representative of typical isolated
high-rises in Japan, it is shown that while isolation is feasible under US design
levels, requirements are much more stringent and some changes from the
Japanese design would be required to make the design acceptable under the US
code.

INTRODUCTION

To-date, just under 200 isolated high-rise buildings, ranging from 60 to 180 meters tall,
have been constructed in Japan, with the most common implementation being for concrete
residential condominium buildings. While tall buildings, due to their long periods, already
have reduced response levels compared to low or mid-rise buildings, the addition of base
isolation is meant to further increase the performance. A specific target is to limit maximum

a) Civil Engineering Dept, McMaster University, 1280 Main St W, Hamilton, Ontario, L8S-4L8 Canada
b) Building Design Dept, Osaka Main Office, Takenaka Corporation, 4-1-13, Hommachi, Chuo-ku, Osaka,
541-0053 Japan
c) Research and Development Institute, Takenaka Corporation, 1-5-1, Ohtsuka, Inzai, Chiba 270-2329, Japan
d) Disaster Prevention Research Institute, Kyoto University, Gokasho, Uji, Kyoto 611-0011, Japan

1
floor accelerations to below 0.2 g to avoid damage to acceleration sensitive content,
including overturning of furniture, which is more likely to be unbraced in residential homes.
Acceleration induced damage to building content was observed in fixed-base, high-rise
buildings in the 2011 Tohoku earthquake, both in Sendai as well as in Tokyo, 300 km away
(Hoki et al. 2011, Kubo et al. 2012). In these buildings, shelving, shelf content and partition
walls fell and appliances moved to an extent that threatened the safety of nearby people and
required significant clean up. Limiting response levels to ensure continued occupancy in
residential high-rises has large potential benefits for post-earthquake response, as housing
displaced people is a major issue.

By comparison, the use of isolation in the United States has been limited. One of the main
factors is the complexity in the design code for isolated buildings which results in increased
time and cost spent on a project (Ryan et al. 2010). Thus, the majority of isolation use has
mainly remained within essential, public buildings such as hospitals, city halls or 911 centers
(Higashino and Okamoto, 2006). Isolation has not spread in use to typical office or
residential buildings. Perhaps due to limited isolation projects in the US, most practitioners
view the use of isolation for high-rise buildings as impractical and unfeasible and high-rise
isolation has only been implemented in one instance for a historical retrofit (Youssef et al.
2000). These views are clearly at odds with the state-of-practice in the Japanese design
community.

Major issues of concern to the US community can be separated into concerns for the
superstructure performance and isolation layer performance. On the superstructure side, are
there benefits to adding isolation to an already very flexible structure, and could adding
isolation be detrimental in some cases? In traditional base isolation, the period of the
isolation system is typically 3+ times longer than the period of the superstructure. This large
period separation means that the first mode shape includes very little deformation in the
superstructure. Additionally, this means that the correlation between the modes is very small,
limiting higher-mode contributions. Isolated high-rise buildings typically have fixed-base
first mode periods ranging from 2.5 to 4.5 seconds, while the period of the isolation systems
generally range from 4 to 7 seconds. The ratio of the isolation period to the superstructure
period is typically below 2, and thus, less benefit from the added isolation is expected.

In terms of isolator performance, can the large displacement demands for the isolation
layer be accommodated with currently manufactured bearings, and do the large overturning

2
moments resulting from the high center of gravity in tall buildings cause problems with uplift
and high compressive loads in isolation bearings? Isolation consistently reduces the
acceleration response of isolated structures. However, in trade-off, the isolation displacement
is large. In Japan, rubber bearings are the most commonly used isolator type; however, the
size and related displacement capacity of these bearings is limited. Additionally, in rubber
bearings, uplift behavior may cause delamination and reduction in horizontal stiffness, while
extreme compressive loads cause buckling. This may cause for concern for both the behavior
of the isolation system as well as the superstructure as differential settlements can result in
large P-Delta forces in high-rise buildings.

To investigate these issues, the performance of a case study isolated high-rise building
was compared to a fixed-base counterpart. The isolated high-rise model is representative of
already constructed high-rise buildings in Japan, selected after a survey of designs was
completed. Both Japanese and US design codes are considered, and ground motions
conforming to the respective code requirements are used in the study.

USE OF HIGH-RISE ISOLATION IN JAPAN

After the 1995 Kobe Earthquake, there was a large demand in Japan for structures of
increased seismic safety. Before the earthquake, Japan had an existing, well-organized
isolation design code with a history of almost fifteen years. The combination of large demand
and a streamlined design code lead to an explosion in popularity of base-isolated buildings
(Clark et al, 1999), and there are now multiple thousands of isolated buildings. With a
population of over 100 million, a tenth of which live in Tokyo, Japan has welcomed high-rise
buildings and especially high-rise condominiums. Inhabitants of high-rise condominiums
often represent wealthier clients who are willing to spend larger amounts upfront to gain
better performance, as is often the case with isolation.

Thus, isolation has been adopted for the improved performance of high-rise buildings,
and its use is not uncommon. In Japan high-rise buildings are defined as greater than 60 m in
height. As of the end of 2011, there were over 170 isolated high-rise buildings. A histogram
of the building heights is shown in Figure 1. While the majority of the constructed buildings
have superstructures below 100 m, there are buildings up to 175 m tall. For the most part
these buildings have concrete superstructures; less than 10% of constructed isolated high-
rises are steel. Isolation systems are typically natural rubber bearings in combination with

3
either high damping rubber bearings or lead plug rubber bearings. Flat sliding bearings or
linear rolling bearings are sometimes used in the isolation systems along with rubber bearings
to elongate the period of the isolation system. Additional damping is often provided with
steel U dampers, viscous dampers or a combination of damping devices. In general,
supplemental damping is located only in the isolation plane in high-rise isolated buildings.

After the 2011 Tohoku Earthquake, changes have been made in disaster response
policies. As cities are densely populated, it is desired that people remain in buildings directly
after earthquakes to relieve congestion for emergency services, minimize the need for
temporary housing and reduce stress on transportation services. To achieve this goal,
buildings must be fully functional after large earthquakes. This need will continue to drive
the use of isolation in high-rise buildings.

Constructed Isolated High−Rises

80
70
60
50
Number

40
30
20
10
0
60 80 100 120 140 160 180
Height (m)

Figure 1. Heights of constructed isolated high-rises in Japan between 2000 and 2012.

BUILDING DESCRIPTION

SUPERSTRUCTURE

The US design code allows the use of a force reduction factor on design basis earthquake
forces in the design of isolated superstructures (ASCE 2010). Vassilou et al (2013) found
that, while short and mid-rise isolated buildings designed with force reduction factors exhibit
significantly larger than expected post yield displacements, flexible high-rise buildings will
respond following the traditional equal displacement rule in which the ductility is equal to the
force reduction factor. However, due to detrimental effects of P-delta forces in high rise
buildings, force reduction factors may not be conducive to stable behavior in large
earthquakes. In general Japanese superstructures are designed to remain nominally elastic

4
under design Level 2 earthquakes, which have a return period of approximately 500 years
and are comparable to the US design basis motions (Pan et. al, 2005).

This case study uses a design representative of isolated high-rise buildings already
constructed in Japan. The three-bay by three-bay building is 32-stories, totaling 104 m above
the isolation system. The aspect ratio of the building is just over five. The weight of the
building is 220 MN. The lateral resisting system is a concrete moment frame, designed to
have a yield strength of 0.08 g base shear. The superstructure has a fixed-base period of 2.67
s, based on the gross cross-sectional properties. As assuming Rayleigh damping can cause
underestimation of isolated building responses (Ryan and Polanco, 2008), stiffness
proportional damping is assumed with 3% in the 2nd mode of the isolated building. For
comparison purposes, a model for a fixed-base counterpart was developed. The fixed-base
model maintained the same building configuration as the isolated model, with member sizes
increased so that the strength of the building was 0.14 g base shear, a value common for
Japanese reinforced concrete buildings of similar height. The fixed-base building has a
fundamental period of 2.03 s, with 5% Rayleigh damping assumed.

In the Japanese design code, analysis procedures for high-rise and isolated buildings are
the same. A first mode pushover analysis is conducted on a full frame model of the building.
From the pushover analysis, the individual story backbone curves are computed and assigned
to a lumped-mass stick model. This nonlinear lumped-mass model is then used for the
dynamic validation of the design (Otani, 2004, Pan et. al, 2005). While the lumped-mass
model includes higher mode effects, only the first mode is matched to the full frame. Thus, a
full 2D frame model was adopted for dynamic analysis in this study. The building was
modeled in the analysis software SAP2000. As a 2D model was used, the four building
frames were modeled in parallel with their displacement responses linked. Fiber joint hinges
were used to capture the nonlinear behavior of the building.

ISOLATION LAYER

In Japanese practice, all isolation bearings are prequalified before they can be made
commercially available. The qualifications are done through a government agency with
testing requirements similar to the prototype tests required in the US design code. The
manufacturer can then list all the qualified bearings in a catalogue which provides bearing
properties, range of applicability and modeling guidelines, and no project specific testing of

5
bearings is mandatory, unlike in US design. However, similar to US design quality control
tests, all bearings undergo simple tests to verify their properties before installation. As exact
bearing properties may not be known until these tests are completed, designs are done with
upper and lower bounds of bearing properties, similar to US practice.

Currently, in Japanese designs, rubber bearings are almost exclusively used as the
principal isolation device. Thus, building height is limited by the axial load capacity of the
rubber bearings. The largest rubber bearings available in Japan are 1,600 mm in diameter
which have a 1,000 mm displacement limit (Ministry of Land, Infrastructure and Transport,
2007). The rubber bearings with the largest displacements capacity are 1,500 mm in diameter
which have a 1,200 mm displacement limit. Larger bearings have not been produced because
of both the limited size of manufacturing facilities and the issue of proper vulcanization in
extremely large rubber bearings.

The isolation system for the case study is composed of natural and high damping rubber
bearings with additional oil dampers and is typical of Japanese design. A four by four grid of
rubber bearings are used, with one isolator under each column of the moment frame, as seen
in Figure 2. The 2D building model includes behavior in the X-direction, indicated in the
isolation plan. The isolation system for the case study includes bearings of 1,100 and 1,200
mm in diameter. Their properties for design (Japanese Ministry of Land, Infrastructure and
Transport, 2007) are given in Table 1. The upper and lower bounds for analysis take into
account variations in properties due to manufacturing variation, aging effects and temperature
variation. The natural rubber bearings selected have a variation of +26% to -15% in stiffness
from that reported in Table 1. The high damping rubber bearings used have a variation of
+41% to -25% in stiffness with a variation of -13% to -3% in equivalent damping from that
reported in Table 1.

The catalogue provides guidelines for modeling the high damping rubber bearings. The
horizontal behavior is represented using a bilinear model in which properties change based
on the displacement of the bearing. The properties for the bearings based on the provided
equations from the manufacturers are shown in Figure 3(a). The horizontal behavior of the
bearings was modeled using the “rubber bearing” link element in SAP2000. As the properties
for this element are set prior to the analysis, if the maximum displacement was significantly
far from that used to determine the bearing properties, the bearing element properties were

6
changed and the analysis was rerun. As in Japanese design practice, axial load dependencies
were not considered for the horizontal bearing behavior.

The catalogue also provides modeling guidelines for the vertical behavior of the bearings.
Initially, the tensile stiffness is 1/15th of the compressive stiffness, once a stress of 0.3 MPa is
reached, the tensile stiffness then decreases and the stiffness further decreases at 1 MPa, as
shown in Figure 3(b). The vertical behavior of the bearings was modeled separately using a
“nonlinear plastic” link element with a pivot model. The bearing manufacturer limits the
allowable tensile stress in the bearings to 1 MPa for Level 2 earthquakes. However, this value
is provided based on safety limits, data provided by the manufacturer based on tests done on
bearings up to 800 cm in diameter shows the bearings have stable behavior when cycled past
1.5 MPa (Japanese Ministry of Land, Infrastructure and Transport, 2007). The allowable

Figure 2. Layout of the isolation system

Table 1. Isolator properties

Shear Horiz. Comp.
Bearing Shape Total rubber Area Allowable
Modulus stiffness stiffness
type factor height (mm) (mm2) disp. (mm)
(N/mm2) (kN/mm) (kN/mm)
1200 mm 11286
0.62 37 248 893 - 6340
 HRB x102
1100 mm 9480
0.62 37 252 882 - 5240
 HRB x102
1200 mm 11305x
0.39 37 240 960 1.85 4570
 NRB 102

7
 1100 mm 1200 mm
8 8

(kN/mm)
eq 4 4
K

0 0
0 1 2 3 4 0 1 2 3 4
1400 1400
Qd (kN)

700 700
0 0
0 1 2 3 4 0 1 2 3 4
50 50
(kN/mm)
K1

25 25
0 0
0 1 2 3 4 0 1 2 3 4
0.2 0.2
α

0.1 0.1
0 0
0 1 2 3 4 0 1 2 3 4
Strain Strain
(a) (b)
Figure 3. (a) Horizontal properties for the bilinear model for the high damping rubber bearings where
Keq is the equivalent stiffness, Qd is the yield strength of the bearing, K1 is the initial stiffness, and α
is the ratio between the post yield and initial stiffness. (b) Vertical tension model for all rubber
bearing types.

0
Pressure (MPa)

1200 HDR
−20
1100 HDR
NRB
−40

−60
0 1 2 3 4
Strain

Figure 4. Allowable stresses in the rubber bearings at increasing levels of strain.

compressive stress is dependent on the strain in the bearing, shown in Figure 4 for the three
bearing types used.

For traditional isolation or low-rise or mid-rise structures the fundamental period of the
isolated building can be approximated at the period of the isolation layer (Naeim and Kelly,
1999). However, this no longer holds true for high-rise buildings where the period separation
is significantly smaller. For the isolated case study building, the fundamental period, which is
displacement dependent due to the nonlinear behavior of the high damping rubber bearings,
is 5.24 s at 200 mm displacement and 5.84 s at 500 mm displacement. The four viscous
dampers are linear with a damping coefficient of 2.5 kNs/mm. The total equivalent damping
of the isolation system varies from 30 – 35% with displacement. Roughly half of the damping
comes from the hysteretic damping in the high damping rubber bearings while rest is due to
the supplemental viscous dampers.

8
 EVALUATION UNDER JAPANESE DESIGN MOTIONS

GROUND MOTIONS

The Japan design code is based predominantly on the behavior under Level 2 motions,
defined as having a PGV of 500 mm/s and a return period of approximately 500 years.
Japanese design typically uses six unidirectional ground motions, the “standard set” of Taft
(1952, EW component), El Centro (1940, NS component) and Hachinohe (1968, NS
component), all scaled to 500 mm/s PGV, as well as three motions spectrally matched to the
design response spectrum, modified for the local site properties. Full details of the Japanese
response spectrum can be found in Higashino and Okamoto (2006).

The three spectrally matched motions used in the case study are referred to as Kokuji
Kobe L2, Kokuji Random L2 and Kokuji Hachi L2, the time histories for which are shown in
Figure 5. “Kokuji” is a reference to the code specification for ground motion selection and
creation (Japanese Ministry of Construction, 2000). Kokuji Kobe L2 and Kokuji Hachi L2
are derived from the phase content of the 1995 Kobe and the 1968 Hachinohe ground
motions, respectively. Kokuji Random L2 is derived from random phase input. The response
spectra for all six motions as well as the design spectrum are shown in Figure 5. The 5%
damped Japanese design spectrum is roughly 2.5 times larger than the 30% damped shown. It
can be seen that the methods of scaling result in similar spectra for all the standard set
motions and similar spectra for all the code matched motions. Thus, the variation between the
building responses under the six motions is not expected to be large.

To verify that allowable axial loads in bearings are not exceeded, designers sometimes
run vertical analyses separately and combine the axial loads with those from the
unidirectional analysis to compare against the specified limits. Alternatively, designers
sometimes add on a static load value of approximately 0.3 g to the maximum axial responses.
In the case study the second method will be used.

Although not explicitly required in the design code, the maximum bearing responses
under a Level 3 earthquake motion are often checked. Under this level, the bearing tension
may exceed the 1 MPa tension limit; however, displacements and compressive stresses
should remain within the manufacture specified limits. The Level 3 motion is obtained by
amplitude scaling the Level 2 motion by 1.5 (similar to the US relationship between design

9
 Kokuji Kobe L2 30% damped response spectra
0.4 0.8 El Centro

Accel (g)
0.2 Taft

Psuedo Accel (g)

0 0.6 Hachinohe
−0.2 Kokuji Kobe L2
0.4
−0.4 Kokuji Random L2
0 20 40 60 80 Kokuji Hachi L2
Kokuji Random L2 0.2
0.4 Design spectrum
Accel (g)

0.2
0
0 0 2 4 6 8
600
−0.2
−0.4
0 30 60 90 120

Disp (mm)
Kokuji Hachi L2 400
0.4
Accel (g)

0.2
0 200
−0.2
−0.4 0
0 30 60 90 120 0 2 4 6 8
Time (s)

Figure 5. Time histories of the Japanese spectra matched motions and response spectra of all
Japanese design motions.

basis and maximum considered motions). For the case study, the two Japan motions with
highest levels of response were rerun for the isolated building at Level 3 scale.

BUILDING PERFORMANCE

The fixed and isolated high-rise building models were analyzed under the six Japanese
Level 2 ground motions, using the isolator properties given in Table 1 for the isolated
building. The peak drift ratio, defined as the maximum drift of all stories and peak base
shears for the fixed and isolated buildings are listed in

Table 2, as well as the isolation layer displacement for the isolated building. As expected,
the variation between responses to the standard set motions and to the spectrally matched
motions is minimal. For both the isolated and fixed-base buildings, there was no significant
yielding in the superstructure. For the isolated building, the axial stresses in the bearings were
well within the manufacturer provided limits. For the motion resulting in the largest axial
load fluctuation, Kokuji Hachi L2, the tension stresses remained below 30% of the maximum
allowed and compressive stresses remained below 40% of the maximum allowed.

The isolated buildings exhibited peak story drifts of less than half of the fixed-base
building. However, to begin with fixed-base roof drifts ranged from 0.4 to 0.6%, below levels

10
at which damage to nonstructural partition walls occurs, roughly 0.6% (ATC-58, 2007).
Major improvement in behavior is seen when the roof level accelerations for the fixed and
isolation buildings are compared. The roof level response spectra are shown in Figure 6. The
mean peak floor acceleration is reduced from 0.61 g in the fixed-base case to 0.16 g in the
isolated building. This is below the target level of 0.2 g, at which level content may topple or
fall (ATC-58, 2007). For the equipment frequency range, roughly 2-10 Hz, the response
spectra are consistently reduced by the use of isolation, and little to no damage would be
expected to acceleration sensitive content. In a full building design, upper and lower values
of bearing properties would be used as well; however, as the behavior is well within the
allowable responses, this effect is not investigate here for Japanese motions.

Table 2. Peak isolation displacement and roof drift under Japanese Level 2 design ground motions
Kokuji Kokuji Kokuji
Motion: El Centro Taft Hachinohe
Kobe L2 Random L2 Hachi L2
Isolator disp.
180 170 170 270 250 260
(mm)
Isolated peak
0.21 0.25 0.17 0.25 0.23 0.26
story drift (%)
Isolated base
4.5 4.0 4.6 5.3 5.4 5.4
shear (%g)
Fixed-base peak
0.44 0.39 0.39 0.57 0.53 0.49
story drift (%)
Fixed base
21.8 23.1 19.5 23.2 22.1 23.4
shear (%g)

5% damped roof spectra

3 35
Mean Isolated
Mean Fixed−base 30
2.5
Mean
25 Isolated
2
Mean
20 Fixed
SA (g)

Story

1.5
15
1
10

0.5 5

0 0
−1 0 1 0 0.2 0.4 0.6 0.8
10 10 10
Frequency (Hz) Peak Floor Acceleration (g)

Figure 6. Roof level response spectra and mean peak floor accelerations vs height for the Japan Level
2 design motions in the isolated and fixed-base buildings.

11
 Motions Kokuji Kobe L2 and Kokuji Hachi L2 which resulted in the largest responses
were amplified to the Level 3 scale and the isolated building was analyzed under the
amplified motions, now referred to as Kokuji Kobe L3 and Kokuji Hachi L3. Under these
motions the peak roof drift ratio was less than 0.4% and the isolator displacement was less
than 430 mm. While it is not required that the tension stress limit be observed for the Level 3
motions, the maximum tension stress remained within 85% of 1 MPa. The maximum
compression stress remained within 50% of the allowable limit. Even after 0.3 g load was
added to the compressive loads on the bearings, they were still within their allowable limit.
This shows good behavior of both the isolation system and superstructure under Japanese
Level 3 motions. Thus, when isolation is selected within a Japanese design code context,
major reduction in acceleration response is gained and there is no concern about the about the
stability of the bearings or building under large earthquakes.

EVALUATION UNDER US DESIGN MOTIONS

GROUND MOTIONS

To evaluate the performance of the isolated high-rise under US code requirements, a suite
of ten ground motions, listed in Table 3, was selected. The motions were selected from
among those gathered and scaled for the PEER Transportation Systems Research Program
(TSRP) (Baker et al. 2011). The TSRP has several groups of motions. The motions used for
this study were selected from a set of 40 site-specific ground motions for a location in
Oakland, CA 7km from the Hayward Fault on a soft rock site. Of the ten motions selected,
five are pulse-type motions. As the frame was modeled in 2D, only the fault normal
components of the records were used, with the exception of one motion, for which the fault
normal and fault parallel components were run separately and the responses vectorally added
to investigate the contribution of 3D loading, which is required in the US design code.
Vertical components were not included.

The TSRP report gives different scale factors for each motion (listed in Table 3) to match
the 2% in 50 years level hazard spectrum, which is the code defined maximum considered
event (MCE), and the 10% in 50 years hazard spectrum, which is taken here as the design
basis event (DBE). Baker et al. used the uniform hazard spectra to match the motions. The
spectra modified for 30% damping is shown in Figure 7, the 30% damped spectra is found

12
from the 5% damped spectra by dividing by 1.7, the ASCE 7 damping factor (ASCE, 2010).
It is worth noting that the motions always have a scale greater than (or in one case equal to)
one; that is, they are always increased from their original recording. The DBE level has a
return period of 475 years, similar to the Japanese Level 2. The DBE scaled motions are used
to examine superstructure performance and compare against the fixed-base counterpart. The
US building code requires that isolator stability is checked for the MCE level (ASCE, 2010).
In addition, the high-rise superstructure should be checked for stability under the MCE level
(Tall Building Initiative, 2010). As more than seven ground motions are used in the study, the
average of the values can be used for design. One motion was run with nominal as well as
upper and lower bound bearing properties to investigate the expected variation in response.

The left side of Figure 7 shows the DBE level ground motion response spectra. Compared
to the Japanese design motions, there is much larger variation between the spectra. One
contributing factor is that the motions from the TSRP report were amplitude scaled rather
than spectrally matched, which results in smaller variation (Pant et al, 2013). The use of

Table 3. Selected ground motions with scaling values.

Record NGA Pulse DBE MCE
Earthquake Year Station
# record # Period scaling scaling
Imperial Valley-
1 179 1979 El Centro Array #4 4.61 1.46 2.48
06
Imperial Valley- El Centro
2 184 1979 5.86 1.43 2.43
06 Differential Array

3 316 Westmoreland 1981 Parachute Test Site 3.58 2.15 3.66

Superstition Westmoreland Fire

4 728 1987 - 2.28 3.88
Hills-02 Station

5 768 Loma Prieta 1989 Gilroy Array #4 - 2.26 3.84

Saratoga – Aloha
6 802 Loma Prieta 1989 4.47 1.75 2.97
Ave
Newhall – Fire
7 1044 Northridge-01 1994 - 1.00 1.70
Station
Sylmar – Converter
8 1085 Northridge-01 1994 3.49 1.45 1.45
Station East

9 1605 Duzce, Turkey 1999 Duzce - 1.08 1.85

Chi-Chi,
10 2734 1999 CHY074 - 1.67 2.85
Taiwan-04

13
 30% damped DBE response spectra 30% damped mean response spectra
1.2 1.2
1 US DBE
0.9 2 0.9 US MCE
3 Japan motions

SA (g)
SA (g)

0.6 4 0.6 US DBE

5 design spectrum
0.3 6 0.3
7
0 0
0 2 4 6 8 8 0 2 4 6 8
1200 9 800
10
900 600
Disp (mm)

Disp (mm)
600 400

300 200

0 0
0 2 4 6 8 0 2 4 6 8
Period (s) Period (s)

Figure 7. (Left) Response spectra of the DBE level motions (Right) Comparison of the mean spectra
of the US and Japan design motions.

amplitude scaling versus spectral matching changes project to project in the US, and thus, the
variation in the selected US motions for this study is on the large side. The right side of
Figure 7 compares the mean response spectra of the US DBE level and MCE level motions to
the Japanese standard set motions and spectrally matched motions. The displacements
expected from the US DBE level earthquake motions are 1.5 to 2 times larger than for the
Japanese motions at the long periods (4 – 7 s) common in isolated high-rises.

BUILDING PERFORMANCE

The fixed and isolated high-rise building models were analyzed under the ten DBE level
ground motions, using the isolator properties given in Table 1 for the isolated building. The
peak story drift ratios and the peak base shears for the fixed and isolated buildings are listed
in Table 4 as well as the isolation layer displacement for the isolated building. As expected,
the variation in the responses is much larger than for the Japan code motions. The isolated
building remained nominally elastic in all DBE level motions except Motion 1, in which
minor yielding occurred. The corresponding fixed-base building exhibited minor yielding in
Motions 1 and 6 with cracking occurring in many of the other motions. Stable behavior was
maintained under all DBE level motions.

14
 For most of the DBE motions, the use of isolation reduced the roof drift of the building;
however the degree of reduction varied significantly more than for the Japan motions,
ranging from over 50% to zero in the case of Motion 1. With the exception of this motion,
isolation kept peak story drifts well below values at which damage would be seen for drift
dependent nonstructural elements, while damage would be expected in the fixed-base
building for the majority of the motions. As with the Japan motions, major decreases in
acceleration response are seen for the isolated buildings, shown in Figure 8. The mean peak
floor acceleration was decreased from 0.8 g in the fixed-base building to 0.23 g, protecting
acceleration dependent nonstructural content. By adopting isolation, the roof level response
spectra are decreased consistently over the full range of content frequencies, compared to the
fixed-base roof spectra.

The isolated building model was further analyzed under the ten MCE level motions. The
peak isolator displacements and peaks story drifts are listed in Table 4. The superstructure
remained nominally elastic for all but four motions, Motions 1, 2, 6 and 9. With the exception
of Motion 1, all peak story drift values remained below 1%. Motion 1 resulted in significant
yielding concentrated in the lower levels of the superstructure where peak story drifts reached
almost 4%. The Tall Building Initiative suggests that the mean of the peak story drifts should
not exceed 3% while the maximum from all motions should not exceed 4.5%. Thus, the
superstructure behavior is deemed acceptable under the US code for MCE level motions.

Table 4. Peak isolation displacement and roof drift under US design ground motions

Motion number: 1 2 3 4 5 6 7 8 9 10

Isolator disp. (mm) 850 570 330 370 180 460 370 370 600 180
Isolated peak story
0.77 0.35 0.29 0.26 0.36 0.46 0.37 0.38 0.35 0.20
drift (%)
DBE level

Iso. base shear (%g) 18.7 10.8 6.7 7.1 5.8 8.3 8.7 5.8 9.5 3.9
Fixed-base peak
0.78 0.60 0.64 0.53 0.74 0.81 0.68 0.65 0.43 0.37
story drift (%)
Fix. base shear (%g) 34.3 27.8 33.4 24.8 30.4 36.5 31.7 33.0 24.1 26.3

Isolator disp. (mm) 1090 820 570 680 320 730 640 620 740 310
MCE level

Isolated peak story

3.91 0.95 0.48 0.44 0.57 0.84 0.60 0.62 0.69 0.31
drift (%)
Iso. base shear (%g) 25.1 21.1 10.5 12.1 9.5 14.7 14.8 11.8 17.2 6.7

15
 5% damped roof spectra
4 35
Mean Isolated
3.5 30
Mean Fixed−base
3 Mean
25 Isolated
2.5 Mean
20 Fixed
SA (g)

Story
2
15
1.5
10
1

0.5 5

0 0
−1 0 1
10 10 10 0 0.2 0.4 0.6 0.8
Frequency (Hz) Peak Floor Acceleration (g)

Figure 8. Roof level response spectra and mean peak floor accelerations vs height for the US DBE
level design motions in the isolated and fixed-base buildings.

For the MCE level motions, the average of the isolator displacement response is 650 mm.
The fault perpendicular component was run for Motion 6, resulting in a displacement of 430
mm. When added to the displacement from the fault normal component by vector sums, the
displacement for Motion 6 increases by 20%. The fault normal component of Motion 6 was
also run with upper and lower bounds on bearing properties. Lower bounds on properties
resulted in an increase in displacement of 10%. If the average of the isolator displacement
responses is increased by 30% to adjust for lower bound properties and bidirectional loading,
the average displacement would be 860 mm, still below the maximum allowable
displacement of the bearings.

Bearing vertical loads

For the Japanese design motions all vertical loads were within the specified design limits;
however, this was not the case for the US design motions. For the US DBE level design
motions, two motions resulted in axial stresses larger than specified. Under Motion 6 the
tension limit was exceeded by 4%, with a maximum tensile stress of 1.04 MPa. Motion 1
resulted in larger than allowable values for both tension and compression, with the tension
limit exceeded by up to 30% and the compression limit exceeded by up to 50%.

For the MCE level, the tension limit of 1 MPa was exceeded under all but one motion. It
should be remembered that this limit is not strictly enforced for the Japanese Level 3. The
details of how the limits were exceeded are listed in Table 5. With the exception of Motion 1,

16
the maximum that the tensile stress 1.35 MPa. Previous tests on rubber isolators under
combined shear and tension have demonstrated that these bearings can undergo tensile strains
that greatly exceed the cavitation strain. According to Kelly and Konstantinidis (2011), the
reason for this is that when a bearing is displaced laterally, even by a slight amount, any
tension results in a rotation of the middle layers which gives the vertical load a component
along the layer, thus inducing shear deformation. Consequently, the bearing displaces
vertically by an amount that is considerably larger than that which would be possible under
pure tension.

The vertical tension stresses versus the horizontal bearing displacement for the bearings
in Frame D (Figure 2) under Motions 2, 6 and 9 are shown in Figure 9. These motions
resulted in tensile limits being exceeded by more than 20%. If the overturning forces are in
phase with the isolation layer displacement, the maximum tension forces would occur at large
bearing displacements. This is the case for the majority of the instances of exceedance.
However, as seen in Figure 9 some bearings are in tension while passing through zero
displacement. Limited experimental data exists on the tensile capacity of rubber bearings
while instantaneously passing through a state of zero shear. Future research on this is
necessary.

The manufacturer allowed compressive stresses were exceeded under the MCE level of
Motions 1 and 2. The vertical stresses versus the horizontal displacement for the bearings in
Frames C and D are shown for these two motions in Figure 10 and 11. As Motions 1 and 2
are recordings from the same earthquake event, similarities between the behaviors in the two
motions are expected. In both motions, there was only one instance of the compressive limit

Table 5. Bearing tension under US MCE level motions

Record number 1 2 3 4 5 6 7 8 9 10

Limit exceeded? Yes Yes Yes Yes Yes Yes Yes Yes Yes No

Number of bearings exceeded 12 8 4 4 2 8 3 2 11 -

Amount exceeded by (%) >100 <35 <6 <4 <2 <30 <10 <12 <24 -
Number of times exceeded
2 2 1 1 1 1 1 1 4 -
(max of all bearings)
Max duration of exceedance (s)
1.8 1.25 0.45 0.40 0.15 0.75 0.50 0.55 1.05 -
(max of all bearings)

17
 Bearing D1 Bearing D2 Bearing D3 Bearing D4
Axial Stress (MPa) 1.5 1.5 1.5 1.5
Motion 2

1 1 1 1

0.5 0.5 0.5 0.5

0 0 0 0
−1000 0 1000−1000 0 1000 −1000 0 1000−1000 0 1000
1.5 1.5 1.5 1.5
Axial Stress (MPa)
Motion 6

1 1 1 1

0.5 0.5 0.5 0.5

0 0 0 0
−1000 0 1000−1000 0 1000 −1000 0 1000−1000 0 1000
1.5 1.5 1.5 1.5
Axial Stress (MPa)
Motion 9

1 1 1 1

0.5 0.5 0.5 0.5

0 0 0 0
−1000 0 1000−1000 0 1000 −1000 0 1000−1000 0 1000
Horiz Disp (mm) Horiz Disp (mm) Horiz Disp (mm) Horiz Disp (mm)

Figure 9. Vertical tension stress versus horizontal bearing displacement for the bearings of Frame D
for three motions. Dotted line represents maximum allowed stress by the bearing manufacturer.

being exceeded. In Motion 2 this occurred in only bearings in Line 4 (Figure 2) of the
isolation system. In Motion 1, which consistently resulted in the most extreme behavior, the
compressive limits were exceeded by large amounts in bearings in both Lines 3 and 4, and
the bearings exceeded their maximum allowed displacement. While the manufacturer
provided limits have safety factors included, implications of buckling behavior are large. At
buckling, axial load capacity is lost and redistribution of the axial loads to other bearings
could cause concern for their stability. Concern also arises if bearings on one side of the
building decrease significantly in height which could exacerbate P-delta effects in the rise-
rise superstructure.

The US design code relies on the average response of the motions for design check.
Averaging the maximum and minimum stresses on the bearings and comparing them to the
design limits at the average displacement of 650 mm, no bearing exceed their compression
limit and only the corner bearings exceed their limits in tension by up to 15%. Using bearing

18
 Bearing C1 Bearing C2 Bearing C3 Bearing C4
0 0 0 0

Axial Stress (MPa)

−20 −20 −20 −20

−40 −40 −40 −40

−60 −60 −60 −60

−1000 0 1000 −1000 0 1000 −1000 0 1000 −1000 0 1000
Bearing D1 Bearing D2 Bearing D3 Bearing D4
0 0 0 0
Axial Stress (MPa)

−20 −20 −20 −20

−40 −40 −40 −40

−60 −60 −60 −60

−1000 0 1000 −1000 0 1000 −1000 0 1000 −1000 0 1000
Horiz Disp (mm) Horiz Disp (mm) Horiz Disp (mm) Horiz Disp (mm)

Figure 10. Vertical stress versus horizontal bearing displacement for the bearings of Frames C and D
for Motion 1. Dotted line represents maximum allowed stress by the bearing manufacturer.

Bearing C1 Bearing C2 Bearing C3 Bearing C4

0 0 0 0
Axial Stress (MPa)

−20 −20 −20 −20

−40 −40 −40 −40

−60 −60 −60 −60

−1000 0 1000 −1000 0 1000 −1000 0 1000 −1000 0 1000
Bearing D1 Bearing D2 Bearing D3 Bearing D4
0 0 0 0
Axial Stress (MPa)

−20 −20 −20 −20

−40 −40 −40 −40

−60 −60 −60 −60

−1000 0 1000 −1000 0 1000 −1000 0 1000 −1000 0 1000
Horiz Disp (mm) Horiz Disp (mm) Horiz Disp (mm) Horiz Disp (mm)

Figure 11. Vertical stress versus horizontal bearing displacement for the bearings of Frames C and D
for Motion 2. Dotted line represents maximum allowed stress by the bearing manufacturer.

upper bound properties for Motion 6 resulted in an increase in bearing tensile loads of 5%
and an increase in bearing compressive loads of 20%. Three dimensional analysis would
predominantly affect the corner bearings; considering the fault parallel analysis of Motion 6,
increased corner bearing compression by 20% and would ensure extreme tensile loads in one

19
or more of the corner bearings. If the averages of the tension and compression loads are
factored considering these values and compared to the design limits at 860 mm of
displacement (which was found using the lower bound properties), all corner bearings further
exceed the tension limits and 1100 mm diameter high damping bearings and the natural
rubber bearing on the sides of the frames exceed their compression limits, with the natural
rubber bearing exceeding their limits by up to 15% and the 1100 mm diameter high damping
bearings exceeding their limits by up to 70%. If a further 0.3 g is added to the average
compressive stresses to account for the effects of vertical motions, the remainder of the
natural rubber bearings would reach their compressive limit. Thus, the vertical stability of the
bearings may not be acceptable under the US design code.

ISSUES FOR FURTHER EXPLORATION

The case study showed that the isolated high-rise building significantly decreased story
drifts and floor accelerations. At the same time, multiple issues for future consideration were
encountered. Under the US MCE level motions the mean peak bearing displacement was
within the maximum allowed displacement of the system. However, the mean bearing axial
loads exceeded the maximum allowed tensile and compressive stress limits for some of the
bearings, with largest tensile loads exerted in the corner bearings. Thus, further investigation
is needed into the dynamic behavior of a group of rubber bearings in which some bearings
momentarily exceed their design limits either in tension or compression. The current US
design code is strict on isolator performance, and the same criteria are adopted for the
performance of isolators under the DBE and MCE earthquake levels. As the MCE level
earthquake is a very rare event, it may be practical to allow some level of damage to the
bearings without sacrificing the safety of the structure, as in the design of fixed-base
buildings. To explore the possibilities, more research is needed to quantify the failure
behavior of isolators as well as the overall performance evaluation of base isolated structures
in which some damage to isolators occurs.

If the high-rise building presented is to be acceptable under the current US design code,
the isolation system must be redesigned. There are multiple methods that could be explored,
including adopting uplift restraining sliding bearings or weakening the isolation plane could
to limit the accelerations transmitted to the superstructure, reducing overturning moments.
However, as the latter option would require increasing bearing displacements. As the

20
maximum displacement of the rubber bearings is currently limited by the ability to
manufacture bearings over 1,600 mm in diameter, the use of other isolation devices such as
sliding friction pendulum bearings may be considered.

CONCLUSIONS

The application of base-isolation to high-rise buildings was investigated using a case

study building model representative of high-rise buildings in Japan. The building, a 32-story
concrete moment frame, was isolated on a combination of natural rubber and high damping
rubber bearings with additional viscous dampers in the isolation layer. A fixed-base
counterpart was designed with identical layout but increased member sizes for comparison
purposes. Both buildings were analyzed under commonly used Japanese design motions as
well as a set of US design motions with design base earthquake (DBE) scales. The isolated
building was then analyzed under the US design motions with maximum considered
earthquake (MCE) scales to evaluate the suitability of the design to US design code levels.

Large differences between Japanese and US design codes were highlighted by the case
study. While Japanese design uses Level 2 ground motions for design validation, the US code
requires that the stability of the isolation system and high-rise superstructure be confirmed
under the MCE level. Japan design uses six ground motions which do not vary significantly
from design to design and for which building response does not vary significantly. By
comparison, the US design code requires motions to be chosen specifically for the project
and can result in motions that produce a large distribution of building response.

Results showed that the adoption of isolation decreased both story drifts and accelerations
in the superstructure. Peak story drifts were reduced by roughly 30-50% to levels well below
those where damage occurs to drift sensitive non-structural components such as partition
walls or cladding, which can save significantly on repair costs. Isolation resulted in major
decreases in floor level accelerations, with the mean peak roof accelerations reduced by
almost 75%. The mean peak roof acceleration of the isolated building was 0.16 and 0.23
under the Japanese Level 2 motions and the US DBE level motions respectively. At these
levels of acceleration, little to no damage would be expected in acceleration sensitive
building content.

Under Japanese Level 3 motions, the isolation system and superstructure response was
within all design limits. For the US MCE motions, superstructure response and bearing

21
displacements were within design limits. However, large overturning forces resulted in
tensile and compressive forces beyond those allowed by the bearing manufacturer, and thus,
the design presented may not be acceptable within the US code. Further research is needed on
the behavior of bearings under large tensile loads and to understand the effect of individual
bearings exceeding their design limits on the group behavior of bearings.

The implementation of isolation for high-rise buildings under the US design code may
still be feasible with redesign of the isolation system. Protecting the operability and
occupancy of high-rise buildings is of primary concern, especially for condominium
buildings in urban areas in which a large number of people would be displaced and require
temporary shelter after an earthquake. As isolation can limit superstructure as well as
nonstructural content response, it is of hope that protective systems of this nature can be
adopted for high-rise buildings in design communities outside of Japan.

REFERENCES

Applied Technology Council, 2007. ATC-58 Performance Assessment Calculation Tool (PACT).
ASCE, 2010. ASCE 7-10 Minimum Design Loads for Buildings and Other Structures, Reston:
American Society of Civil Engineers.
Baker, J.W., Lin, T., Shahi, S.K. and Jayaram, N., 2011. New Ground Motion Selection Procedures
and Selected Motions for the PEER Transportation Research Program, Report No. PEER 11/03,
Pacific Earthquake Engineering Research Center, University of California, Berkeley, CA.
Clark, P.W., Aiken, I.D., Nakashima, M., Miyazaki, M. and Midorikawa, M. 1999. The 1995 Kobe
(Hyogo-ken Nanbu) earthquake as a trigger for implementing new seismic design technologies in
Japan. Lessons Learned Over Time, Learning From Earthquakes, Volume III, EERI.
Higashino, M. and Okamoto, S. (eds), 2006. Response control and seismic isolation of buildings,
Taylor & Francis, New York.
Hoki, K., Hayashi, K., Itou, M., Furukawa, S., Enokida, R., Chung, Y.L. and Nakashima, M. 2011.
Seismic performance and damage of building structures in urban areas observed in the 2011 Off
the Pacific Coast of Tohoku Earthquake. Proceedings, Kyoto University Disaster Prevention
Research Institute Annual Meeting , Uji, Japan. (in Japanese)
Japanese Ministry of Construction, 2000. Notification No. 1461-2000: Stipulation of criteria for
structural calculations performed to confirm the safety from the perspective of structural capacity
of high-rise buildings. MC, Japan. (in Japanese)
Japanese Ministry of Land, Infrastructure and Transport. 2007. Appendix of approval certificate of
the seismic isolator No. MVBR-0341 (Bridgestone high damping rubber bearing; X0.6). MLIT,
Japan. (in Japanese).
Kelly, J.M. and Konstantinidis, D.A. 2011. Mechanics of rubber bearings for seismic and vibration
isolation, Wiley, New York.
Kubo, T., Hisada, Y., Aizawa, K. Omiya, K. and Koizumi, H. 2012. Damage report and seismic
intensity survey of super high-rise buildings in the Tokyo metropolitan area after the Tohoku
Earthquake. Journal of the Japan Association of Earthquake Engineering. 12: 5th article. (in
Japanese)

22
Mansoor, A. and Mosqueda G, 2012. Experimental simulation of base-isolated buildings pounding
against moat wall and effects on superstructure. Earthquake Engineering and Structural
Dynamic, 41: 2093-2109.
Naeim, F. and Kelly, J.M. 1999. Design of Seismic Isolated Structures: From Theory to Practice,
Wiley, New York.
Otani S. 2004. Japanese seismic design of high-rise reinforced concrete buildings: an example of
performance based design code and state of practices. Proceedings, 13th World Conference on
Earthquake Engineering, Vancouver, BC, Canada.
Pan, P., Zamfirescu, D., Nakashima, M., Nakayasu, N. and Kashiwa, H., 2005. Base-isolation design
practice in Japan: Introduction to the post-Kobe approach, Journal of Earthquake Engineering,
9(1), 147:171.
Pant, D.R., Constantinou, M.C. and Wijeyewickrema, A.C., 2013. Re-evaluation of equivalent lateral
force procedure for prediction of displacement demand in seismically isolated structures.
Engineering Structures, 52, 455:465.
Ryan, K.L., Ardent, L.A. and Larsen, D., 2010. Evaluation of design review process and requirements
for seismic-isolated buildings. Earthquake Spectra, 26(4), 1101:1116.
Ryan, K.L. and Polanco, J., 2008. Problems with Rayleigh damping in base-isolated buildings.
Technical Note, Journal of Structural Engineering, 134(11), 1780:1784.
Tall Buildings Initiative, 2010. Guidelines for the Performance-Based Seismic Design of Tall
Buildings, Report No. PEER 10/05, Pacific Earthquake Engineering Research Center, University
of California, Berkeley, CA.
Vassiliou, M.F, Tsiavos, A. and Stojadonvic, B., 2013. Dynamics of inelastic base isolated structures
subjected to analytical pulse ground motions. Earthquake Engineering and Structural Dynamics.
(available on early view)
Youssef, N., Nuttall, B., Hata, O., Tahtakran, O. and Hart, G. C. (2000), Los Angeles City Hall.
Structural Design of Tall Build., 9, 3:24.

23
 Constructed Isolated High−Rises
80
70
60
50
Number

40
30
20
10
0
60 80 100 120 140 160 180
Height (m)
 1100 mm 1200 mm
(kN/mm) 8 8
Keq

4 4
0 0
0 1 2 3 4 0 1 2 3 4
1400 1400
Qd (kN)

700 700
0 0
0 1 2 3 4 0 1 2 3 4
50 50
(kN/mm)
K1

25 25
0 0
0 1 2 3 4 0 1 2 3 4
0.2 0.2
α

0.1 0.1
0 0
0 1 2 3 4 0 1 2 3 4
Strain Strain
 0
Pressure (MPa)

1200 HDR
−20
1100 HDR
NRB
−40

−60
0 1 2 3 4
Strain
 Kokuji Kobe L2 30% damped response spectra
0.4 0.8 El Centro
Accel (g)

0.2 Taft

Psuedo Accel (g)

0 0.6 Hachinohe
−0.2 Kokuji Kobe L2
0.4
−0.4 Kokuji Random L2
0 20 40 60 80 Kokuji Hachi L2
Kokuji Random L2 0.2
0.4 Design spectrum
Accel (g)

0.2
0
0 0 2 4 6 8
600
−0.2
−0.4
0 30 60 90 120
Kokuji Hachi L2 Disp (mm) 400
0.4
Accel (g)

0.2
0 200
−0.2
−0.4 0
0 30 60 90 120 0 2 4 6 8
Time (s)
 5% damped roof spectra
3 35
Mean Isolated
Mean Fixed−base 30
2.5
Mean
25 Isolated
2
Mean
20 Fixed
SA (g)

Story
1.5
15
1
10

0.5 5

0 0
−1 0 1 0 0.2 0.4 0.6 0.8
10 10 10
Frequency (Hz) Peak Floor Acceleration (g)
 30% damped DBE response spectra 30% damped mean response spectra
1.2 1.2
1 US DBE
0.9 2 0.9 US MCE
3 Japan motions

SA (g)
SA (g)

0.6 4 0.6 US DBE

5 design spectrum
0.3 6 0.3
7
0 0
0 2 4 6 8 8 0 2 4 6 8
1200 9 800
10
900 600
Disp (mm)

Disp (mm)
600 400

300 200

0 0
0 2 4 6 8 0 2 4 6 8
Period (s) Period (s)
 5% damped roof spectra
4 35
Mean Isolated
3.5 30
Mean Fixed−base
3 Mean
25 Isolated
2.5 Mean
20 Fixed
SA (g)

Story
2
15
1.5
10
1

0.5 5

0 0
−1 0 1
10 10 10 0 0.2 0.4 0.6 0.8
Frequency (Hz) Peak Floor Acceleration (g)
 Bearing D1 Bearing D2 Bearing D3 Bearing D4
Axial Stress (MPa) 1.5 1.5 1.5 1.5
Motion 2

1 1 1 1

0.5 0.5 0.5 0.5

0 0 0 0
−1000 0 1000−1000 0 1000 −1000 0 1000−1000 0 1000
1.5 1.5 1.5 1.5
Axial Stress (MPa)
Motion 6

1 1 1 1

0.5 0.5 0.5 0.5

0 0 0 0
−1000 0 1000−1000 0 1000 −1000 0 1000−1000 0 1000
1.5 1.5 1.5 1.5
Axial Stress (MPa)
Motion 9

1 1 1 1

0.5 0.5 0.5 0.5

0 0 0 0
−1000 0 1000−1000 0 1000 −1000 0 1000−1000 0 1000
Horiz Disp (mm) Horiz Disp (mm) Horiz Disp (mm) Horiz Disp (mm)
 Bearing C1 Bearing C2 Bearing C3 Bearing C4
0 0 0 0
Axial Stress (MPa)

−20 −20 −20 −20

−40 −40 −40 −40

−60 −60 −60 −60

−1000 0 1000 −1000 0 1000 −1000 0 1000 −1000 0 1000
Bearing D1 Bearing D2 Bearing D3 Bearing D4
0 0 0 0
Axial Stress (MPa)

−20 −20 −20 −20

−40 −40 −40 −40

−60 −60 −60 −60

−1000 0 1000 −1000 0 1000 −1000 0 1000 −1000 0 1000
Horiz Disp (mm) Horiz Disp (mm) Horiz Disp (mm) Horiz Disp (mm)
 Bearing C1 Bearing C2 Bearing C3 Bearing C4
0 0 0 0
Axial Stress (MPa)

−20 −20 −20 −20

−40 −40 −40 −40

−60 −60 −60 −60

−1000 0 1000 −1000 0 1000 −1000 0 1000 −1000 0 1000
Bearing D1 Bearing D2 Bearing D3 Bearing D4
0 0 0 0
Axial Stress (MPa)

−20 −20 −20 −20

−40 −40 −40 −40

−60 −60 −60 −60

−1000 0 1000 −1000 0 1000 −1000 0 1000 −1000 0 1000
Horiz Disp (mm) Horiz Disp (mm) Horiz Disp (mm) Horiz Disp (mm)