AIJ., Design Recommendations For Seismically Isolated Buildings, 2016 PDF

20180130salehinalireza
ISBN978-4-8189-5000-9
Design Recommendations
for Seismically Isolated Buildings
AIJ-2016
Architectural Institute of Japan

@Seismicisolation
@Seismicisolation
Copyright © 2016 Architectural Institute of Japan. All Rights Reserved.

──
Design Recommendations for Seismically Isolated Buildings
──
5-26-20, Shiba, Minato-ku, Tokyo 108-8414, JAPAN
Tel: +81-3-3456-2051
Fax: +81-3-3456-2058
https://www.aij.or.jp/
──
Issued May 25, 2016
ISBN978-4-8189-5000-9 C3052
──
Edited and published by Architectural Institute of Japan
──
Notice
Copyright of this publication is proprietary to the Architectural Institute of Japan and is protected
under the Copyright Act of Japan and other acts and conventions. Private use by the individual or
other use (including reproduction, modification, or distribution) of the publication beyond the scope
permitted by the Copyright Act is prohibited without the permission of the right holder.
The Architectural Institute of Japan has made every effort to assure the accuracy of this publication
and shall not be liable for any decision or act that the User makes by using the information contained
in this publication.
@Seismicisolation
@Seismicisolation
Preface to the English Version

In Japan, it was in the early 1980s that laminated rubber bearings were first put to practical
use and the research and development of seismically isolated structure became active.
Although the first seismically isolated building, a house, was built in Yachiyo City, Chiba
Prefecture in 1982, the spread of seismically isolated structures did not proceed.
Therefore, in 1986 the Subcommittee for Seismically Isolated Structures (Chair: Hideyuki
Tada) was established by the Architectural Institute of Japan (AIJ) in order to clarify the
problems of seismically isolated structures at that time and the underlying logical composition
of the research development, as well as to present the appropriate scope of utilization of the
seismically isolated structures. The subcommittee conducted its activities assertively and
published the Design Recommendations for Seismically Isolated Buildings (first edition) in
1989. These design recommendations were based on following principles, which have
retained in all subsequent revised versions:
“This book has been created with an emphasis on the development of objective judgment
materials and criteria for supporting the decisions of structural designers. When exploring the
possibility of new structural systems, it is important to recognize the essence of the design,
and not to focus on the formal completeness of the design method. That is, where uncertain
facts are involved, it is considered that the guidelines indicate only a decision and should
avoid the convenience of the structural designer as much as possible. Because this kind of
decision may apparently make the design easier, it would dampen the creativity of the
designer and inhibit the progress of technology.”
After publishing the first edition in 1989, the Subcommittee for Seismically Isolated
Structures (Chair: Hiroshi Akiyama) has continuously striven to collect and develop
seismically isolated structure-related technologies, and, based on these results, the revised
second edition of Design Recommendations for Seismically Isolated Buildings was published
in 1993. In this book, the subcommittee confirmed that the seismically isolated structure is an
excellent choice among the various types of earthquake-resistant structures, capable of clearly
achieving high performance. The subcommittee also revealed that seismically isolated
structures should be subject to special restrictions as compared to normal buildings in terms of
building scale, structural system, maintenance management system, and so on.
After the 1995 Kobe earthquake, the continued use and functional maintenance of many
buildings became very difficult, and various problems such as repair huge costs emerged.
Since the Kobe earthquake, seismically isolated structures have come to be employed in a
large number of constructions, and significant progress has been observed in the design of
seismic isolation systems and the development of seismic isolation devices. To reflect these
developments, the Subcommittee for Seismically Isolated Structures (Chair: Akira Wada)
published the revised third edition of Design Recommendations for Seismically Isolated
Buildings in 2001.
@Seismicisolation
@Seismicisolation
Since then, the application area of seismic isolation technology has expanded such as that it is
now used in high-rise buildings. In addition, issues such as the response evaluation by long-
period ground motion have arisen. Therefore, in 2013 the Subcommittee for Seismically
Isolated Structures (Chair: Mineo Takayama) collected the latest findings and published the
revised 4th edition of Design Recommendations for Seismically Isolated Buildings, which
summarized the latest findings to support the design of seismically isolated structures.
The seismically isolated structure is a relatively new technology, and there are some building
owners and structural designers who have a sense of distrust for the new technology. The
reason why the seismically isolated structure has been able to develop in such circumstances
includes a range of factors, including the AIJ’s efforts in disseminating information about the
new technology to society by publishing the Design Recommendations; the Japan Society of
Seismic Isolation (JSSI), founded in 1993, developing various efforts for the healthy spread of
seismic isolation technology; and the seismically isolated structure having proved its
performance by demonstrating it in the 1995 Kobe Earthquake and the Tohoku Earthquake of
2011.
This English version is a translated version of the first and second parts of the Design
Recommendations for Seismically Isolated Buildings (4th edition). In the third and fourth
parts of the Japanese version of Design Recommendations, descriptions of the design
examples and design documents are included, respectively, but these have been omitted from
this translation.
We hope that this book will help to spread the use of seismic isolation technology in foreign
countries. We will be happy if the environmental improvement that have been made enable
many structural designers to tackle the seismic isolation design on a daily basis, leading to
protect the citizens and cities of the world from the threat of earthquakes.
In preparing this book, we have had cooperation from many private companies and JSSI, and
engineers both in the translation and financially, with a total budget of about 3.5 million yen,
including 0.5 million yen from the International Activity Fund of the AIJ. In particular, we
would like to express deep gratitude to Mr. Masanori Tasaka (Nikken Sekkei Ltd.), Mr.
Akitsugu Muramatsu (Taisei Corporation), Mr. Ryotaro Kurosawa (Kurosawa Construction),
Mr. Shigenobu Suzuki (Bridgestone Corporation), Mr. Yasushi Ichikawa (Nippon Steel &
Sumikin Engineering Co., Ltd.), Mr. Yasuo Tsuyuki (Kayaba System Machinery Co., Ltd),
Mr. Toru Takeuchi (Tokyo Institute of Technology), Mr. Nagahide Kani (JSSI), and Mr.
Mineo Takayama (Fukuoka University). For any inquiries about the contents of this document,
please contact Mineo Takayama (mineot@fukuoka-u.ac.jp).
April, 2016
@Seismicisolation
@Seismicisolation
Preface
This is a revised version of the third edition of Design Recommendations for Seismically
Isolated Buildings published in 2001.
The first edition was published in 1989, followed by the second one in 1993. The publication
of the series corresponded to the burgeoning of seismic isolation technology at the time. The
series is the first attempt to describe device properties used for seismically isolated building
structures from the ultimate state performance to the durability. The Hyogo-ken-Nambu
Earthquake in 1995 increased public awareness of the high performance of seismic isolation.
Since then, the number of application projects had been on an exponential curve until the
third edition was published in 2001. At that time, this technology had already been applied to
high-rise buildings of more than 100 meters. On the occasion of earthquake events such as
Tokachi Oki Earthquake 2003, Niigata-ken Chuetsu Earthquake 2004, and Fukuoka-ken
Seiho Oki Earthquake 2005 to name a few, a huge amount of data were acquired and agreed
well with the analytical expectations conducted in advance, which further accelerated
expansion of the application projects. A task committee was formed under the Steering
Committee for Structural Dynamics in 2009 to revise and elaborate the third edition and to
reflect the latest findings and innovations.
In the dawn of the technology’s history, pioneering engineers developed the devices,
designed the structures, and decided the safety margin. As time passed, computer software
for dynamic analysis has matured and useful technical manuals for devices have been
prepared. Seismic isolation is now regarded as a mature technology rather than a cutting edge
one. In fact, it has been applied to a wide variety of structures, from residential houses to
super high-rise buildings. In spite of the maturity of seismic isolation technology, the
intensity level of the ground motion for structural design has become the main issue from the
seismological point of view. This trend has been accelerated by the latest event, the Tohoku
Earthquake in 2011. The practical issue faced by engineers who would like to design
seismically isolated buildings is now how to select an appropriate device according to the
increasing intensity level of an earthquake.
The task committee hopes that the fourth edition will help structural engineers to expand the
seismic isolation technology to create a safer and better society by natural disaster mitigation.
October, 2013
@Seismicisolation
@Seismicisolation
Introduction
These Recommendations relate to seismically isolated structures. Such structures contain a
seismic isolation level that includes isolators, which mainly support the building and can
deform flexibly in the horizontal direction during an earthquake, and dampers, which reduce
the displacements and absorb the seismic input energy. Above this level is the normal
building (superstructure).
The major supplements for seismically isolated structures after publication of the third
edition to the present are the revision of the Building Standards Law (amended 1998,
enforced in 2000), and the Great East Japan Earthquake (2011).
In the revision of the Building Standards Law, notifications and recommendations for
seismically isolated structures were established, and Ministerial approval was required for
seismic isolation devices. In the notifications, the term “seismic isolation material” is used.
However, in these current Recommendations, the term “seismic isolation device” is used,
because it is considered that isolators and dampers, which are the most important constituents
of seismically isolated structures, should be dealt with in the same way as ordinary structural
members. Various types of seismic isolation devices are currently on the market, and their
performance has improved. The development of various types of seismic isolation devices
contributes to expanding the range of application of seismically isolated buildings. On the
other hand, matters of concern are the ensured margin in the energy absorption capacity of
seismic isolation devices due to long-period, long-duration seismic motions, and in the large
response deformations due to pulse waves near faults. It is necessary for seismic isolation
devices to be used only after their ultimate performance has been sufficiently confirmed.
In addition to time history response analyses discussed in the first edition, the
Recommendations described response prediction methods based on the energy method as
seismic response prediction methods for seismically isolated buildings. In this revised edition,
a response prediction method by equivalent linearization is introduced, in addition to the
energy method. Although many seismically isolated buildings exhibited a sufficient
effectiveness of seismic isolation in the Great East Japan Earthquake, problems were
reported in some seismic isolation expansion joints and hysteretic dampers. Furthermore, the
damage caused by the tsunami in that disaster was enormous. The design of seismically
isolated structures against tsunamis is a task for future investigation.
Seismically isolated structures exhibit sufficient effectiveness of seismic isolation with

respect to the predicted seismic motions. Seismically isolated structures are simple structural
systems, and the effect of the properties of seismic motions on the response of such structures
is large. The seismic forces acting on seismically isolated structures are uncertain, so in the
design of seismically isolated structures it is essential to provide an appropriate margin. It is
desirable that the response of a superstructure does not become excessive; that is, the
@Seismicisolation
@Seismicisolation
response to the seismic motions does not exceed those predicted. On the other hand, for small
seismic motion input and strong winds, it is desirable that the oscillations of seismically
isolated buildings are within an allowable range.
Seismic design starts by setting the input seismic motions, but at present it is almost
impossible to accurately predict the seismic motions at a site where a building is to be
constructed. However, structural engineers have carried out designs under these uncertain
conditions. This edition was produced with an emphasis on providing objective
decision-making material to support the decisions of structural engineers. The main points
that have been strengthened in comparison with the previous editions are as follows:
・Introduction of the most recent knowledge regarding the properties of seismic isolation
devices
・Introduction of new knowledge regarding input seismic motions, such as long-period,
long-duration seismic motions.
・Evaluation of the response of super high-rise buildings to strong winds
・Seismic isolation effects based on the seismic observation records from the Great East
Japan Earthquake (not contained in the English edition)
・ Upgrades of examples of seismically isolated buildings (not contained in the English
edition)
In this edition, incorporating new knowledge as much as possible was prioritized, so in some
cases duplication of the contents of the previous edition was eliminated. This edition should
be considered to be a supplement to the previous edition.
This edition consists of Part I “Design Recommendations” and Part II “Commentaries.” The
Design Recommendations describes the properties of seismically isolated structures, the
basic requirements of design, and the basic concepts of design. The Commentaries include
explanations and material to deepen understanding of the Recommendations. No single
approach can achieve all design targets, so various options are available. The independent
actions of the structural engineer must be to comprehensively assess the options and to
produce a single design.
From this point of view, the Design Recommendations are as objective as possible, and the
methods of realizing a specific design are contained in the Commentaries, so that it is
possible to appropriately add or modify the contents in accordance with the developments of
future technology. Chapters 1 and 3 of the Commentaries section have been written so that
they can be easily read by designers other than structural design specialists.
October 2013
@Seismicisolation
@Seismicisolation
Members of Committees Concerned
Research Committee on Structures
Chairman Mitsumasa Midorikawa

Secretaries Kenichi Kato, Hitoshi Shiohara, Izuru Takewaki
Members (Omitted)
Managing Committee on Earthquake and Structural Dynamics
Chairman Yuji Miyamoto

Secretaries Susumu Ohno, Kenichi Kato, Ichiro Nagashima
Members (Omitted)
Sub Committee for Seismically Isolated Structures
Chairman Masaru Kikuchi

Secretaries Masahito Kobayashi, Akihiro Kondo
Members Tetsushi Inubushi, Nobuyuki Ogino, Hideaki Kato, Ryu Shimamoto,
Mineo Takayama, Junji Toyama, Norio Nakanishi, Ippei Hata,
Hiroki Hamaguchi, Yasuyoshi Hitomi, Akira Fukukita, Goro Miwada
Working Group for English Edition of the Design Recommendations
Chairman Mineo Takayama

Secretaries Masahito Kobayashi, Ippei Hata
Members Yasushi Ichikawa, Nagahide Kani, Ryotaro Kurosawa,
Shigenobu Suzuki, Masanori Tasaka, Toru Takeuchi,
Yasuo Tsuyuki, Akitsugu Muramatsu
@Seismicisolation
@Seismicisolation
Contents
Part I Design Recommendations for Seismically Isolated Buildings
Chapter 1 Outline of Seismically Isolated Buildings 1

1.1 Characteristics of Seismically Isolated Structures 1
1.2 Outline of Seismic Isolation Devices 3
1.3 Applicability of Seismically Isolated Structures 4
Chapter 2 Design – General 5

2.1 Seismic Isolation Performance Considerations 5
2.2 Basic Scheme 6
2.3 Arrangement and Selection of Seismic Isolation Devices 7
2.4 Building Equipment Design 9
2.5 Maintenance Management 9
Chapter 3 Structural Design 10

3.1 Basic Considerations 10
3.2 Design Input Earthquake Ground Motions 10
3.3 Design Wind Load 11
3.4 Design Policy 12
3.5 Prediction of Seismic Response 12
3.6 Response Analysis Models 14
3.7 Ensuring the Seismic Margin 15
3.8 Points of Consideration Regarding Seismically Isolated Structures 15
@Seismicisolation
@Seismicisolation
Part II Commentaries
Chapter 1 Summary of Seismically Isolated Buildings 17

1.1 Seismically Isolated Buildings in the 21st Century
The Development of Seismic Design and the Flexible Versus Stiff Controversy 17
1.2 Fundamentals of a Seismically Isolated Building 23
1.3 Applicability of Seismically Isolated Structures 28
Chapter 2 Seismic Isolation Devices 30

2.1 Types of Devices and Their Function 30
2.2 Mechanical Properties of Each Device 34
2.3 Combination and Selection of Seismic Isolation Devices 74
2.4 Seismic Isolation Devices and Systems for Detached Housing 78
2.5 Design and Construction for Installation of Devices and for Fire Resistant Covering 81
2.6 Quality and Durability of Seismic Isolation Devices 85
Chapter 3 Design – General 90

3.1 History of Seismically Isolated Buildings 90
3.2 Effectiveness of Seismically Isolated Buildings 92
3.3 Points to Note Regarding Planning and Design 94
3.4 Points to Note Regarding Construction Planning 98
3.5 Quality Control of Seismic Isolation Devices 100
3.6 Maintenance of the Seismic Isolation Level and Seismic Isolation Devices 102

4.1 Design Targets 103
4.2 Input Ground Motions 116
4.3 Seismic Response Prediction by Simplified Analysis Methods 147
4.4 Response Prediction by Time History Response Analysis 176
4.5 Wind Design 200
4.6 Points to Note Regarding Special Seismically Isolated Structures 210
@Seismicisolation
@Seismicisolation
Part I Design Recommendations
@Seismicisolation
@Seismicisolation
@Seismicisolation
@Seismicisolation
Part I Design Recommendations for Seismically Isolated Buildings
Chapter 1 Outline of Seismically Isolated Buildings
1.1 Characteristics of Seismically Isolated Structures
(1) Principle of seismically isolated structures
The scale, mass, strength, and stiffness of a building are vastly greater than the size,
mass, and flexibility of people. The day-to-day life of people is protected within the
stable building space created by that mass and stiffness. It is most important that these
structures can endure the gravity loads that constantly act on them, and it is also
important that they can withstand external disturbances such as earthquakes and
typhoons, although their frequency is less.
Almost all earthquake damage is caused by the collapse of structural members due to
oscillations caused by ground motions. In this situation, the structure cannot support its
own weight, and the space in which people are living is crushed. The first principle of
seismic design is that a building supports its own weight, and the second principle is the
earthquake input energy is absorbed by the strength and the ductility of the structure.
Conventional seismic design emphasizes protecting the function of the building, the
asset value, and the safety of people against moderate earthquake ground motions, and
preventing the collapse of the structure against severe earthquake ground motions in
order to protect human life. In the case of preventing collapse, the structure itself is
responsible for maintaining the mechanisms of vertical support and absorbing the
seismic input energy, and so cracking of the concrete and yielding of the steel are
permitted. When the accelerations, velocities, and inter-story drifts of a building are
relatively large, collapse may occur, and residual deformation may be observed after the
earthquake, and then generally it is difficult to restore the functions of the building, and
the repair cost is rather large.
A seismically isolated structure is a structure in which a seismic isolation level is

provided at the foundation or at an intermediate story of the building. The structure
above the seismic isolation level is almost completely isolated from the horizontal
component of the earthquake ground motions. The seismic isolation level incorporates
seismic isolation devices, such as isolators that are capable of following large horizontal
deformation while stably supporting the weight of the building, and dampers that absorb
the seismic input energy. This structural method is capable of complying with the two
principles of seismic design stated above. For large earthquake ground motions, the
horizontal force acting on the superstructure is sufficiently small compared with that of
ordinary structures, and therefore elastic design can be easily carried out for the
horizontal forces on the superstructure. As a result, the building function is maintained
during large earthquake ground motions, the asset value is not reduced, and it is possible
to continue to protect the safety of the occupants.
@Seismicisolation
@Seismicisolation
2 Part Ⅰ Design Recommendations for Seismically Isolated Buildings
(2) Performance and effect of seismic isolation in structures
Seismic measurement records have been obtained for many seismically isolated
buildings, such as the University of Southern California Hospital in the Northridge
Earthquake (1994), the WEST Building in the Southern Hyogo Prefecture Earthquake
(1995), the Suisen residence in the Niigata Prefecture Chuetsu Earthquake (2004), and
several buildings in the Great East Japan Earthquake (2011). From these records it has
been verified that the horizontal acceleration produced in the superstructures were in the
range of 1/3 to 1/4 or less of the input acceleration. In a conventional structure, the
response acceleration of the superstructure is 2 to 3 times the acceleration at the ground
surface, and so the internal oscillations of a seismically isolated building are almost an
order of magnitude smaller than those of an ordinary building. The vertical motions
produced in the superstructure of a seismically isolated structure are slightly amplified,
similar to the motions of an ordinary building. But, by reducing the horizontal motions
sufficiently, the safety and feeling of security inside the building is greatly increased. In
conventional seismic design, the aim is that a structure will not collapse even though it
has been subject to large deformation, and so increasing the ductility of the structure is
emphasized. In addition, the aim is to distribute the plastic deformation throughout the
whole building, so that the seismic input energy is absorbed by the whole building.
However, the oscillations of the structure depend on the characteristics of the earthquake
ground motions; hence, it is difficult to uniformly distribute the plastic deformation
throughout all stories. Larger damage is frequently concentrated in a particular story, and
it is difficult to know this in advance. In contrast, a seismically isolated structure is
designed by assuming that the deformation during an earthquake is concentrated in the
seismic isolation level, and so the performance of these structures is more predictable.
As a result of many years of research, it is now possible to increase the deformation

capacity of a framed structure. However, because external walls, internal walls, and
other non-structural components are normally made from brittle materials, the design in
ordinary structures has involved increasing the width of the joints between panels and
providing slack in connections. In the case of superstructures, even when they are
subject to large earthquake ground motions, the inter-story drift that occurs can be
sufficiently reduced. Therefore, non-structural members can be designed and constructed
without considering inter-story drift due to earthquake ground motions, except in special
cases.
(3) Design of seismically isolated structures
Design, including architectural design, is a creative activity. It is not merely selecting a

member from a menu in accordance with a prescribed procedure. Likewise, the design
of seismically isolated structures is creative. It is important for the structural engineer to
have in-depth consideration based on his or her own imagination, as well as physics and
engineering. Seismically isolated structures should be designed by thoroughly
determining their behavior during an earthquake by referring to design guidelines,
earlier designs developed by others, damage during past earthquakes, numerical
analysis, and the results of structural experiments. Several topics remaining unknown,
however, such as potential earthquake ground motions that will occur and the behavior
of seismically isolated structures at that time, so efforts should be made to understand
these factors as much as possible.
@Seismicisolation
@Seismicisolation
1.2 Outline of Seismic Isolation Devices
(1) Characteristics of devices and performance evaluation
In a conventional frame structure, members such as columns and beams are rigidly
connected, so the mechanical behavior of one affects the other. Therefore, even if test
results are obtained for members that have been separately extracted, it is not a simple
task to derive the behavior of the structure as a whole from these test results. The
deformation of seismic isolation devices during an earthquake is almost pure horizontal
deformation; in this case, if each individual seismic isolation device was extracted and
tested, it would be comparatively easy to determine the characteristics of the structure as
a whole by combining those results.
For elastic test specimens, the similarity rule with respect to size is valid. However, the
ultimate failure phenomena are affected by the test specimen size, so tests should be
carried out on full-scale test specimens as much as possible. It is possible to carry out
individual tests on seismic isolation devices of actual size under the conditions of actual
environment, actual displacement, and actual velocity. This is the advantage of
seismically isolated structural systems.
(2) Isolator
The isolator is required to stably support the mass of the building over a long period of
time. Even when horizontal deformation occurs during an earthquake, the isolator must
stably support the vertical load and minimize any settlement in the vertical direction
associated with the horizontal deformation. A clear relation exists between the horizontal
force acting on the isolator and the horizontal displacement, but it should be verified that
the relation used in design would be exhibited in the actual situation. For the design, an
allowable horizontal deformation is indicated for the variation in stiffness of the isolator
and for each compression load or tension load (tension deformation) acting on the
isolator, so the fluctuations in stiffness should be clearly taken into account during the
design.
(3) Dampers
When a seismically isolated structure is subjected to earthquake ground motions, the

input energy is accumulated in the elastic strain energy of the isolator and is absorbed by
the plastic deformation or the frictional resistance in the damper. The role of the damper
is to reduce the response displacement produced when subjected to earthquake ground
motions and to ultimately absorb all the input energy. The relation between the force
acting on the damper and the deformation should be correctly evaluated to ensure that
the relation corresponds to the actual behavior. Also, when it is anticipated that the
damper will be subjected to repeated deformation over a long period of time, it is
required that the damper has durability in repeated loading and can ensure a sufficient
margin in energy absorption capacity.
@Seismicisolation
@Seismicisolation
1.3 Applicability of Seismically Isolated Structures
Seismically isolated structures can be applied to almost all buildings, ranging in scale
from detached houses to high-rise buildings with several tens of stories, and from
ordinary frame structures to large-span structures, regardless of the fundamental natural
period of the superstructure, provided they are properly designed by using appropriate
seismic isolation devices. Seismic isolation is also an effective method for seismic
retrofit. In particular, it is effective for historical buildings when it is necessary to leave
the superstructure intact, or when daily life and business work must continue even
during the retrofitting work.
Marginal space (clearance) should be provided around a building with a seismically

isolated structure. The clearance should correspond to the amplitude of the oscillations
during an earthquake. Because of this, it is difficult to apply seismically isolated
structures to buildings in densely built cities with other buildings close by. However,
some methods can be applied in these situations, such as constructing an artificial
foundation using seismically isolated structures for a whole block. If these problems can
be overcome, it is possible to apply seismic isolation to most buildings without
restricting the types of buildings.
@Seismicisolation
@Seismicisolation
Chapter 2 Design - General 5
Chapter 2 Design – General
2.1 Seismic Isolation Performance Considerations
(1) Seismic safety considerations
Buildings are rarely subject to earthquakes. So, when an earthquake does occur, various
problems may emerge. The seismic damage incurred by an ordinary building structure
can be considered in several stages, such as inability to maintain the function of the
building, furniture or fixtures falling down or toppling, slight damage to the structure,
injuries to occupants, damage to the structure that cannot be repaired, loss of the asset
value of the building, loss of human life, and collapse of the building.
Seismically isolated structures aim to minimize these losses as much as possible, but
structural damage that could occur when subjected to earthquake ground motions should
be considered and ranked according to the judgment of the structural engineer. The
following eight items represent the possible damage to structural and non-structural
components. To avoid such damage, an appropriate safety margin should be provided in
the performance of the seismically isolated structure.
a) Expansion joint damage
b) Damage to piping in the seismic isolation level
c) Exceeding the deformation limit or the energy absorption capacity of the seismic
isolation device including the damper
d) Exceeding the deformation limit of the seismic isolation device (isolator)
e) Exceeding the clearance limit (impact with an adjacent structure, pavement or wall)
f) Plastic deformation of the superstructure
g) Losing the vertical load bearing capacity of the seismic isolation device
h) Plastic failure of the foundation structure, and losing the vertical load bearing
capacity
(2) Evaluation of seismic isolation performance and earthquake ground motions for design
The overall intended performance of seismically isolated structures is to ensure safety

during a major earthquake. At the same time, the performance is evaluated from several
points of view, such as the reduction in response acceleration in the occupied parts
during an earthquake, the extent to which the function is maintained after the
earthquake, and the reduction in the horizontal force acting on the superstructure and the
inter-story drift.
If the earthquake ground motions used in the design are set too high, the number of
dampers in the seismically isolated structure will be excessive. So, even though safety
will be increased in a severe earthquake, the effectiveness of a reduction in the response
@Seismicisolation
@Seismicisolation
acceleration, the effectiveness of a reduction in the horizontal force acting on the

superstructure and the inter-story drift when subject to a moderate earthquake will be
deteriorated. Therefore, the seismic isolation performance is evaluated both for severe
and moderate earthquake ground motions.
When a structure is subject to large earthquake ground motions that exceed the design
earthquake, the displacement response depends on the characteristics of the seismic
isolation level and the earthquake ground motions. Therefore, a comprehensive
judgment should be formed when setting the earthquake ground motions for the design
earthquake and the target performance.
2.2 Basic Scheme
(1) Conditions for the superstructure
In a conventional seismic-resistant structure, plastic deformation is allowed in the frame

itself and energy is expected to be absorbed during an earthquake, so inevitably a certain
amount of damage is allowed in the structural frame. In a seismically isolated structure,
the deformation is concentrated in the seismic isolation level and the seismic isolation
devices absorb almost all the seismic input energy, so the superstructure does not need to
absorb energy. Therefore, no damage develops in the superstructure during an
earthquake. The necessary condition for design is that the responses of the
superstructure are limited to the elastic range for the horizontal forces produced during
an earthquake. If establishing this condition proves to be difficult, it is necessary that the
superstructure exhibits sufficient strength and stiffness so that excessive plastic
deformation is not produced. Therefore, the design to ensure ductility, which is carried
out in the seismic design of an ordinary structure, can be simplified in the case of a
seismically isolated structure. By utilizing these characteristics, it is possible to reduce
the inter-story drift during an earthquake and to increase the degrees of freedom for the
purposes of the design plan, materials used, and architectural detailing.
(2) Ground conditions
When constructing ordinary buildings, an important point is to select a construction site

with good ground conditions. If the ground conditions are not good, piles or other
effective foundations as well as ground improvement are used for foundation
construction. However, when constructing a seismically isolated structure, it is important
to prevent differential settlement. Problems such as damage to the foundations or to the
supports for the weight of the building. must not occur due to vibration during an
earthquake.
A seismically isolated structure is isolated from the short period seismic inputs but
responds to the longer period seismic inputs. The challenging components are the longer
period components that are 2 seconds up to about 10 seconds. However, earthquake
ground motions in this period range are amplified depending on the condition of the
ground, from several hundred meters below ground up to the ground surface level. So, it
is necessary to know the properties deep underground as well as near the surface.
Because the vibration characteristics of seismically isolated structures are extremely
nonlinear with large damping, the resonant response may not be induced by the
earthquake ground motions at a certain period. However, for earthquake ground motions
@Seismicisolation
@Seismicisolation
with long period components, the vibration amplitude of the seismic isolation level
could become excessive.
Earthquake ground motions that occur near faults produce large response deformation in
seismically isolated structures. For faults and their activity near a site, great attention
should be paid to the past occurrences of earthquakes, the latest information collected,
and detailed site investigations conducted.
(3) Location of the seismic isolation level
When designing a seismically isolated structure with a basement, two methods can be
used to provide the seismic isolation level: (1) a method in which the whole building
including the basement layers is supported by seismic isolation devices (base isolation),
and (2) a method in which the seismic isolation level is provided between the above
ground layers and the underground layers (intermediate-level isolation). In the case of
method (1), retaining walls around the building should be provided. In the case of the
(2), large relative displacement is produced between the basement portion and the
portion that includes the first floor and higher, so the design should take these into
consideration.
In the case of intermediate level seismic isolation, the seismic isolation level is provided
at some story of the building. Therefore, attention should be paid to amplification of the
input earthquake ground motions due to the structure below the seismic isolation level as
well as the response values of the lower structure. The design of elevators, stairs, and
piping should consider the relatively large deformation at the intermediate seismic
isolation level, and fireproof coating should be applied to the seismic isolation devices
(in particular, the isolators).
(4) Site clearance
Unlike ordinary buildings during an earthquake, a displacement of several tens of

centimeters is produced at the seismic isolation level at the base of the building.
Therefore, sufficient clearance should be provided around the seismically isolated
building.
2.3 Arrangement and Selection of Seismic Isolation Devices
(1) Arrangement of seismic isolation devices
The properties of seismic isolation devices include bearing capacity, deformation

capacity, stiffness, and energy absorption capacity. It is necessary to arrange these
seismic isolation devices in a well-balanced manner with respect to the total weight of
the superstructure and the vertical load acting on the base of each column. During an
earthquake, an additional axial load acts on the isolators due to the overturning moment
and the vertical earthquake ground motions. The seismic isolation devices should be
selected and arranged so that the variation in properties due to the additional axial load
is small. To reduce the torsional deformation of the seismic isolation level, the devices
should be arranged so that the center of mass of the building and the center of stiffness
of the seismic isolation level coincide over the full deformation range.
@Seismicisolation
@Seismicisolation
In an ordinary structure, by using structural members having a certain design margin, the
damage of the overall structure may be small. On the other hand, in a seismically
isolated structure, the deformation during an earthquake is concentrated in the seismic
isolation level. Because the seismic input energy is intentionally absorbed in this
portion, it is necessary that the seismic isolation level has the mechanical properties,
mechanical performance, and deformation capacity margin set in the design.
(2) Selection of seismic isolation devices
The seismic isolation devices are important structural members for maintaining the
safety of the whole building. It is therefore necessary that they absorb the seismic input
energy while stably supporting the weight of the building, even when subjected to large
horizontal deformation during an earthquake. Seismic isolation devices that will
properly maintain their properties and performance should be designed and selected.
Although about 30 years have passed since the first seismically isolated structures were
constructed, eventually the JIS (K 6410) standard for laminated rubber bearings was
established, and it was mandated that all seismic isolation devices obtain Ministerial
Approval. Seismic isolation devices are generally selected from approved products or
from manufacturers’ catalogs, but the engineer should independently check their
mechanical performance and durability.
(3) Ensuring performance of seismic isolation devices
It is necessary that the performance of seismic isolation devices associated with the
design and construction of seismically isolated structures is correctly evaluated.
Therefore, the evaluation tests should apply the actual displacement and velocity
produced during a potential earthquake under the conditions of actual size, actual
environment, and actual load. The seismic isolation devices used in large-scale
seismically isolated buildings have inevitably becomes larger. It is desirable that
full-size seismic isolation devices are checked by using the actual displacements and
actual velocities. When full-size devices under actual conditions cannot be checked,
various dependencies and ultimate performance should be checked by using
appropriately reduced scale test specimens.
Two types of performance evaluation tests for seismic isolation devices can be used. The
first test determines the behavior up to ultimate failure, with the objective of obtaining
the necessary design information (property evaluation tests, device performance
verification). The second test confirms the quality and performance of seismic isolation
devices actually used in a building (product performance verification). The mechanical
performance of seismic isolation devices is determined based on the materials and heat
treatments used in their fabrication. Therefore, the performance of a complete product
should be tested. As a rule, tests for all products should be carried out until appropriate
testing equipment and testing methods are established.
@Seismicisolation
@Seismicisolation
2.4 Building Equipment Design
Like ordinary buildings, seismic isolation buildings are supplied with electricity, water,
sewage systems, and gas, which are connected to the outside. In some cases hot water,
steam, and chemical substances are also supplied. It is necessary to design and construct
piping with appropriate allowances for deformation in accordance with the importance
of the building equipment. The allowances correspond to the large relative horizontal
displacement produced in the seismic isolation level during an earthquake. When the
piping contains liquid, vibrations of the pipes between supports during an earthquake
can become problematic.
It is presumed that the seismic performance of seismically isolated structures is higher

than that of the surrounding lifelines. If the building performance is required to be
continuously maintained after a major earthquake, preparations should be made so that
the function can be independently maintained for several days without dependence on
the lifelines.
It is required that the expansion joints provided between the seismic isolation level and
the ground surface, or between the seismic isolation level and the seismic-resistant
building are capable of following the horizontal motions in two directions and the
vertical motions of the seismically isolated building. In order that the expansion joints
perform correctly, it is essential that the performance is verified numerically and/or
experimentally in advance, and that appropriate construction methods are applied.
2.5 Maintenance Management
Not only buildings but also all structures require continuous maintenance management
in order to maintain their initial performance over a long period of time. When designing
a building, the materials incorporated, the structural members, and the overall structure
must be used in such a way that they do not require elaborate maintenance. To preclude
special maintenance management, isolators and dampers must have durability that is
equal to or higher than that of ordinary building members. It is also necessary that the
overall seismically isolated structure is designed and constructed by using these
materials and members so that the same level of maintenance management as that of an
ordinary building is needed.
However, because the safety of building structures during an earthquake depends on the
performance of the isolators and dampers and the large horizontal displacement
produced in the seismic isolation level, the building owner should be informed of the
necessity to perform basic visual inspections, periodic inspections, and post-earthquake
inspections.
@Seismicisolation
@Seismicisolation
Chapter 3 Structural Design
3.1 Basic Considerations
The fundamental goals when designing seismically isolated buildings are to minimize
the seismic forces acting on the superstructure and to maintain the deformation of the
isolators within the allowable limits.
In the case of low- and medium-rise seismically isolated buildings, the amount of
energy to be absorbed by the superstructure is negligible compared with that of the
seismic isolation level. Therefore, if the superstructure can resist the seismic forces
corresponding to its rigid body behavior, and if the superstructure is free from various
restrictive conditions, it does not need the energy absorption capacity that must be
ensured in an ordinary building. Also, the adverse effect of torsional vibrations
produced by the mass and stiffness eccentricity of the superstructure is basically
prevented by eliminating the torsional source of the seismic isolation level. Therefore,
compared with an ordinary building, the design freedom in a seismically isolated
building is significantly greater.
In an ordinary building, the seismic load is dominant, so excessive load bearing capacity
is provided with respect to loads other than seismic loads. However, in a seismically
isolated building, the seismic loads are reduced and the wind load becomes dominant,
and so the design margin with respect to long-term loads is reduced. This means that in
a seismically isolated building, more careful consideration should be given to the design
of loads other than seismic loads to ensure construction accuracy.
The input earthquake ground motions in the design are typically determined by the
engineer, with consideration of the importance of the building, and the social
requirements.
3.2 Design Input Earthquake Ground Motions
Accurately evaluating a future earthquake that could have a major impact on a building
and accurately predicting the resulting earthquake ground motions at the site are not
easy tasks at the present time, so setting the design input earthquake ground motions for
a building is a major problem, regardless of whether the building is seismically isolated.
Recognizing the above, it is desirable that the design input earthquake ground motions
for a seismically isolated building are appropriately set in accordance with the required
function and importance of each individual building. This is accomplished by
incorporating seismology and seismic engineering knowledge, while paying attention to
the fact that the period of seismically isolated buildings are in the rather long period
range of earthquake ground motions.
The properties of earthquake ground motions are maximum acceleration, maximum

velocity, spectrum properties, phase properties, duration, time domain envelope, and
non-stationary properties of the spectrum. Some of the factors that affect these
properties of the input earthquake motions acting on a building include the following:
a) Seismic activity of construction site and surrounding area
@Seismicisolation
@Seismicisolation
b) Source properties
c) Properties of propagation path from source
d) Properties of deep ground
e) Properties of surface strata and ground
By appropriately evaluating the effect of these factors on each individual site, the input
earthquake ground motions can be determined. The period of a seismically isolated
building is relatively long, so when setting the input earthquake ground motions, special
considerations should be given to the spectral properties in the long period range, the
total input energy, and the phase properties.
Consequently, many assumptions and a certain amount of judgment must inevitably be

made when setting the design input earthquake ground motions. After sufficiently taking
into consideration the inherent uncertainties, the engineer should set the design input
earthquake ground motions in accordance with the function and importance of the
building, and also should pay attention to the seismic margin of the building.
Only then can the engineer set the design earthquake ground motions, the design
response spectrum, and the design energy spectrum.
Actual earthquake records and simulated earthquake ground motions that are
compatible with the design response spectrum or the design energy spectrum level can
be used as the ground motions in the time history response analysis. To appropriately
evaluate the effect of the phase properties and the non-stationary properties of the
ground motions on the seismic response, it is desirable to use several types of ground
motions.
3.3 Design Wind Load
The wind loads for the building are determined by taking into consideration the shape of
the building, the structural properties, and the conditions around the location of the
building. The wind load coefficients used for determining the wind loads may be
indicated in “Recommendations for Loads on Buildings” (AIJ publication, 2004), public
notifications of the Ministry of Land, Infrastructure and Transport, or results obtained
from wind tunnel tests.
The following is a summary of the characteristics of the external wind disturbances

acting on a building.
a) External wind disturbances have an average component as well as a fluctuating

component.
b) The predominant period of the fluctuating component of external wind disturbances

is longer than that of external seismic disturbances.
@Seismicisolation
@Seismicisolation
c) External wind disturbances may generate aerodynamically unstable vibrations on

the structure.
d) The duration of strong external wind disturbances such as storms can last several
hours or more for the passage of one typhoon.
The higher the building is, the stronger the effect of external wind disturbances.
Recently, many high-rise seismically isolated buildings have been constructed, so it is
becoming indispensable to investigate their wind resistance.
3.4 Design Policy
(1) The seismically isolated building under consideration consists of a lower structure, a
seismic isolation level that includes isolators and dampers, and a superstructure.
(2) The structural design for loads and external forces other than seismic loads should be
carried out in accordance with the methods used for ordinary buildings. However, in
some cases, wind loading on high-rise seismically isolated buildings and detached
seismically isolated houses can produce large displacements in the seismic isolation
level, so the effect of wind loading on the seismic isolation devices should be
investigated.
(3) The seismic isolation level and the superstructure should be designed based on various
analyses on the response of the building subject to the design seismic input in the
horizontal direction. These analyses can be carried out independently for two horizontal
directions of the building.
(4) The isolators and dampers should be arranged so that torsion is not produced in the
seismic isolation level when the superstructure is considered to be a rigid body.
(5) The maximum displacement and the maximum axial load on the isolator, and the
required energy absorption of the dampers should be confirmed, as obtained from
response analysis using the design seismic input, and they should be within their
respective allowable values.
(6) For the superstructure, an allowable stress design may be performed under the design
seismic loads.
(7) The lower structure should be designed in the same way as that for an ordinary building
to ensure that excessive settlement, inclination, and horizontal movement do not occur.
In particular, if differential settlement occurs, horizontal displacement will be produced
in the seismic isolation level due to the low horizontal stiffness of the level. Also, it is a
possibility that the axial load in the isolators will fluctuate and the deformation capacity
will be reduced. Therefore, careful design is essential.
3.5 Prediction of Seismic Response
(1) Prediction of the seismic response of seismically isolated buildings due to horizontal
seismic input is carried out by a time history response analysis (see (2)) or by a
simplified analysis method (see (3)).
@Seismicisolation
@Seismicisolation
(2) An appropriate vibration model among the models described in section 3.6 should be
used for performing a time history response analysis. In this case, it is assumed that a
certain number of earthquake ground motions are used.
(3) The seismic response prediction by using a simplified analysis method assumes that the
stiffness of the superstructure of a seismically isolated building is normally extremely
large compared with the stiffness of the seismic isolation level. The results of the
seismic response analysis when the analysis model is a superstructure that is assumed to
be a single concentrated mass are considered in order to obtain a comprehensive
response prediction. The following methods can be used for performing this analysis.
a) Response prediction based on energy balance
With this method, the maximum displacement of the seismic isolation level is
obtained by assuming the seismic energy input is concentrated in the seismic
isolation level. Depending on the properties of the seismic isolation devices, the
method may assume dampers with perfectly elastic-plastic restoring force
characteristics and perfectly elastic isolators, or dampers with a viscous damping
mechanism.
b) Response spectrum method
The building, modeled as a nonlinear single concentrated mass model, is converted

into an equivalent linear single concentrated mass system having an equivalent
period obtained from the equivalent stiffness at the predicted maximum
displacement amplitude of the seismic isolation level and the equivalent viscous
damping coefficient per cycle. Based on the acceleration response spectrum of the
equivalent linear system, the maximum displacement of the seismic isolation level
is obtained.
When a simple analysis method is used, the response of the superstructure can be
determined based on the shear force coefficient distribution derived from the
structural characteristics of the superstructure and the seismic isolation level.
(4) Seismic isolation level of seismically isolated structure has a very low horizontal
stiffness as compared to that of the superstructure. The translational first mode in
horizontal direction appears dominantly in which the superstructure behaves like a rigid
body. Frequency of the first mode of isolated structure is very low. Generally, dynamic
amplification effect in the low frequency region is small. Torsional first mode has the
same aspect as that of translational first mode. Amplification of the torsional vibration
in the seismically isolated structure is also very small. There is low possibility of a large
dynamic amplification of torsional vibrations induced by a deviation between the center
of stiffness and the center of mass of the superstructure. Therefore the stresses in each
part of the frame of the superstructure can be calculated using seismic forces in
accordance with the seismic shear force, obtained from a static analysis taking the
torsional deformation of the building into consideration.
@Seismicisolation
@Seismicisolation
3.6 Response Analysis Models
(1) Response analysis models for seismically isolated structures can be set in several levels
ranging from simple to detailed models, depending on the purpose.
a) Single concentrated mass model
The superstructure is modeled as a rigid body. The response of the seismic isolation
level is obtained when subjected to horizontal earthquake ground motions in one
direction.
b) Multiple concentrated mass model
In this model, the seismic isolation level is the first story, and each story of the
superstructure is modeled as a concentrated mass. Both the behavior of the seismic
isolation level and the response of each part of the superstructure are obtained.
c) Torsional vibration model
This model assumes rigid floors, which account for horizontal displacements in two
directions and torsional displacements as the components of displacement of each
floor. The torsional vibrations and the response to horizontal inputs in two
directions are obtained.
d) Three-dimensional model
In this model, the whole building is a three-dimensional frame structure and the
vertical motions of each column are taken into consideration. The response
including the effects of irregular arrangements of seismic shear walls and uplift on
isolators can be obtained.
e) Vertical motion and uplift model
Each story of the superstructure is modeled as a concentrated mass with flexual and
shear springs, shear springs and axial springs that are provided at the position of the
isolation devices at the lowest level. By connecting these devices with rigid links,
the overturning moment and the fluctuations of the axial load on the isolators
produced by the vertical motions can be obtained. This model is capable of
reproducing the behavior of a seismically isolated building by recognizing that the
tension forces generated in the isolators are simpler than those in the
three-dimensional model.
(2) The seismic isolation devices (isolators and dampers) are normally modeled as a single
spring having their own hysteresis properties. However, to accurately reproduce the
behavior of the isolators and dampers, it is effective to further divide each of these into
elements reflecting their mechanical properties so that they are modeled as compound
mechanisms of these elements.
@Seismicisolation
@Seismicisolation
3.7 Ensuring the Seismic Margin
(1) The response of the seismically isolated building is greatly affected by the properties of
the seismic isolation level. Because the seismic isolation level is composed of only the
seismic isolation devices, their performance can be examined by conducting various
tests and fluctuations in the results of the response analysis are reduced. At the seismic
isolation level, the seismic isolation devices only take loads, so it is necessary to note
that only the calculated margin is used. Therefore, the design should account for
variations, such as manufacturing tolerances, changes with temperature, and changes
with time. Moreover, the structural engineers should consider the amount of fluctuation
caused by undetermined factors. In contrast, in an ordinary building, the allowable
stresses are calculated by neglecting various walls and non-structural members, so that
generally the margin is not the one calculated.
(2) For setting the input earthquake ground motions, many parameters greatly govern the
response of seismically isolated buildings. A diversified investigation should sufficiently
take into consideration the amount of uncertainty of these parameters. However, the
prediction of earthquake ground motions is still being researched in seismic
engineering; few engineers at this time understand this research and can use it for
setting the design earthquake ground motions. As a result, the collapse mode of the
building with respect to input earthquake ground motions that exceed those predicted
should be checked, and a design margin including fail-safe behavior should be ensured.
3.8 Points of Consideration Regarding Seismically Isolated Structures
(1) If the superstructure and the lower structure become plastic, the energy absorption in the
seismic isolation level is reduced, and the expected seismic isolation effect cannot be
obtained. As a rule, parts other than the seismic isolation level should be designed to
remain elastic even in a large earthquake.
(2) In seismically isolated structures, evaluation of the energy absorption capability of the
superstructure during an earthquake is not important in the design. When an appropriate
design margin has been provided, the structural rules for ensuring deformation
capability can be set freely in accordance with design judgment.
(3) In buildings that are greatly affected by wind loading, such as high-rise seismically
isolated buildings and detached seismically isolated houses, the seismic isolation effect
during a major earthquake should be ensured, while also satisfying the conflicting
requirement of preventing excessive deformation of the seismic isolation level due to
strong wind loading.
(4) In some cases of seismically isolated buildings with a large aspect ratio, excessive axial
forces may be produced in the seismic isolation devices and foundations due to not only
the overturning moment caused by the horizontal forces, but also the effect of rocking
vibrations. Therefore, it is important to investigate building overturning.
(5) In long planar buildings, torsional input can be generated by phase shifts due to
differences in the transmission path of the earthquake ground motions. As a result, a
torsional response is excited in the building. Normally the torsional stiffness of the
seismic isolation level is small, so the horizontal and torsional motions are coupled to
@Seismicisolation
@Seismicisolation
produce an elastic plastic response. Even when the response torsional angle is small, a
large horizontal deformation could be produced in the seismic isolation devices in the
corners of the building. Therefore, torsional vibrations should be extensively
investigated.
(6) When designing the seismic isolation of a building with a long period, the seismic
isolation effect may be increased by setting a small design shear force (yield shear
force) of the seismic isolation level and increasing the seismic isolation period.
However, these settings increase the vibrations due to long period earthquake ground
motions and strong winds, and so the energy absorption capacity of the dampers should
be ensured.
@Seismicisolation
@Seismicisolation
Part II Commentaries
@Seismicisolation
@Seismicisolation
@Seismicisolation
@Seismicisolation
Chapter 1 Summary of Seismically Isolated Buildings
Section 1.1 Seismically Isolated Buildings in the 21st Century – The Development of
Seismic Design and the Flexible Versus Stiff Controversy
(1) The beginnings of seismic design
Seismic design as seismic technology was introduced after the 1923 Great Kanto
Earthquake. During the subsequent three-quarters of a century, seismic design
progressed rapidly. Along with widening the scope of its application, the design concepts
diversified. With the development of computers, and measurement and control
equipment, response analyses have progressed dramatically, and design methods have
diversified and become more sophisticated. On the other hand, the destruction wrought
by successive major earthquakes continually emphasize that our seismic design methods
were not yet fail-safe.
Seismic design started as a method in which the forces to resist earthquakes were added
to structural frameworks originally designed for gravity loads. As a result of repeated
structural damage wrought by earthquakes and progress in response prediction
technology, the importance of seismic design in the total design expanded, and seismic
design became dominant in structural engineering in Japan.
During this time, continual attempts were made to realize seismically isolated structures
as a method to fundamentally isolate the building from the horrors of earthquakes.
However, preventive methods had a germination period of about half a century while
seismic design became established. The preconditions for the design of seismically
isolated structures, namely, the maturity of seismic design theory to support theories as
well as the development of reliable structural devices in the form of rubber isolators,
were not established until the 1970s.
The progress of seismic design and the establishment of seismically isolated structures
have faced many obstacles; that is, each has followed a tortuous path. This is because
seismic design is not a simply application of theory. There is also a legal obligation to
protect the welfare of the public, which has been a primary factor from the start. Seismic
design has to be rationally supported by scientific principles. On the other hand, as law it
has to be sufficiently simple and clearly understood by everyone. Facts, science, and
theory are not yet unified. The possibility always exists that incorrectly understood facts
could be logically stated. This fundamental dilemma, which forms an undercurrent to
seismic design, is the source of continuous dispute.
@Seismicisolation
@Seismicisolation
18 Part Ⅱ Commentaries
This dispute is commonly referred to as the "flexible versus stiff controversy” (i.e., the
first of five). The argument started as an academic dispute in reaction to the introduction
of the seismic coefficient method and the recommendation of stiff structures
immediately after the Great Kanto Earthquake in 1923.
Seismic design introduced the seismic coefficient method immediately after the Great
Kanto Earthquake. In the seismic coefficient method, the horizontal force applied to a
building was considered to be 10% of the weight of the building, and the stress on the
building was designed to stay within the elastic region under the effect of gravity and
seismic forces. Elastic vibration theory had already been proposed by the early 20th
century, and it was generally known what forces a building was subjected to in
horizontal vibratory motions. Although introduction of the seismic coefficient method
had great significance as a seismic design method, its academic significance was small.
Thereafter, the value of the seismic coefficient in the seismic coefficient method was
doubled (20% of the weight), and it was established in the 1950 Building Standards Law,
where it has remained to this day. The seismic coefficient method is excellent for its
practicality and its logic, but its scientific basis was not yet sufficient. Systematic
formulation of the seismic coefficient method was completed during the 1960s with the
introduction of nuclear power plants and the period of high economic growth in Japan,
during which time structures with high importance were designed to be stiff and strong.
(2) Progression of “flexible” versus “stiff” controversy
The period of high economic growth from the late 1960s prompted the construction of
high-rise buildings in cities, and discussion regarding the feasibility of super high-rise
buildings in an earthquake country led to the second "flexible versus stiff controversy.”
As a result, it became clear that when the natural period of a building is increased, the
intensity of the seismic forces applied to the building decreases in inverse proportion to
the natural period. This was a victory for the "flexible side" in the "flexible versus stiff
controversy.”
With the development of electronic computers and the enhancement of seismic records,
it became possible to understand more details of the behavior of buildings during an
earthquake. As a result, it became clear that the design of low- and medium-rise
buildings, which are midway between the two extremes, was not so simple from both the
force and deformation aspects. In other words, as long as a short period building remains
elastic, the horizontal forces produced in the building expressed as a seismic coefficient
could reach 1.0 (100% of the weight). It became clear that a building designed for a
horizontal force equivalent to a seismic coefficient of 0.2 could resist a major earthquake
@Seismicisolation
@Seismicisolation
only if the proper amount of plastic deformation capability was provided. The defects of
the seismic coefficient method became apparent after the Niigata Earthquake (1963), the
Earthquake off the Coast of Tokachi (1968), and the Earthquake off the Coast of Miyagi
Prefecture (1978). The theme of the third "flexible versus stiff controversy" was how to
provide the energy absorption capability in buildings.
In the 1981 amendment of the Building Standards Law, it became obligatory to provide
energy absorption capability as a condition fulfilled by the building in addition to using
the conventional seismic coefficient method. In this way, the basic framework of seismic
design was completed by ensuring the basic strength by the seismic coefficient method
and ensuring the required plastic deformation capability. However, determining the
deformation capability related to the nonlinear region beyond the elastic region is not
easy, even with the present advanced analysis technology. On the other hand, as analysis
technology advances, the number of options has increased, and it is necessary to face the
fact that no objective means can verify the validity of all results.
Furthermore, compared with the initial scheme of simply adding seismic resistance to
structural frameworks that were originally designed for gravity loads, the design of
structural frameworks considering the absorption of energy by plastic deformation was
very complex. As a result, in the design of structural frameworks, it seems that seismic
resistance is the main load and gravity is secondary.
The main theme of the third "flexible versus stiff controversy" was the very practical
problem of how to incorporate a building's energy absorption capability in the plastic
region into seismic design. As seismic design was becoming more complex, the
application of seismically isolated structures as a means of fundamentally mitigating the
destruction caused by earthquakes was investigated in earnest in Japan. In
seismic-resistant structures, the structural framework is expected to have strength and
plastic deformation capability by the use of seismic-resistant elements. Therefore, in an
earthquake, some kind of damage would be inflicted on the structural framework. In
contrast, proponents of seismically isolated structures asserted that the building itself
would remain undamaged. The veracity of this was not accepted in the seismic design
society, and it was impossible to guarantee seismically isolated structures since not much
positive proof was found in strongly seismic countries such as Japan. Against the
resistance of the seismic design world, the validity of seismically isolated structures was
established on the basis of the evidence from seismic design. The availability of
excellent structural elements in the form of rubber isolators was indispensable to
achieving this validity.
@Seismicisolation
@Seismicisolation
(3) Southern Hyogo Prefecture Earthquake (Kobe Earthquake)
In the 1980s, seismic design was becoming more advanced and complex. This was an
extremely important period for seismically isolated structures, which had passed through
the research stage and had reached the stage of practical application. The principle of
seismically isolated structures can be described as follows:
"Seismically isolated structures are two-layered structures, with the building portion as
the superstructure, and the seismic isolation level and the foundations as the lower layer.
The seismic isolation level consists of rubber bearing isolators that constitute flexible
horizontal springs, and of dampers that are capable of absorbing all the seismic input
energy. The superstructure is sufficiently stiff compared with the seismic isolation level,
and the superstructure consists of masses in a structural system. The whole seismically
isolated structure is configured as a lumped mass model, and can be configured as a
single mass vibrating system."
The high vertical support capacity and horizontal elastic deformation capability of
rubber bearings enable seismically isolated structures with long natural periods
exceeding 4 seconds, and so the combination of the elasticity of isolators and the
plasticity of dampers functions as a highly efficient energy absorption mechanism.
Viewed another way, seismically isolated structures provide the simplicity and clarity
required of seismic design, as well as the advantages of long period structures, and their
ability to provide energy absorption is beyond doubt. The concept of seismically isolated
structures appeared on the scene as if it were an opposing concept to seismic design, but
after the fourth "flexible versus stiff controversy," these structures were recognized in the
seismic design society as being superior. In recent years, the trend was to find structural
forms that can exhibit better performance; these structures are referred to as "vibration
control structures.” It is no exaggeration to say that the trend towards vibration control
structures was ignited by the trend of seismically isolated structures.
In any case, a fundamental difficulty in seismic design is to ensure the required strength
and energy absorption capability by using structural frameworks that support gravity,
and it is hard to provide simplicity and clarity. When the Southern Hyogo Prefecture
Earthquake (Kobe Earthquake) occurred in the early hours of January 17, 1995, the
achievement of seismic-resistant structures in Japan became clear.
The impact associated with the Kobe Earthquake was enormous. The following is a
summary of the lessons learned from the damage caused by that earthquake.
@Seismicisolation
@Seismicisolation
• Seismic design technology had not reached perfection, but its technological progress
over three-quarters of a century was clearly recognized.
• The progress of seismic design technology was offset by the expansion of the sphere
of human habitation, and it was not easy to reduce the absolute scale of earthquake
damage.
• It is not possible to prevent urban disasters by only improving the seismic design
technology of individual buildings.
After inspecting the collapsed buildings immediately after the earthquake, it was
reported that the defect of current seismic design technology was the insufficient
consideration given to vertical seismic motions. However, the cause of buildings
collapsing was damage to columns by horizontal seismic motions, such that the columns
lost their bearing capacity and released the potential energy of the gravitational field. It
has been shown that in the Kobe Earthquake, when compared with the enormous power
of the horizontal seismic motions, the effect of the vertical seismic motions was normal.
In steel structures, the obvious failure mode causing damage to the members was
unexpected and raised a fundamental doubt regarding the current seismic design, which
considers the source of seismic performance to be energy absorption by plastic
deformation of the structure. Some of the expressed opinions on design methods were
that seismic design should return to elastic design. However, the many full-scale
structural tests, carried out during the dawn of Japanese super-high-rise buildings in the
early 1960s, showed that steel structures did not easily fail and separate, and brittle
fractures were rare. The cause of this damage was the commonplace use of steel
materials under severe conditions due to the priority placed by society on economics and
labor saving. Furthermore, it was accepted that the structures had sufficient resistance as
seismic elements, but because the seismic motions were extremely large, the structures
could not withstand the loads and subsequently collapsed. An ambiguous expectation
among specialists was that if the design of seismic forces conformed to the current
Building Standards Law, buildings would be safe even in major earthquakes, since actual
buildings have varying amounts of reserve capacity. We fully realized that the
uncertainty of seismic motions must be squarely faced. The strength of the seismic
motions near the epicenter greatly exceeded the standard level projected in the Building
Standards Law, depending on the location. Furthermore, the damage to various
structures confirmed that the road to perfection of the seismic design was still long.
(4) The fifth "flexible versus stiff controversy"
@Seismicisolation
@Seismicisolation
While it is true that seismic design technology was progressing at a certain speed, the
speed of expansion of the sphere of human habitation was faster, and consequently
technology appeared to be regressing.
It is impossible to achieve disaster-resistant cities by only pursuing the seismic resistance

of individual buildings. The disaster in Kobe thoroughly showed us the various aspects
of a disaster in a modern city. However, if it is possible to completely guarantee the
seismic resistance of individual buildings, it will clearly be possible to prevent urban
disasters. It is confirmed that the improvement of seismic design technology is the
fundamental technical task.
On the occasion of the Kobe Earthquake, two seismically isolated buildings, constructed
at locations that were not very far from areas affected by the disaster, were verified to
exhibit the expected performance. The 20th century came to a close with the rise and
acceptance of seismically isolated structures in the seismic design society. However, it
seemed unrealistic to assume that in the future all buildings would be seismically
isolated. However, with verification of the superiority of seismically isolated structures,
seismic-resistant structures would be substantially improved as the number of cases of
seismically isolated structures increased. Thus, the fifth "flexible versus stiff
controversy" is developing around the performance of seismically isolated structures,
and it is a difficult battle for seismic-resistant structures. The struggle for
seismic-resistant structures is at a crucial stage.
@Seismicisolation
@Seismicisolation
Section 1.2 Fundamentals of a Seismically Isolated Building
(1) Mechanism of seismically isolated structures
Figure 1.2.1 schematically shows the most common seismically isolated structural
system. If the superstructure is sufficiently stiff compared with the seismic isolation level,
the superstructure can be treated as a single mass system model.
Figure 1.2.1 also schematically shows (1) the mechanism for supporting the
superstructure in the vertical direction, (2) the mechanism for exhibiting a restoring force
in the horizontal direction, and (3) the mechanism for absorption of energy in the relative
displacement between the superstructure and the foundations in the seismic isolation
level. These basic structural elements, referred to as the seismic isolation devices,
characterize a seismically isolated structure.
Namely, the superstructure is flexibly connected to the foundations by mechanisms (1),

(2), and (3).
Superstructure
(2)
Isolation level
(3) (1) (2) Basement (1) (3)
Fig. 1.2.1 Schematic diagram of a seismically isolated structure
Using variations of mechanisms (1), (2), and (3), various seismic isolation systems can
be produced. Furthermore, by selecting the properties of mechanisms (1), (2), and (3) in
the same system, it is possible to produce a system having different properties.
The following is a typical seismic isolation system that is currently used.
a) A rubber bearing is used for functions (1) + (2), and separate damping devices are
used for function (3).
b) Functions (1)+(2)+(3) are incorporated into a single rubber bearing into which a
lead (tin) plug is inserted, or a bearing made of highly damping rubber.
@Seismicisolation
@Seismicisolation
c) Sliding bearings are used for functions (1) + (3), and a separate spring device is used
for function (2).
The following is a specific explanation of the effect of seismic isolation when the
superstructure is flexibly connected to foundations with this type of system.
The fundamental natural period of a reinforced concrete structure is about 0.02H seconds
(H: building height [m]), so the natural period of 15 m height is about 0.3 seconds. It is
easy for a structure supported on isolators to have a primary natural period of about 3 to
4 seconds. When the primary natural period changes from 0.3 seconds to 3 to 4 seconds,
normally the size of the structural response to seismic motions is given by the relation
shown in Fig. 1.2.2. As can be seen, a major difference in the seismic forces (response
acceleration) acts on the structure and the amplitude properties (response displacement)
between the case where the structure is connected to the foundations (example: natural
period 0.3 seconds) and the case where it is supported on an isolator (example: natural
period 3 to 4 seconds). The magnitude of the response also varies depending on the
amount of damping. The change in the amount of response due to the flexible connection
is the result of the characteristics of the seismically isolated structure; in other words, the
change is the effectiveness of seismic isolation.
It is difficult to simultaneously reduce both the response acceleration and the response
displacement. It is important to appropriately set the natural period of the seismically
isolated structure and the amount of the damping in accordance with the properties of the
seismic motions used in the design.
Small
Displacement Response
Damping
Acceleration Response
Large
Small Damping Damping
Large Damping
Period(s) Period(s)
Fig. 1.2.2 Example of earthquake response spectrum
@Seismicisolation
@Seismicisolation
(2) Outline of seismic isolation devices
(a) Isolators
Normally a rubber bearing, which is the main seismic isolation device, is used as the
isolator.
A rubber bearing is a lamination of rubber and steel plates alternately stacked, as shown
in Fig. 1.2.3. The elastic modulus of rubber is low but it is virtually incompressible, so if
the rubber is in thin layers and the protrusion in the lateral direction as a result of axial
compression forces is constrained by steel plates, the amount of displacement in the
axial direction is small, and the rubber bearing has strong compression resistance. When
a horizontal force acts on this rubber bearing, it deforms due to rubber’s elastic modulus
(apparent shear deformation). The result is a bearing that is stiff in the vertical direction
and flexible in the horizontal direction.
The performance of the rubber bearing and, in particular, its horizontal deformation
capability while supporting a vertical load are important. To ensure the horizontal
deformation capability, the number of rubber layers may be increased, and then the
height of the rubber bearing may be increased, but this decreases the compression axial
load bearing capacity (buckling load), and significantly increases the dependency of the
horizontal stiffness on the axial load. In other words, by selecting the shape it is possible
to produce a rubber bearing with various properties. But, when selecting the shape, it is
important to investigate the first shape factor and the second shape factor, as described
below. The second shape factor greatly affects the stability of the rubber bearing, so it is
necessary to select it with caution.
First shape factor: the ratio of the area of one layer of rubber that is constrained by steel
plates to the free surface area (in the field of vibration prevention rubber, this is simply
referred to as the shape factor). This is related to the vertical stiffness and stability.
Second shape factor: the ratio of the rubber diameter to the total thickness of all rubber
layers (a coefficient is introduced for the design of the rubber bearing for seismic
isolation). This is related to the loading capacity and horizontal stiffness.
Flange Plate
tR : Thickness of rubber sheet (mm)
Rubber D : Diameter of rubber sheet (mm)
Steel Plate n : Number of rubber layers
First shape factor : D/(4tR)
Second shape factor : D/(ntR)
Fig. 1.2.3 Laminated rubber bearing
Rubber bearings are used in many seismically isolated buildings. Isolators other than
rubber bearings, such as sliding bearings or roller bearings, are also used. Sliding
bearings and roller bearings on their own do not have restoring force, so in most cases
they are used in conjunction with rubber bearings. In contrast to rubber bearings in
which the shearing force increases with horizontal deformation, the shear force (friction
force) in sliding bearings and roller bearings is constant after sliding. Therefore, to
@Seismicisolation
@Seismicisolation
reduce the horizontal stiffness of the seismic isolation level and to increase the natural
period of the seismically isolated building, sliding bearings or roller bearings are also
used. In sliding bearings, energy absorption due to friction (damper effect) can be used.
Furthermore, integrated devices combine the isolator function and the damper function.
A typical example is a rubber bearing in which a lead plug is inserted into the center, or
a bearing made of high damping rubber.
The restoring force properties of the separate damper type (separate function type), the
integrated type (combined function type), and the sliding bearing type have their own
individual characteristics, and their differences appear in the results of seismic response
analysis. So, it is necessary for the structural engineer to evaluate and select the devices
in accordance with the design conditions.
Isolators are stiff in the vertical direction but flexible in the horizontal direction, so
when installing them during construction, it is necessary to maintain the horizontal
accuracy.
(b) Dampers
Dampers are essential for seismically isolated structures to suppress excessive

deformation of the isolators and to absorb the seismic energy.
Types of dampers include hysteretic dampers that use the plastic deformation energy of
metals such as steel and lead, or Coulomb friction energy, and viscous dampers that use
oil, a viscous fluid or a visco-elastic material. Hysteretic dampers and viscous dampers
each have their own advantages and disadvantages, so it is also necessary for the
structural engineer to evaluate and select them in accordance with the design conditions,
such as the necessary damping capability, the allowable maximum deformation, the
environmental temperature, and the building use. To increase the torsional stiffness, it is
desirable that the dampers are placed as close as possible to the external periphery of the
plane of the seismic isolation level.
The reduction in the response shear coefficient is related to the increase in the natural
period by the isolators and the appropriate setting of the amount of damping. It is clear
from the response prediction method using the energy balance1) that an optimum yield
shear force coefficient (= ratio of yield shear force of the damper group to the total mass
of the superstructure) enables dampers to have perfect elastic-plastic restoring force
properties. Figure 1.2.4 shows the response prediction curve1) obtained from the energy
balance (see Section 4.3.1). The vertical axis in this figure is the base shear coefficient
of the seismic isolation level, and the horizontal axis is the displacement of the seismic
isolation level. In the figure, (a) is the case in which only hysteretic dampers are used,
and (b) is the case in which viscous dampers (20% of the damping factor) are added to
the hysteretic dampers. The downwardly convex curves in the figure are the response
for a specific seismic isolation period Tf, and the vertical curves are the response for a
constant damper yield shear force coefficient as. The point of intersection of the two is
the response of the seismically isolated building. In each case, although the base shear
coefficient decreases as the seismic isolation period increases, the base shear coefficient
@Seismicisolation
@Seismicisolation
increases if the amount of damping is too great or too little. An ideal amount of damping
for each seismic isolation period is seen.
Tf (sec)
2 2
0.35 0.35
Max. Shear Coefficient of Isolation level a1
Max. Shear Coefficient of Iaoltion Level a1

VE=150cm/s VE=150cm/s
0.30 0.30 Damping Factor 0.2
3
0.25 0.25
Tf (sec)
0.20 0.20 3
4
0.15 0.15
2%
5 4
0.10 0.10
6 5
12%
6
0.05 0.05
15%
10%
15%
8%
as
7%
6%
3%
4%
5%
3%
as
9%
5%
7%
0.00 0.00
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Max. Displacement δmax (cm) Max. Displacement δmax (cm)
(a) Case of hysteresis damper only (b) Case of hysteresis damper + viscous damper
(20% of damping factor)
Fig. 1.2.4 Maximum predicted deformation of isolation level
based on energy balance method
@Seismicisolation
@Seismicisolation
Section 1.3 Applicability of Seismically Isolated Structures
Seismically isolated structures can be applied to a wide range of structures ranging from
small-scale housing to super high-rise buildings. However, it is necessary to consider the
following points at the planning and design stage in order for a seismically isolated structure
to be built.
The response of a seismically isolated building is greatly affected by the predicted seismic
motions. It is therefore desirable to set the design seismic motions and the design criteria by
consulting with the building owner. In recent years, concerns have been raised over the
continuation of oscillations for a long time due to long period ground motions, as well as the
concern over the large response due to short period pulse waves measured close to the
epicenter faults.
The amount of deformation generated in the seismic isolation level in response to the
predicted seismic motions, and a way to ensure seismic isolation clearance with respect to the
maximum deformation expected (see Fig. 1.2.5), are important. If the seismic isolation
clearance is small, seismic motions that exceed the design seismic motions could have an
impact against the retaining wall or similar structure. Depending on the impact velocity, a
large force could act on the superstructure due to the shock during impact, so it is desirable to
avoid impact. Consideration should also be given to how much margin is to be provided with
respect to seismic motions that exceed the design seismic motions used in the design of
seismically isolated buildings.
Horizontal Clearance
Vertical Clearance
Fig. 1.2.5 Seismic isolation clearance
In the initial stages of application of seismically isolated structures, it was thought that the
number of stories in the superstructure was limited and that would be difficult to apply
seismically isolated structures on soft ground. Recently, however, the application of
seismically isolated structures to high-rise buildings has been expanding, and examples of
applications with aspect ratios (slenderness) of more than four can be seen. Seismically
isolated buildings are now being designed on soft ground after sufficient investigation and
study of the ground properties and input ground motions.
@Seismicisolation
@Seismicisolation
In a high-rise seismically isolated building, a tension force can act on the seismic isolation
devices, so various measures are required to keep these tension forces below the allowable
value. Furthermore, in a high-rise seismically isolated building, because the superstructure is
flexible, it cannot be regarded as a single mass and the effect of a higher-order modal
response in the seismic response must be considered. In addition, it is necessary to be careful
about the response during strong winds. In all, careful consideration is required for the design
of high-rise seismically isolated buildings.
In seismically isolated buildings whose scale is large in plan, the possibility of generation of
torsional response increases. It is known that the torsional response of a seismic isolation
level can be reduced by aligning the stiffness center of the seismic isolation level with the
center of gravity of the superstructure (reducing the eccentricity). However, even when the
torsion is small, the response deformations are larger at the corners than at the center of
gravity. It is necessary to take this into consideration when setting the horizontal clearances.
Seismically isolated structures exhibit their seismic isolation effect with respect to horizontal
seismic motions, not vertical seismic motions. In some three-dimensional seismically
isolated buildings2), a seismic isolation effect has also been expected in the vertical direction,
but this has not become commonplace. In cases where the building houses equipment that is
sensitive to vertical seismic motions, it is necessary to implement separate vibration
reduction or vibration prevention measures against vertical vibrations.
Although seismic isolation devices use materials that have sufficient durability, and the
occurrence of corrosion can be seen depending on the environment (in particular, a humid
environment) of the seismic isolation level. It is important to check whether any fault in the
seismic isolation devices has occurred after the devices have been subjected to an earthquake
or typhoon. It is necessary to install a "scribed type displacement meter" to record the
deformation of the seismic isolation level to verify the integrity (residual capacity) of the
seismically isolated building. It is essential to avoid constructing ancillaries around the
building that could obstruct the deformation and movements of the seismically isolated
building, or to overlook the deformation of the seismic isolation level when repairing piping.
To maintain the function of the seismically isolated building, it is necessary to carry out daily
inspections as well as periodic inspections.
References
1） Hiroshi Akiyama：Earthquake-Resistant Limit-State Design for Buildings, University of Tokyo Press, 1985
2） Osamu Takahashi, Hiromasa Aida, Junji Suhara, Yasuo Tsuyuki, Takashi Fujita：Development of 3Dimensional
Seismic Isolation Buiding, Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan, B-2,
pp.435-436, 2008
@Seismicisolation
@Seismicisolation
Chapter 2 Seismic Isolation Devices
Section 2.1 Types of Device and Their Function
A seismically isolated structure is a structure in which a seismic isolation level is provided

between the lower structure (foundation) and the superstructure to insulate the superstructure
from the horizontal component of seismic motions. The seismic isolation level can follow
large horizontal deformation while stably supporting the weight of the building, and
incorporates seismic isolation devices including isolators with appropriate restoring force and
dampers that absorb the seismic input energy.
As shown in Fig. 2.1.1, isolators can be broadly classified into rubber bearings, sliding
bearings, and roller bearings. A natural rubber bearing uses mainly natural rubber. A high
damping rubber bearing, which uses specially blended rubber, has an energy absorption
function. A plugged rubber bearing provides energy absorption by introducing a damping
plug such as lead, tin, iron particles and rubber, into the center of the rubber bearing.
・Natural Rubber Bearing

Rubber
Bearing ・High Damping Rubber Bearing
・Plugged Rubber Bearing
（Lead, Tin, Granular-Polymer）
・Elastic Slider
Isolator Sliding
・Rigid Slider
Bearing
・Friction Pendulum
・Steel Ball Bearing

Roller bearing
・Linear motion type
Fig. 2.1.1 Classification of isolators
Isolators can be classified based on whether they perform single or compound functions. The
compound function type operates not only as an isolator but also as a damper, such as a
high-damping rubber bearing or a plugged rubber bearing, as described later. Sliding bearings
can be classified as intermediate between the compound function type and the single function
type, depending on the extent of their damping function. Among the sliding bearings, curved
sliding bearings can be considered to be the compound function type consisting of an isolator
and damper. Among the roller bearings, a direct acting roller bearing consists of a
combination of direct acting devices. An example is a bearing device consisting of a steel
block, around which steel spheres circulate, and a bearing rail, arranged in a cross shape,
which is capable of withstanding large axial tension forces.
@Seismicisolation
@Seismicisolation
Dampers can be classified into a hysteretic damper, hydraulic damper, and viscoelastic
damper, as shown in Fig. 2.1.2. Hysteretic dampers can be classified into those that use
energy absorption by plastic deformation of the material, such as steel and lead dampers, and
those that use energy absorption due to frictional force, such as frictional dampers.
Hydraulic dampers include oil dampers that use the damping force from the pressure
difference on the front and rear of a piston produced by the resistance force when oil flows
through a valve, and viscoelastic dampers that use the shear resistance force of a highly
viscous material as the damping force. Oil dampers are normally given properties that are
linearly proportional to the velocity by varying the shape of the orifice of a valve, in which
the force is proportional to the exponent of the velocity. When a certain pressure is exceeded,
a bypass valve opens and it becomes possible to obtain bilinear properties. Viscous dampers
produce a relative velocity in a high viscosity material that is placed between flat plates at a
constant spacing. Viscous dampers consist of two types: those with resistance plates placed
horizontally (flat type viscous dampers), and those with resistance plates placed vertically
(wall type viscous dampers).
Cylindrical shaped fluid viscous dampers, an American product, use the resistance of a
silicone fluid and an orifice, or they have an amplification mechanism that uses the rotational
drive of a ball screw (a rotary damping tube). Other viscoelastic dampers, which use a
viscoelastic substance conventionally used in vibration control devices, can also be applied to
seismically isolated structures.
・Steel Damper
Hysteretic ・Lead Damper
・Friction Damper
・Oil Damper
Damper Hydraulic
・Viscous damper
Viscoelastic
Fig. 2.1.2 Classification of dampers
Hysteretic dampers obtain the required damping force by a comparatively simple mechanism.
The restoring force properties vary depending on the material, but dampers using mild steel
display a smooth spindle form. Frictional dampers include types in which frictional surfaces
are brought into contact at a constant force by disk springs or bolt tension, and types that
absorb energy by a frictional force while supporting the load of the building. Sliding bearings
can be considered to be one type of frictional damper. Their restoring force properties display
an almost perfectly rigid plastic shape. It is necessary to be cautious about various
@Seismicisolation
@Seismicisolation
dependencies such as the durability of the frictional surfaces and the frictional coefficient.
Of the hydraulic dampers, oil dampers exhibit a damping force that is proportional to the
velocity, and viscous dampers exhibit a damping force that is almost proportional to the
exponent of the velocity. The restoring force properties exhibit a smooth elliptical shape,
which has the advantage of reducing the response acceleration of the superstructure.
However, caution is necessary regarding the effect of ageing of the viscous material used,
dependence of the temperature and velocity of the damping force.
In seismic isolation of detached houses, restoring devices that do not have any support
function, such as displacement constraint devices for strong winds, are sometimes installed
as seismic isolation devices.
As described above, seismic isolation devices or seismic isolation systems that are a
combination of seismic isolation devices are required to be capable of supporting the
building, as well as to be capable of horizontal displacement and seismic energy absorption.
The vertical support capability of seismic isolation devices can be obtained by confirmation
tests to determine the ultimate compression load that can be stably supported. The ultimate
deformation capability in the horizontal direction is also basically confirmed by testing, but is
also necessary to confirm that the load support capability and the seismic energy absorption
capability are not greatly affected by the amount of horizontal deformation. In recent years, it
has been reported that for a high-damping rubber bearing, when horizontal forces in two
directions are applied, torsional deformation accumulates, and failure displacement is
reduced compared with a displacement applied in one direction. If using seismic isolation
devices with no possible reduction in the ultimate performance when loads are applied in two
directions, it is necessary to assure a design margin in the ultimate performance.
The seismic energy absorption capability has been verified to date by repeated deformation
tests carried out on seismic isolation devices. However, a concern is that for long-period and
long-duration ground motions, the oscillations will continue for a long period of time, so the
energy absorption capability due to repeated deformation should be re-verified. By
considering the energy balance with respect to a single lumped mass model having a seismic
isolation level with bilinear restoring force properties, the following equation is established
between the equivalent velocity VE of the seismic input energy and the cumulative plastic
deformation δp.
VE2
δp = (2.1.1)
2 gα s
Here, as: yield shear force coefficient (= seismic isolation level yield load / building mass).
@Seismicisolation
@Seismicisolation
100
80
αs= 0.01
60
δp (m)
40 0.03
0.05
20
0.10
0
0 100 200 300 400 500
VE (cm/s)
Fig. 2.1.3 Relation of energy input and cumulative plastic deformation
Figure 2.1.3, which is based on equation (2.1.1), shows that as the equivalent velocity of the
seismic input energy on the horizontal axis becomes larger, the cumulative plastic
deformation of the seismic isolation devices increases. Even at the level of VE = 400 cm/s, if
the yield shear force coefficient for the seismic isolation level is 0.03 or higher, it can be seen
that the necessary cumulative plastic deformation is 30 m or less. In the case that a sine wave
is applied to the seismic isolation devices in a vibration test, a cumulative deformation of
30 m is equivalent to about 38 repetitions if the vibration amplitude is ±0.2 m, and 19
repetitions if the amplitude is ±0.4 m. It is necessary to confirm that the energy absorption
performance is sufficient by repeated tests on the seismic isolation devices.
@Seismicisolation
@Seismicisolation
Section 2.2 Mechanical Properties of Each Device
2.2.1 Rubber Bearing
(1) Structure and characteristics
The structure of a rubber bearing has thin steel plates (intermediate steel plates) that are
alternately laminated with thin rubber sheets. When the rubber layers are thick, they are
displaced downwards considerably when they are compressed. It can be easily seen that the
supporting capability with respect to compressive loads is low. If the thickness of each
rubber layer is small, the lateral spread of the rubber can be greatly reduced. Therefore, it is
possible to reduce the downward displacement under compressive loads. During shear
deformation, the steel plates do not constrain the shear deformation. The horizontal stiffness
keeps the flexibility of the rubber itself.
Figure 2.2.1 shows the cross section of a rubber bearing. The parameters that determine the
shape of the rubber bearing are the rubber diameter D, the rubber layer thickness tR, and the
number of rubber layers n. These parameters are incorporated into the first shape factor S1
and the second shape factor S2.
The first shape factor S1 is the same as the shape factor in the conventional field of
anti-vibration rubber and is defined by the following equation.
Constrained area of rubber (compressed area)

S1 =
Free surface area of one rubber layer (side area)
The second shape factor S2 is a coefficient that is newly introduced for a rubber bearing and
is defined as:
Rubber diameter
S2 =
Total rubber thickness
S1 is mainly related to the vertical and bending stiffness, and S2 is mainly related to the
buckling load and horizontal stiffness. In the case of a laminated rubber bearing with a
circular cross section, S1 and S2 can be calculated from equation (2.2.1). The constrained
area is the area of the rubber layer that is constrained by the intermediate steel plates. If the
bearing has a central hole, or if the central hole is filled with a lead plug, attention has to be
paid to the calculation of the constrained area and the free surface area.
πD 2 4 D D (2.2.1)
S1 = = , S2 =
πDt R 4t R nt R
When a rubber bearing is subjected to compression, the rubber tends to deform to the
outside, and this deformation is constrained by the intermediate steel plates. In addition, the
incompressibility of the rubber material (Poisson's ratio of about 0.5) forms a triaxial
@Seismicisolation
@Seismicisolation
compression state (hydrostatic compression state) in the center of the rubber layer. This is
to say that, conceptually, the rubber layer resembles confined water within the rubber’s
molecular network, and water will not leak from the rubber layer. Therefore, the amount of
deformation due to compression is extremely small. The rubber bearing is capable of
exhibiting the same high compression stiffness as a single reinforced concrete column
having the same cross-sectional area.
Diameter (D)
Flange Plate
Height Steel Plate（Thickness ）

H = ( n − 1)t S + nt R Rubber（Thickness ）
Fig. 2.2.1 Section of a rubber bearing
The various types of rubber bearing isolators include the following:

• natural rubber bearing
• high-damping rubber bearing
• lead plug rubber bearing
• tin plug rubber bearing
Laminated rubber bearings except for natural rubber bearings have the function of a
damper. High-damping rubber bearings absorb seismic energy by a combination of viscous
and plastic properties as a result of a special rubber blend. Lead plug and tin plug rubber
bearings have a lead or tin cylindrical plug inserted into the center of the rubber bearing.
They absorb seismic energy by plastic deformation of the lead or tin following the shear
deformation of the rubber bearing. The diameter of the lead plug is usually about 1/5 of the
external diameter of the rubber bearing.
A circular shape, which has no directionality, is usually used as the plane shape of the
rubber bearing. Recently, a square rubber bearing has also been used. A square rubber
bearing should be used after confirming that the directionality of properties, such as
hysteresis in the horizontal direction, and failure, is sufficiently small.
(2) Horizontal performance
(a) Hysteresis properties
When a shear force acts on a rubber bearing, the intermediate steel plates do not
constrain the shear deformation (with no change in volume) of the rubber layers, and
flexible horizontal stiffness of the rubber sheets is exhibited. A laminated rubber bearing
maintains its load supporting capacity even when it is highly deformed. This is because
the triaxial compression stress state continues to be formed in the portion where the
uppermost and lowermost laminated rubber layers overlap (referred to as the effective
supporting portion), as shown in Fig. 2.2.2. Most of the compression load is supported
by this effective supporting portion, and the compression stress (reaction force) in this
@Seismicisolation
@Seismicisolation
portion becomes extremely large. On the other hand, generation of the tensile reaction
force is very limited. The constraint confining effect on "water" becomes smaller, and it
is considered that the compression stiffness becomes smaller in proportion to the
cross-sectional area of the effective supporting portion when shear deformation is
applied. Still, the vertical displacement is at an extremely small level. As a result of this
load mechanism, a rubber bearing can deform greatly in the horizontal direction while
supporting large vertical loads.
Fig. 2.2.2 Vertical stress distribution by FEM analysis of rubber bearing

(Compressive stress: 30 MPa)
Figure 2.2.3 shows examples of hysteresis curves for various types of laminated rubber
bearings. The compressive stress in the tests was 10 or 15 N/mm2, and test results were
obtained up to a shear strain of 300%. The shear strain is represented as the ratio of the
horizontal displacement to the total rubber thickness. The shear stress on the vertical
axis of the figures is calculated by dividing the restoring force by the cross-sectional
area of rubber bearing. The high-damping rubber bearing and the lead plug rubber
bearing have a damping function and exhibit bilinear-shaped hysteresis curves. The
modeling of the restoring force properties is as follows: for the natural rubber bearing, a
linear model is usually used, and for the high-damping rubber bearing and the lead plug
rubber bearing, a bilinear model adapted to each shear strain is frequently used (which is
why it is called a corrected bilinear type).
2 2
Specimen:600×6.0-26 Specimen:800×6.0-26 Specimen:800×5.4-37
1 Axial Stress:15N/mm2
Shear Stress (N/mm2)
Axial Stress:10.5N/mm2
Axial Stress:15N/mm2
S1=31.7, S2=5.1 1 S1=33.3, S2=5.1 1 S1=36.1, S2=4.00
0 0 0
-1 -1
-1
-2 -2
-300 -200 -100 0 100 200 300 -300 -200 -100 0 100 200 300 -300 -200 -100 0 100 200 300
Shear Strain (%) Shear Strain (%) Shear Strain (%)
(a) Natural rubber bearing (b) Lead rubber bearing (c) High-damping rubber bearing
Fig. 2.2.3 Hysteresis curves for rubber bearings
Figure 2.2.4 schematically shows the restoring force properties of a laminated rubber
bearing. The graph shows the hysteresis curves for different compressive stresses (σ1 ≤
σ2 ≤ σ3). The buckling stress for a laminated rubber bearing is almost proportional to the
elastic shear modulus G of the rubber multiplied by the shape factor, GS1S2. As the
compressive stress is increased to the buckling stress, the initial horizontal stiffness and
deformation capability of the laminated rubber bearing are reduced.
@Seismicisolation
@Seismicisolation
σ1
σ2
c b
σ3 Shear Strain γ
Fig. 2.2.4 Example of restoring force characteristics of a rubber bearing
In hysteresis curve a, the shear force increases strongly in proportion to the horizontal
deformation up to 200% to 300% of the shear strain and exhibits uniform horizontal
stiffness. Thereafter, the horizontal stiffness gradually increases (referred to as
hardening) until final failure of the rubber layer. For a laminated rubber bearing with a
second shape factor of about 5 or more, the shear strain at breaking is often 400% or
more. To induce breaking in the rubber layer, it is necessary that the strength of adhesion
between the intermediate steel plates and the rubber layers is sufficiently large.
Hardening is the result of strain hardening of the rubber material, and it causes an
increase in the floor response of the superstructure, while it also has the effect of
suppressing the response displacement of the seismic isolation level.
The hardening phenomenon does not appear in hysteresis curve b. After a certain
deformation, the horizontal tangent stiffness becomes negative and the restoring force is
lost. This type of phenomenon is caused by buckling of the laminated rubber bearing.
Generally, buckling is the point at which the horizontal stiffness becomes zero, indicated
by ○ in the figure. However, at the point with zero tangent stiffness, the load supporting
capacity is maintained. Therefore, the point which means actual instability is considered
to be the point indicated by ◆, where the shear force becomes ze
In hysteresis curve c, the initial stiffness is negative in the small deformation region.
Thereafter, although the stiffness becomes positive, the deformation capability is
reduced. This is because buckling occurs at the initial deformation and the apparent
restoring force follows due to strain hardening of the rubber material. In this case,
although it exhibits a certain resistance force, the laminated rubber bearing is in the
buckled state. It should be noted that it has no load supporting capacity. Compressive
stress with zero horizontal stiffness corresponds to the buckling stress σcr indicated by
equation (2.2.6). It should also be noted that σcr has no direct relation to the buckling
behavior in the large deformation region.
(b) Dependence on compressive stress
The horizontal stiffness KH of a laminated rubber bearing can be obtained from the
following equation, which is based on the solution to the buckling problem, where a
horizontal force and a compression force act on an elastic body (Haringx theory1)).
@Seismicisolation
@Seismicisolation
P2
KH =
 qH 
2k r q tan   − PH (2.2.2)
 2 
P P
q= 1 + 
kr  ks 
Here,
P: compressive load
H: total thickness of rubber layers and intermediate steel plates (= TR+TS)
TS: total thickness of intermediate steel plates (= (n-1)tS)
TR: total thickness of rubber layers (= ntR).
The effective shear stiffness and the effective bending stiffness are obtained from the
following equations,
H , k r = Erb I H (2.2.3)
k s = GA
TR TR
where
E r Eb ,  2  , G: elastic shear modulus of rubber
E rb = E r = 3G  1 + κS12 
E r + Eb  3 
A: cross-sectional area, I: second moment of area for the cross section
Eb: bulk modulus, κ: correction coefficient depending on the rubber hardness.
The correction coefficient κ for rubber hardness approaches 1 as the rubber becomes
flexible, and is in the range 0.93 to 0.85 for normal rubber with an elastic shear modulus
of G = 0.29 to 0.44 MPa2). Usually, it is heuristically determined for each product
together with the bulk modulus3).
In equation (2.2.2), considering that tan(qH/2) ≅ qH/2 holds when the compression load
approaches zero, the horizontal stiffness KH0 with zero compressive load can be
obtained from the following equation. This is the same as the equation that considers
only the shear stiffness of the rubber layers.
GA (2.2.4)
KH0 =
TR
Equation (2.2.4) is extremely effective for evaluating the horizontal stiffness of the
laminated rubber bearing. However, when applying the equation, it is necessary that the
shape of the laminated rubber bearing has a small variation in horizontal stiffness
against the variation of the compressive load, as shown in Fig. 2.2.5. This is essential in
the design of the laminated rubber bearing.
@Seismicisolation
@Seismicisolation
(c) Buckling load
The buckling load Pcr is obtained by setting KH = 0 in equation (2.2.2), in other words,
when qH = π.
1  4π 2 k 
(2.2.5)
Pcr = k s  1 + 2 r − 1
2  H ks 
 
Substituting ks and kr in equation (2.2.3) into equation (2.2.5), and considering that Er ≅
2GκS12, D = S2Tr, and that S12S22 is an extremely large value, the equation can be
approximated as follows:
2
Pcr 1 H  Tr  κ (2.2.6)
σ cr = = G  πS1 S 2  = ζGS1 S 2
A 2 Tr  H
(
2 1 + 2κS12 G E b )
where
κ .
ζ =π
(
8 1 + 2κS12 G Eb )
When applying equations for the horizontal stiffness or the buckling load to a
high-damping rubber bearing, it is necessary to substitute the equivalent elastic shear
modulus at each shear strain for the elastic shear modulus of the rubber. Furthermore,
for calculating the buckling load, the intermediate steel plates are considered to be rigid.
If it is not possible to consider the intermediate steel plates as sufficiently rigid, a
reduced buckling load is obtained depending on the diameter of the laminated rubber
bearing, the rubber thickness, and the compressive stress. If the intermediate steel plates
are thin and extremely flexible, the buckling load is reduced to about 1/2 compared with
the case that they are rigid, which requires caution in the design4).
KH/KH0
1
0.8
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
σ /σcr
Fig. 2.2.5 Influence of compressive stress dependence of horizontal stiffness
@Seismicisolation
@Seismicisolation
σ /σcr
1
0.8
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
δ / D = γ / S2
Fig. 2.2.6 Stability limit deformation
The relation between the horizontal stiffness of the laminated rubber bearing and the
compressive stress (equation (2.2.2)) can be approximated by the following equation.
2
KH  σ 
= 1 −  
 (2.2.7)
KH0  σ cr 
Here, σcr is the buckling stress of the laminated rubber bearing, and it is determined by
the shape and the material properties. The compressive stress dependency of the
horizontal stiffness depends on the shape and the material properties. By focusing on the
buckling stress in the equation above, the dependency can be evaluated. Figure 2.2.5
shows the variation in the horizontal stiffness from equation (2.2.7). It is basically
necessary to take into account the properties and the design condition of the rubber
bearings. These become the measure for sound use of rubber bearings, that is, to
determine the range of the long-term and the short-term compressive stress with respect
to the buckling compressive stress σcr.
For example, if the variation in the compressive stress is within the range of 0 to 0.45
with respect to the buckling stress, the variation in horizontal stiffness is -20% or
smaller.
If it is assumed that the whole compression load is supported by the effective supporting
portion (Ae) when the laminated rubber bearing is deformed in shear, the stability limit
can then be considered to occur when the compressive stress at the portion is equal to
the buckling stress. The broken line in Fig. 2.2.6 has been drawn based on this.
Likewise, in the literature on the buckling load subjected to shear deformation5), it has
been shown that the buckling load decreases corresponding to the amount of shear
deformation δ. Stable shear deformation can be obtained from the following equation.
σ δ γ (2.2.8)
= 1 − cr = 1 −
σ cr D S2
where γ is the shear strain (γ = δ/TR).
@Seismicisolation
@Seismicisolation
Based on the results of property tests on rubber bearings6), it has been confirmed that
equation (2.2.8) generally errs on the side of safety. However, little data is available in
the range where the shear deformation is small and the compressive stress is high. The
colored area in Fig. 2.2.6 shows an example of the parameter range for a laminated
rubber bearing. In the design of seismically isolated buildings, it is important that
laminated rubber bearings are used within the range where they can exhibit stable
performance. However, even within this range, the hysteresis properties of laminated
rubber bearings vary due to compressive stress dependence and strain dependence. It is
necessary to investigate how much this variation affects the response of seismically
isolated buildings.
The stability of the laminated rubber bearing can be evaluated based on the buckling
stress, as described above. However, the stability of the hysteresis properties varies
depending on the type, shape, and rubber properties of the laminated rubber bearing.
Therefore, when setting the design displacement and design compressive stress, it is
essential to make a judgment based on test data for the actual product to be used.
In addition, laminated rubber bearings have velocity dependency, temperature

dependency, repeated cycling dependency, and variation due to manufacturing. It is
necessary to verify these dependencies by tests. Repeated deformation for a long time in
earthquakes with a long period and a long duration is a concern. It is necessary to
confirm the energy absorption performance of laminated rubber bearings having a
damper function. Furthermore, horizontal creep deformation occurs in lead plug rubber
bearings or high-damping rubber bearings in response to strong winds on high-rise
buildings, and the buildings are subjected to repeated deformation over a long period of
time. It has been confirmed that the hysteresis curve differs from the initial state, for
example, in the reduction of yield resistance due to repeated deformation. It is necessary
to appropriately reflect this effect in the design. As a general rule, full-size products
should be used in tests, but in cases where reduced scaled specimens are used due to the
performance of the testing machine, it is necessary to use precisely scaled models of the
actual products.
(3) Vertical performance
(a) Compression properties
The compressive stiffness of a laminated rubber bearing KV can be calculated from the
following equation using the longitudinal elastic modulus Ec.
E cb A (2.2.9)
KV =
TR
where E cb =
E c Eb
E c + Eb
, ( )
E c = 3G 1 + 2κS12 .
It has been shown experimentally that the compressive stiffness of a rubber bearing
subjected to shear deformation decreases in almost direct proportion to the effective
supporting area. If the shear deformation is large, the effective supporting area
@Seismicisolation
@Seismicisolation
decreases, then the compression stiffness is reduced at that time. However, it has also
been found that if the shear deformation is large, the reduction is not simply
proportional because of strain hardening of the rubber material. Therefore, based on
equation (2.2.9), the lower limit of the compression stiffness during shear deformation
can be obtained from the following equation.
Ae
K Ve = KV (2.2.10)
A
where Ae is the effective supporting area.
When the shear deformation is δ, the effective supporting area, calculated by the
following equation, is the area of the overlapping portion of the top and bottom surfaces
of the circular laminated rubber bearing.
  2 
2 δ δ  δ 
Ae = 1 −  1 −   + sin −1    A
 π
  D  D  D   (2.2.11)

 δ δ 

.
≅ 1 − 12
D  A  ≤ 0.6
D 
A laminated rubber bearing with a central hole can be conveniently dealt with by using
the first shape factor, which takes the central hole into account.
It has also been observed in compression failure tests of laminated rubber bearings that
the constraint of the rubber layers was lost as a result of tensile failure of the
intermediate steel plates, and thus the result was complete failure7). From these test
results, it can be seen that the ultimate compression resistance of a laminated rubber
bearing is governed by the strength and thickness of the intermediate steel plates.
Figure 2.2.7 shows the amount of vertical deflection of a natural rubber bearing when
subjected to shear deformation. In addition to the compressive deformation due to the
axial compression force, deflection is produced in the vertical direction in the laminated
rubber bearing due to the shear deformation. It is necessary to be careful about this
effect when setting vertical clearances.
12
Vertical Displacement (mm)
Specimen:600×6.0-26
10 S =31.7, S =5.1
1 2 s=25MPa
6
s=15MPa
4
2 s=5MPa
0
-300 -200 -100 0 100 200 300
Shear Strain (%)
Fig. 2.2.7 Vertical deformation due to horizontal deformation of rubber bearing
@Seismicisolation
@Seismicisolation
(b) Tension properties
When tension acts on a laminated rubber bearing, the relation between the tensile load
and the tensile deformation is bilinear. This is because when tension acts on a laminated
rubber bearing, a negative pressure state exists within the rubber, which causes voids
within the rubber layers. The tensile stiffness is usually unaffected by the magnitude of
the shear deformation. On the other hand, if a large tensile strain has been experienced,
the tensile stiffness is reduced8). The tensile forces act on the laminated rubber bearing
through the installation bolts and flanges of the laminated rubber bearing. Therefore, the
initial tensile stiffness is affected by the thickness of the bearing flanges after the
laminated rubber bearing has been subjected to large deformation, and the effect of the
flange thickness on the tensile load-deformation relation becomes relatively small9). As
the size of the laminated rubber bearing is increased, it has been reported that the tensile
strain at failure decreases10). Therefore, it is necessary to sufficiently verify the scale
effects and degradation that occur after being subjected to tensile deformation.
Compared with the compressive stiffness, the initial tensile stiffness is small, at about
1/5th to 1/10th. As the tensile deformation increases, the stiffness decreases to about
1/100th. The initial tensile stiffness is affected by the shape of the laminated rubber
bearing, and the thickness of the flange, and varies depending on the range of tensile
deformation over which the stiffness was evaluated. It is necessary to evaluate the
stiffness based on tests on the laminated rubber bearing that is actually used.
As the result of tests to evaluate the properties in the tensile direction, it has been shown
that the tensile deformation capability of the laminated rubber bearing is high.
Consequently, the number of design cases where the tensile load is allowed to act has
increased. It is considered that the performance in the tensile direction can be evaluated
from the tensile force and tensile strain. Tensile force is transmitted through the flanges
of the laminated rubber bearing. For the dimensions of the flanges that are normally
used, bending deformation is produced, so the tensile stress is not uniformly transmitted
to the laminated rubber bearing. Therefore, it is considered appropriate to evaluate the
properties of the laminated rubber bearing on the tensile side by using the tensile strain.
Even when the tensile stress is governed by the allowable pullout force, it is valid to
convert the allowable tensile stress in accordance with the properties of the rubber
material. For example, a method in which the allowable value is determined by a
multiple of the coefficient of the shear modulus G of rubber can be considered.
Figure 2.2.8 shows the relation between the tensile strain and the shear strain for a
laminated rubber bearing. The tensile strain is calculated by dividing the tensile
deformation by the total rubber thickness of rubber bearing. Although the type and size
of the laminated rubber bearings used in the tests were different, by focusing on the
natural rubber bearing only, it can be seen that generally the same trend is exhibited. In a
simple tensile test of a laminated rubber bearing, the deformation capability has a tensile
strain of 300% or more. When tensile deformation is applied after the shear deformation
has been applied, the failure tensile strain is reduced by the amount of the shear strain.
In the figure, the tensile stress at the maximum tensile strain is also shown. It can be
@Seismicisolation
@Seismicisolation
seen that up to a shear strain of about 200%, the tensile resistance capability can be
exhibited for a tensile stress of 2 MPa or more.
Black mark : Ruptured

White mark: not Ruptured
300 >4MPa NRB
250 HDR
Tensile Strain (%)

LRB
200
3MPa
150
2MPa 2.5MPa
100
50
1.6MPa
0
0 100 200 300 400
Shear Strain (%)
Fig. 2.2.8 Tensile limit strain of a laminated rubber bearing

(NRB: natural rubber bearing, HDR: high damping rubber bearing, LRB: lead plug rubber bearing)
(4) Others
The primary role of the intermediate steel plates is to constrain the rubber layers. During
compression, the stress acting on the intermediate steel plates is in the tensile direction and
does not have a large magnitude. However, during shear deformation, out-of-plane bending
deformation is generated in the intermediate steel plates to support the compression load in
the effective compression part. Partial plastic strain is assumed. The thickness of the
intermediate steel plates used for the laminated rubber bearing is 0.3 to 0.8 times the
thickness of the rubber layer, and about 0.5 on average. Frequently, the thickness of the
intermediate steel plates is constant, even if the diameter of the laminated rubber bearing is
large, so the diameter increases as the intermediate steel plates become relatively thin. For
out-of-plane bending deformation, the ratio of the diameter to the thickness of the
intermediate steel plates is a problem. When the intermediate steel plates are thin, negative
effects have been reported, such as reduction in the buckling load due to the presence of the
central hole4). It is considered that the thickness of the intermediate steel plates has a large
effect on the mechanical properties, particularly on the ultimate properties of the laminated
rubber bearing.
Finite element analysis of the laminated rubber bearing has shown that when the bearing
has a central hole, the stress produced in the intermediate steel plates increases by a factor
of 2 compared with when the bearing has no hole, and the area of the yielded part
increases11). In this way, the central part of the rubber plays an extremely important part in
the load resistance mechanism, and it is desirable not to have a central hole. If a central hole
is necessary, it is essential that the size be as small as possible.
The European standard “Structural bearings Part 3: Elastomeric bearings”

(EN1337-3:2005)12) prescribes the following equation for the thickness ts of the
intermediate steel plates.
@Seismicisolation
@Seismicisolation
K p ⋅ FZd ⋅ (t1 + t 2 ) ⋅ K h ⋅ γ m
tS = （2.2.12）
Ar ⋅ f y
Here, Fzd: vertical axial load, Ar: effective supporting area, t1 and t2: thicknesses of the
rubber layers above and below the intermediate steel plate, respectively, fy: yield stress of
the steel material, Kp: stress correction coefficient (= 1.3), Kh: coefficient depending on
whether a central hole exists (=1: without central hole, =2: with central hole), and γm: safety
factor (here, 1.0 is recommended).
In (2.2.12), if t1 = t2 = t3, and if the effective supporting area is obtained from equation
(2.2.11),
tS A σ (2.2.13)
= 2.6 K h
tR Ae f y
where σ : compressive stress, without central hole: Kh = 1, with central hole: Kh = 2.
Figure 2.2.9 shows the relation between the amount of horizontal deformation and the
thickness of the central steel plates based on equation (2.2.13). If σ/fy is about 0.02, the
value of ts/tr is about 0.5, which is acceptable. For higher compressive stress, it is necessary
to have thicker intermediate steel plates. Since the application range of equation (2.2.12)
has not been clearly indicated, it cannot be applied directly to the products in Japan, but can
be used as a reference.
Soliδ Line : with Central Hole
Broken Line : without Central Hole
10
σ/fy=0.1
tS / tR
σ/fy=0.05
σ/fy=0.02
0.1
0 0.2 0.4 0.6 0.8 1
δ/D
Fig. 2.2.9 Thickness of intermediate steel plates due to EN1337-3
For the dimensional inspection of the laminated rubber bearing, the diameter, height, and
inclination are checked to verify they are within the prescribed range of values. Until now,
no tests have been conducted on the thickness of one rubber layer or the parallelism
tolerance, which does not appear to be much of a problem. Although the effects of the
thickness of each rubber layer and the dimensional tolerance on the performance of the
laminated rubber bearing have not yet been investigated, it is considered that the uniformity
of the thickness of the rubber layers has a large effect on the buckling performance and the
ultimate performance. Allowable tolerances of the accuracy of the laminated rubber bearing
should be prescribed for the thickness of the rubber layers. Particularly for the laminated
@Seismicisolation
@Seismicisolation
rubber bearing with large diameter (for example 1,000 mm or larger), it is questionable that
better accuracy with an error less than 1/10 mm is ensured. Therefore, efforts are essential
to ensure quality by sufficient performance inspection on full-size products.
It is desirable that the installation parts of the laminated rubber bearing be as parallel as
possible. However, in some cases, an angle of rotation is produced due to tilting as a result
of construction accuracy, manufacturing accuracy or insufficient bending stiffness of the
structural frame (including cases of seismic isolation on column heads or pile heads).
Section 4.6.4 explains the properties in the case that the structural frame for mounting a part
of a rubber bearing is inclined.
References
1）J. A. Haringx：On Highly Compressible Helical Springs and Rubber Rods and Their Application for Vibration-Free
Mountings, Philips Research Reports, Vol.3, 1948 & Vol.4, 1949
2）P. B. Lindley：“Natural Rubber Structural Bearings”, Joint Sealing and Bearing Systems for Concrete Structures, Vol.1,
ACI, 1981
3）P. B. Lindley：“Engineering Design with Natural Rubber”, The Malaysian Rubber Producers Research Association, 1974
4）J. M. Kelly：The Influence of Plate Flexibility on the Buckling Load of Elastomeric Isolators, UCB/EERC-94/03, March,
1994
5）Buckle, I. G. and Liu, H．：Stability of Elastomeric Seismic Isolation Systems, Proceedings of Seminar on Seismic
Isolation， Passive Energy Dissipation and Active Control, ATC-17-1, Vol.1, 1993
6) Yoshitaka Muramatsu, Kiyotaka Inoue, Ryuuichi Katoh, Naoyuki Kamitani, Tohru Sakaguchi, Yoritaka Sasaki, Shigenobu
Suzuki, Shinji Sera, Mitsuru Miyazaki, Haruyuki Kitamura：Test Results of Properties of Rubber Bearings for Building，
Journal of Technology and Design, AIJ, No.20, pp.67-70, 2004.12
7）Mineo Takayama, Keiko Morita, Hideyuki Tada：Ultimate Load Capacity of Natural Rubber Bearings, Journal of
Structural and Construction Engineering, AIJ, No.482, pp.43-51, 1996.4
8）Nagahide Kani, Naoyuki Iwabe, Mineo Takayama, Keiko Morita, Akira Wada：Experimental Study on the Tension
Capacity in the Displaced Position of Elastomeric Isolators (part.1~part.3), Summaries of Technical Papers of Annual
Meeting, Architectural Institute of Japan, pp.559-564, 1999.9
9）Ryuichi Kato, Kenjiro Oka, Mineo Takayama, Hiroaki Kunitsu, Miwa Sadamoto：The Tensile Tests of Natural Rubber
Bearings Focused on the Effect of the Steel Frange Plates, Journal of Technology and Design, AIJ, No.15, pp.99-104,
2002.6
10）Yoshitaka Muramatsu, Ichiro Nishikawa, Ichizou Kawabata, Masaharu Takayama, Yuichi Kimura：Tensile property of
Large-Sized Natural Rubber Bearing, Journal of Technology and Design, No.12, pp.53-56, 2001.1
11) Mineo Takayama, Keiko Morita：Maximum Stress of Interlayer Steel Plates in Elastoric Isolator, PVP-Vol.341, Seismic,
Shock, and Vibration Isolation, ASME, 1996
12) EN1337-3:2005 ”Structural bearings Part 3：Elastomeric bearings”
@Seismicisolation
@Seismicisolation
2.2.2 Sliding Bearings
Sliding bearings are isolators that utilize sliding on low friction materials. These bearings
are classified into rigid sliding bearings and elastic sliding bearings. As shown in Fig.
2.2.10, rigid sliding bearings are isolators consisting of a sliding part (sliding member and a
sliding plate) only. They have a rigid-plastic type of hysteresis loop. Elastic sliding bearings
are isolators consisting of a sliding part and a laminated rubber bearing, as shown in Fig.
2.2.11. Their hysteresis loop is a perfectly elastic-plastic type.
A sliding member utilizes resin such as PTFE (polytetrafluorethylene) or polyamide, which

is mixed with filler material. The sliding plate utilizes stainless steel with a mirror finish or
stainless steel to which coating has been applied.
The laminated rubber bearing is made of elastic rubber (which is either natural rubber or
synthetic rubber) with an elastic shear modulus of 0.6 to 0.8 N/mm2 and a second shape
factor of 10 or larger in many cases.
Flange
Pot plate
40
Horizontal Force (kN)
20
Back metal Rubber plate

Sliding material -20
Sliding plate -40

-300 -200 -100 0 100 200 300
Sole plate
Horizontal Displacement (mm)
(a) Components (b) Example of a hysteresis loop

Fig. 2.2.10 Rigid sliding bearing
Flange Shear key

Connecting plate
100
50
Back metal Laminated rubber

Sliding material -50
Sliding plate -100

-300 -200 -100 0 100 200 300
Sole plate Horizontal Displacement (mm)
(a) Components (b) Example of a hysteresis loop

Fig. 2.2.11 Elastic sliding bearing
@Seismicisolation
@Seismicisolation
Sliding bearings can be classified into three types: high friction, intermediate friction, and
low friction. The friction coefficient under long-term compressive stress is 0.10 to 0.15 for
the high friction type, about 0.07 for the intermediate fiction type, and 0.01 to 0.03 for the
low friction type.
The friction coefficient varies depending on the manufacturing variation, changing aging
dependency on compressive stress, velocity, and temperature, and repeated deformation. It
is necessary to take all of these into consideration in the design.
The manufacturing variation is between ±20% and ±30% for the high friction and
intermediate friction types, and ±40% and ±50% for the low friction type1). Little
experimental data on the variation with aging is available, but it is considered that the
variation is about +10% to +20%.
As shown in Fig. 2.2.12, the friction coefficient of sliding bearings tends to become
exponentially smaller as the compressive stress increases. In the range where the
compressive stress is small, attention has to be paid to the particularly large variation.
The trend of variation in the friction coefficients according to velocity varies depending on
whether the sliding member is PTFE or polyamide resin. As shown in Fig. 2.2.13, in the
case of PTFE, the friction coefficient becomes larger as the velocity increases, but becomes
smaller in the case of polyamide resin. The variation is larger in the case of polyamide
resin. In both cases, the variation tends to be smaller as the velocity increases.
The friction coefficient varies virtually linearly with the environmental temperature, as
shown in Fig. 2.2.14. The friction coefficients can differ by more than 30% between -10°C
and 40°C, so it is necessary to consider the temperatures of the construction location in the
design.
Rapid
Change
Coefficient of Friction
Gradual
Change
0 10 20 30 40
Contact Pressure (N/mm 2)
Fig. 2.2.12 Contact pressure dependency of coefficient of friction (sliding bearing)
@Seismicisolation
@Seismicisolation
：PTFE
：Polyamide resin
0 100 200 300 400 500
Velocity (mm/s)
Fig. 2.2.13 Velocity dependency of coefficient of friction (sliding bearing)

about 20-40%
of changes
-10 0 10 20 30 40
Temperature ( ℃)
Fig. 2.2.14 Temperature dependency of coefficient of friction (sliding bearing)
The variation in friction coefficients with repeated deformation greatly depends on whether
the sliding plate has been coated. In low friction types where the sliding plate has been
coated, the friction coefficient increases slightly as the number of repetitions increases.
After 40 repetitions, the variation is several percent (about +0.001 to +0.002), which is
extremely small. On the other hand, in high friction types where the sliding plate is not
coated, the friction coefficient tends to become smaller as the number of repetitions
increases. After 40 repetitions, the change in the friction coefficient is 30% or more (about
-0.03 to -0.05).
(3) Vertical performance
In sliding bearings, the rubber part is arranged in series with the sliding part (indicated as
shock absorption rubber in the case of rigid sliding bearings, and as the rubber bearing in
the case of elastic sliding bearings). The rubber part governs the vertical performance.
Therefore, the vertical performance of sliding bearings can be obtained by the same method
as that for laminated rubber bearings. Various specific property values depend on the
material and the specifications of the rubber part. Because the thickness of the rubber part is
considerably thinner than that of the laminated rubber bearing, the vertical stiffness tends to
be large and the vertical creep strain tends to be small.
@Seismicisolation
@Seismicisolation
2.2.3 Roller Bearings
Roller bearings are isolators that use the rolling motion of steel balls. They can be classified
into the linear motion type and the flat plate type. As shown in Fig. 2.2.15, the linear type
consists of rails and a block assembled into the linear system, and its characteristic is that it
can be subject to not only a compression load but also a tension load2). The plane type
consists of steel balls arranged in a plane, as shown in Fig. 2.2.16, and it has the
characteristic that the allowable load and the plane shape can be adjusted by the number
and arrangement of the steel balls3).
The hysteresis loop shown in Fig. 2.2.17 is modeled as a rigid plastic bilinear type in a
seismic response analysis. It has no restoring function, so frequently it is used in
conjunction with laminated rubber bearings as a seismic isolation system.
Fig. 2.2.15 Roller bearing “cross linear bearing” (Linear motion type)
When an isolator operated
A normal period
Fig. 2.2.16 Roller bearing “steel ball bearing” (Flat plate type)
@Seismicisolation
@Seismicisolation
80 40
Horizontal Foece (kN)

40 20
0 0
-40 -20
-80 -40
-1000 -500 0 500 100 -600 -300 0 300 600
Horizontal Displacement (mm) Horizontal Displacement (mm)
(a) Linear motion type (vertical load=9.70 MN) (b) Flat plate type (vertical load=7.24 MN)
Fig. 2.2.17 Examples of hysteresis loop of a roller bearing
(2) Horizontal characteristics
The friction coefficient for the linear type under long-term loads is between 0.005 and 0.01.
As shown in Fig. 2.2.18, the friction coefficient becomes larger in direct proportion to the
vertical load. The variation in manufacturing is small at less than +0.002, but it is necessary
to consider the changes in the friction coefficient due to tilting or twisting of the bearing, or
the direction of deformation.
Coefficient of Friction μ
0.005 ～0.01
sustained load
Compressive Load Pv
Fig. 2.2.18 Vertical load dependency of coefficient of friction (linear motion type)
The friction coefficient for the plane type is smaller than that for the linear type and does
not exceed 0.003 even with manufacturing variation, change with age, and various
dependencies. The reason is considered to be that the contact between the steel balls and the
steel plate is closer to a point contact than the contact of the linear type.
The change in the friction coefficient due to repeated deformation has been investigated for
the linear type, and it has been confirmed that the causes of changes in the friction
coefficient are indentations in the rolling surface (rail) in the small deformation region, and
temperature rises in larger deformation regions.
@Seismicisolation
@Seismicisolation
(3) Vertical characteristics
The vertical stiffness of the roller bearing depends on the size (vertical load bearing
capacity) of the members, but in the case of the linear type it is 2 to 5 times that of the
laminated rubber bearing, and in the case of the plane type it is 5 to 15 times that of the
laminated rubber bearing. For example, in the case of a long-term load (axial load) of
10,000 kN, the vertical stiffness of the laminated rubber bearing (G = 0.39 N/mm2, S2 =
5.1) is 4.09×106 N/mm, and the vertical stiffness of the linear type is 1.20×107 N/mm
(factor of 2.9), and the vertical stiffness of the plane type is 4.36×107 N/mm (factor of
10.7).
The vertical creep deformation with age is zero in the case of the plane type made only
from steel components. In the linear type, rubber shims are inserted between the upper and
lower blocks, so a settlement of between 0.08 and 0.36 mm corresponding to 3% of the
rubber thickness (2.5 to 12 mm) is anticipated2).
References
1) Akio Suzuki, Sakae Ueda, Toru Sakaguchi, Shinji Sera, Takashi Nakamura, Yoshitaka Muramatsu, Nobuo Murota,
Takaaki Yamamoto, Hideo Mitamura, Haruyuki Kitamura：An Investigation Report of the Characteristic of Sliding
Bearing for Base-Isolated Building, Journal of Technology and Design, No.19, pp.49-54, 2004.6
2) Hiroyuki Harada, Toru Suzuki, Koichiro Asano：Experimental Study of Seismic Isolator Using Linear Motion Guide
(part.1~part.2), Journal of Structural Engineering Vol.50B, pp.539-552, 2004.3
3) Hisaya Tanaka, Kenji Saito, Yukihiro Nishimura, Mitsuo Miyazaki, Takanori Miyoshi ： The Devlopment of
High-Performance Isolated Buildings Using Steel Bail Bearings：(part.2) Full Scale Performance Test of Steel Ball
Bearings，Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan B-2, pp.745-746,
2000.9
@Seismicisolation
@Seismicisolation
2.2.4 Steel Dampers
Steel dampers are devices that absorb vibration energy by the plastic deformation of
steel. In seismically isolated structures, steel dampers must exhibit an almost equal
resistance force (energy absorption capability) with respect to deformation in all
directions in the horizontal plane and must track large deformations of the seismic
isolation level. Therefore, steel dampers used in seismically isolated structures have a
special shape compared with the vibration control devices used in passive vibration
control structures. In the guideline editions up to the previous edition (third edition),
many examples of steel dampers are loop type steel rod dampers1), which consist of four
steel rods per set formed in a spiral shape, but at present the U-shaped steel damper is the
most common. U-shaped steel dampers are manufactured by cold bending material
equivalent to SN490B with strict control of the composition, followed by heat treating.
The shape is simple compared with the loop type steel rod dampers, and they can be
installed at the same position as laminated rubber isolators. So, the number of U-shaped
steel dampers can be comparatively freely selected, and they can be arranged in
accordance with the shape of the installation location. The shape and deformation
properties of U-shaped dampers are shown in Fig. 2.2.19. The structure and
characteristics, horizontal performance, and various dependencies and durability of
U-shaped steel dampers are described here.
Fig. 2.2.19 Typical unit and deformation of U-shaped steel dampers
U-shaped steel dampers deform by rising or falling in the vertical direction (height
direction) when deformed by moving (like a Caterpillar truck in the in-plane 0°
direction). When deformed in the normal 90° direction, they do not move vertically, but
the U-shaped parts extend by torsion near the connections. In the design, the size of
concrete upstands and piping should take into account the range of deformation in the
vertical direction produced when deformed in the 0° direction.
@Seismicisolation
@Seismicisolation
The U-shaped steel damper as a seismic isolation device is composed of several

U-elements arranged radially or combining some of the right angles of the double
U-elements in order to reduce the directionality of the restoring force properties as much
as possible. Figure 2.2.20 shows the repeated hysteresis curves for four U-shaped steel
dampers2) for each loading direction. The restoring force properties are generally
modelled in a bilinear model or a function approximation model. However, the bilinear
model obtained from the equivalent energy absorption in a certain deformation range
underestimates strain hardening. So, it is necessary to be careful when using the bilinear
model.
P [kN] P [kN]
600 600
δ [m] δ [m]
-1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1
-600 -600
Fig. 2.2.20 Hysteresis loops of U-shaped steel dampers (UD55×4)
When subjected to constant amplitude vibration in one horizontal direction, and when
the horizontal deformation (amplitude on one side) is greater than 1.5 times the height of
the damper, failure in the 90° direction becomes dominant; but, in a practical
deformation range smaller than that, failure in the 0° direction is dominant3). Therefore,
the damper fatigue performance when subjected to deformation in the 0° direction is
important. On the other hand, in the case of estimating fatigue damage when subjected to
random displacement vibrations, analysis of the amplitudes by the rainflow method and
determination of the cumulative damage by Miner's rule is effective, as in normal
checking of steel members for fatigue4).
To date, in accordance with the required performance, several sizes of U-shaped steel
dampers that are geometrically similar are available. The horizontal performance varies
depending on the size, but a mechanical similarity rule has been established between the
different sizes5). In other words, the deformation scales are in proportion to the
dimension ratio, the resistance force scales are in proportion to the square of the
dimension ratio, and the product of the two scales, which is the energy absorption, is the
dimension ratio to the power of 3. Likewise, for the number of repetitions up to failure,
if the horizontal deformation is normalized by the average shear deformation angle
divided by the height of the damper representing the horizontal deformation (dimension
ratio), a general evaluation of the fatigue performance (number of repetitions up to
@Seismicisolation
@Seismicisolation
failure) independent of the size of the damper can be carried out3).
(3) Various dependencies and durability
Regarding the velocity (frequency) dependence of U-shaped steel dampers, by

comparison of quasi-static and dynamic loading test results using full-size test specimens,
it was found that the hysteresis curves for the two are virtually identical, and it was
confirmed that no significant reduction occurs in the fatigue performance or energy
absorption3). Furthermore, regarding the temperature dependence, the yield resistance
and the energy absorption tend to decrease at high temperatures and increase at low
temperatures. However, when used in the practical range of +40°C to -10°C, the standard
variation is about ±3% for 20°C, so the effect is small6).
Regarding durability, experience in the use of steel members as building materials is not
considered a problem, provided that corrosion prevention paint or similar corrosion
treatment is applied. However, if peeling of the paint is seen with plastic deformation
during an earthquake, it is desirable that the paint be repaired to maintain the durability.
Sufficient plastic deformation performance was confirmed for U-shaped steel dampers in
the Great East Japan Earthquake in March 2011, but some loosening of the installation
bolts was observed, and measures against this were discussed. It is considered that the
effect of loosened bolts on the mechanical properties of U-shaped steel dampers is small.
However, for continued use it is desirable that the bolts be retightened. To ease the
concerns of building owners, it is necessary that some measures be taken to prevent
loosening, while maintaining the original objective that a damaged damper can be easily
replaced.
References
1) Hisami Hasegawa, Hideyuki Tada, Toru Takeuchi, Kunio Hayakawa：Experiment on Large-Capacity Steel Damper
for Base Isolation System, Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan, B,
pp.673-674, 1990.10
2) Hiroaki Konishi, Atsushi Watanabe, Kazuaki Suzuki：Experimental Study of U-shaped Steel Damper part.6，
Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan, B-2, pp.665-666, 2000.9
3) Shoichi Kishiki, Yuta Ohkawara, Satoshi Yamada, Akira Wada：Experimental Evaluation of Cyclic Deformation
Capacity of U-Shaped Steel Dampers for Base-Isolated Structures, Journal of Structural and Construction
Engineering, No.624, pp.333-340, 2008.2
4) Yuta Ohkawara, Masahiro Uekusa, Shoichi Kishiki, Satoshi Yamada, Akira Wada：Dynamic Loading Test of
U-Shaped Steel Damper for Seismic Isolation, Summaries of Technical Papers of Annual Meeting, Architectural
Institute of Japan, B-2, pp.1053-1054, 2007.8
5) Hideaki Yoshikawa, Hiroaki Konishi, Kazuaki Suzuki, Atsushi Watanabe：Experimental Study of U-Shaped Steel
Damper part.5, Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan, B-2, pp.663-664,
2000.9
6) Kazuaki Suzuki, Atsushi Watanabe, Mineo Takayama：Experimental Study of U-Shaped Steel Damper part.3，
Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan, B-2, pp.659-660, 2000.9
@Seismicisolation
@Seismicisolation
2.2.5 Lead Dampers
The energy absorption of lead dampers during deformation is associated with the
elasto-plastic hysteresis of pure lead. Yielding can be promoted earlier compared with
steel members, and it is possible to exhibit damping performance at comparatively small
deformations. Furthermore, the shape can be chosen so that excessive plastic
deformation does not result in a localized concentration of repeated deformation.
Currently, the two types of lead dampers available are the U180 type with lead castings
of 180 mm diameter, and the U2426 type with a 260 diameter (partially 240 mm).
However, in recent years, the U2426 type in particular has been used. Figure 2.2.21
shows the shape and external appearance of a lead damper. The direction parallel to the
kink of the lead casting is referred to as the P direction (parallel direction to the drawing
in Figure 2.2.21), and the direction normal to this is referred to as the O direction
(orthogonal direction to the drawing in Figure 2.2.21).
Steel Flange Plate
Welded Part
Reinforced Part
Flexible Part
Lead Casting
Flexible Part
Reinforced Part
Welded Part
Steel Flange Plate
Fig. 2.2.21 Shape and external appearance of a lead damper (U2426 Type)
The shape of the hysteresis curve of lead dampers is rectangular. Because the
dependence on velocity and temperature is small, lead dampers are frequently modeled
as perfectly elasto-plastic bilinear. The horizontal loading directions have no significant
difference, so it can be assumed that directionality does not affect performance.
Furthermore, when modeled as perfectly elasto-plastic bilinear, the energy absorption of
the lead damper in one cycle is given by the following equation. The design values found
@Seismicisolation
@Seismicisolation
from this equation generally agree with the experimental values. Figure 2.2.22 shows an
example of the restoring force curves when vibrated at a frequency of 0.33 Hz.
dW＝4×Qy×（δ－Qy/K1）（2.2.14）
where, dW: energy absorption in one cycle, Qy: yield load, K1: initial stiffness,
δ: horizontal displacement.
400 400
P-direction
Ｐ方向
U2426
U2426型 type O-direction
Ｏ方向
U2426
U2426型 type
300 300
200 200
Force (kN)
U180
U180型
U180
U180型
Force (kN)
100 100
水平荷重(kN)
Horizontal水平荷重(kN)
0 0
Horizontal
-100 -100
-200 -200
-300 -300
-400 -400
-500 -400 -300 -200 -100 0 100 200 300 400 500 -500 -400 -300 -200 -100 0 100 200 300 400 500

水平変位(㎜)
Horizontal水平変位(㎜)
Displacement (mm)
Fig. 2.2.22 Example of the restoring force characteristics of a lead damper
Regarding the total energy absorption performance, a correlation has been found
between the amplitude and the number of repetitions to failure, and both type U180 and
type U2426 lead dampers can be represented by a single fatigue curve. At an amplitude
of ±400 mm, the total energy absorption for type U180 is about 3000 kNm, and that for
U2426 is about 8000 kNm. Furthermore, the energy absorption in one cycle and the
yield load tend to decrease as the number of repetitions (cumulative amount of
deformation) increases.
It is known that cracks occur before the failure of lead dampers due to fatigue, and that
the number of repetitions at failure is about 30 to 40 times the number of repetitions
when cracks occur.
Furthermore, regarding the evaluation of fatigue damage of lead dampers, methods using
Miner's rule, which is a method of estimating the residual performance from the depth of
cracks and number of repetitions.
The design value for energy absorption in one cycle of a lead damper is defined to be
compatible with the experimental value, so modeling one cycle of energy absorption is
not assumed to be a problem. However, in the small displacement range of 50 mm or
less, the energy absorption in one cycle is slightly reduced compared with the design
value. Furthermore, in large deformation, the peak value of the yield shear force
@Seismicisolation
@Seismicisolation
increases to about 2 to 3 times the design value due to hardening.
It has been found that lead dampers have a dependence on vibration frequency. The
energy absorption in one cycle tends to increase as the vibration frequency increases, and
tends to decrease as the vibration frequency decreases. Regarding the temperature
dependence, the following relation is found between the temperature and the energy
absorption in one cycle. According to this equation, the energy absorption per one cycle
changes by about 3.5% for a change in temperature of 10°C.
R = －0.35T+107 (2.2.15)
Here, R: percentage change in the energy absorption in one cycle (20°C as standard) (%)
T: environmental temperature at the time of the test (°C).
Regarding durability, when lead is exposed to damp air, an oxide film forms on the
surface and its color changes within a few days, but this oxide film has a protective
property so that the rate of oxidation decreases as time passes. As a result of this oxide
film, the rate of corrosion when lead is exposed outdoors is 0.0007 mm per year or less,
and the durability of lead under normal conditions of use is not a problem.
@Seismicisolation
@Seismicisolation
2.2.6 Friction Dampers
Friction dampers apply energy absorption by contact sliding of the friction members.
Some of the various types of friction dampers are the friction disk damper (Photo 2.2.1),
the brake damper type (Photo 2.2.2), and the rotational friction type. The friction disk
damper uses spring disks to maintain a constant axial force in order to stabilize the
friction. They have a mechanism that can set and adjust the damping force (friction
force) by using the load applied to the friction surface.
Photo 2.2.1 Installation condition of a friction disk damper
Photo 2.2.2 Installation condition for a break damper
By appropriately selecting the type of friction material, more stable performance over
many repetitions can be expected than with steel dampers. So, it is possible to provide a
mechanism that can enable a constant friction force without dependence on vibration
amplitude or compressive stress. Furthermore, like steel dampers, friction dampers do
not function as dampers before sliding occurs.
Here, a frictional disk damper is introduced as an example of a friction damper.
A disk damper is a friction damper having a mechanism in which a sliding member (ultra
high molecular weight polyethylene) is pressed against a sliding plate (stainless steel) by
a disk spring having a constant reaction force. The disk damper absorbs seismic energy
@Seismicisolation
@Seismicisolation
by sliding friction. Furthermore, stable frictional damping can be obtained with a simple
structure. Figure 2.2.23 shows a cross section of the disk damper.
⑥
⑤
⑦
① :すべり材（超高分子量ポリエチレン）
Sliding member
② :すべり板（ステンレス）
Stainless steel plate
③ :皿ばね
Coned disc springs
④ :埋込みカラー
Embedding collar
③
: Shaft (Inner guide)
⑤ （皿ばね）シャフト
⑥ :フランジ
Flange plate
④
⑦ :セットボルト（据付後・開放）
Set bolt
⑧ :ソールプレート
Sole plate
⑧ ① ②
Fig. 2.2.23 Basic composition of a disk damper
As shown in Photo 2.2.3, a disk spring is a cone shaped steel plate that is frequently used
to prevent loosening of bolts in other mechanical structures. The relation between the
deflection when a load is applied in the direction that flattens the disk spring and the
reaction force is shown schematically in Fig. 2.2.24. By superimposing the disk spring
shown in Fig. 2.2.23, the displacement range δconst, over which load Pave is virtually
constant, is produced.
Photo. 2.2.3 Coned disk spring
Vertical Load (kN)

荷重(kN)
Complete contact
Pave 全密着
δconst
たわみ(mm)
Vertical Displacement (mm)
Fig. 2.2.24 Relation between load and displacement of a disk damper
An example of the restoring force properties of the disk damper (the relation between
resistance force and sliding displacement) is shown in Fig. 2.2.25. As can be seen in this
@Seismicisolation
@Seismicisolation
figure, the restoring force properties of the friction damper can be considered to be
rectangular shaped, and the restoring force properties can be modeled as perfectly rigid
plastic. Here, the intercept load Qd is defined by equation (2.2.16). Furthermore, the
friction coefficient µ = 0.17 to 0.20 is based on the results of tests on disk dampers.
Intercept load Qd = friction force F = friction coefficient µ × compression load N (2.2.16)
400
水平荷重(kN)
FDD-500-980-S （すべり材No3、皿ばね1～8） 090417-2
300
200
Horixontal Force (kN)
100
-100
-200
-300
水平変位(mm)
-400
-150 -100 -50 0 50 100 150
Fig. 2.2.25 Restoring force properties of a disk damper
In disk dampers, motion energy is converted into heat energy produced by friction. This
heat raises the temperature of the sliding plate and the sliding member, and is transmitted
to the surroundings. In a major earthquake, it is necessary to absorb a large amount of
energy, so the temperature of the sliding plate and the sliding member becomes high, and
the friction coefficient tends to decrease. Furthermore, if the loading is stopped and
started again from the condition in which the temperature has dropped sufficiently, the
friction coefficient is restored to almost the same value as the initial value. Conversely,
the friction coefficient decreases after the loading has stopped. This property is the same
as that of many other friction dampers of different forms. From the results of multiple
repetition tests, the frictional resistance force varied within the range of about 170 kN to
50 kN for 206 loading cycles, but exhibited stable restoring force properties.
Furthermore, the friction coefficient gradually decreased. The range of variation of the
friction coefficient was between about 0.2 to 0.07.
The compressive stress dependence of the sliding member (ultra high molecular weight
polyethylene) of disk dampers shows a gradual decrease in the dynamic friction
coefficient as the compressive stress increases, but by using disk springs the compressive
stress used is virtually constant (standard compressive stress: 4.9 N/mm2). Therefore, it
is generally not necessary to consider compressive stress dependence. Furthermore,
regarding velocity dependence, the dynamic friction coefficient tends to increase as the
@Seismicisolation
@Seismicisolation
velocity increases, and the standard value of the changes as a result of the velocity
dependence tests has been established.
Regarding temperature dependence, the variation in the reaction force of disk springs
over the temperature range -10°C to 40°C is small, and so the variation in sliding load
can be neglected.
@Seismicisolation
@Seismicisolation
2.2.7 Oil Dampers
In Japan, it is said that the first practical damper using a fluid was for hydraulic landing
gear, called the oleo struts of aircraft. Their application to automobiles started about the late
1940s and to railway vehicles about the early 1950s.
From about 1955 onward, oil dampers were applied to the general industrial field to protect
equipment from earthquakes, and their use gradually spread to vibration control devices for
various types of machinery. Specific examples include dampers for spherical-shaped tanks,
and dampers for highway bridges, and they started to be applied for the seismic safety of
buildings from about 1980 onward.
Fluid dampers can be made compact relative to the energy to be processed. They also
provide stable properties over a wide range of vibration types and temperatures.
Furthermore, logical design methods for fluid control devices have been established, and at
present these methods are widely used in large-scale buildings and civil engineering
structures. Normally, by combining two linear type characteristics, bilinear characteristics
are obtained. Hydraulic oil has a small dependency on the temperature in viscosity, except
at extremely low temperatures, so the dependency of temperature is small because of the
damping mechanism.
At present three types of oil dampers are used in the building field, as shown in Figs. 2.2.26
through 2.2.28: (1) the uniflow type oil damper, (2) the biflow type oil damper, and (3) the
biflow type single cylinder oil damper.
Photo 2.2.4 Oil damper
High-pressure
高圧側side Oil tank
ﾀﾝｸ室 Hydraulic
ｵｲﾙ fluid
Proportion valve
調圧弁(開) Low-pressure
低圧側 side
Piston Rod
ﾋﾟｽﾄﾝﾛｯﾄﾞ
Extension
伸び
Relief valve
ﾘﾘｰﾌ弁(閉) Piston
ﾋﾟｽﾄﾝ逆止弁(閉) Intake valve
吸込弁(開)
One-way valve
Fig. 2.2.26 Structure of uniflow type oil damper
@Seismicisolation
@Seismicisolation
Hydraulic
Piston Rod Oil tank fluid Intake valve
Extension
伸び
High-pressure side Proportion Piston Low-pressure side

valve
Fig. 2.2.27 Structure of biflow type oil damper
Pressure chamber (compression

圧縮側圧力 side)
室 Cylinder
シリン tube
ダ Proportion valve
Piston
ピストン Rod
ロッド
調圧弁B 調圧弁Ａ
Extension
伸び調圧弁C Extension
伸び
吸油弁C
吸油弁B 吸油弁Ａ
Piston
ピストン Pressure
伸長側圧 chamber
力室
(extension side)
Hydraulic
オイルリザ reservoir
ーバ
Fig. 2.2.28 Structure of biflow type single cylinder oil damper
The damping performance of oil dampers can be broadly classified into two types: those
with linear characteristics proportional to the velocity and those with bilinear
characteristics. The type used depends on the performance of the seismically isolated
structure and the performance of damping force generation mechanisms such as the
pressure adjustment valve, and the pressure adjustment valve spring. The damping
performance of oil dampers is defined by the value of the maximum capacity of the
damping coefficient that is possible for each model size, and so the damping performance
can be selected as desired within this range. Normally, a Maxwell model or a Voigt model is
used as the analysis model.
The damping force generation mechanism in oil dampers can be classified into three types:
(1) damping mechanism proportional to velocity to the power of 2, (2) damping mechanism
proportional to velocity to the power of 2/3, and (3) damping mechanism proportional to
velocity to the power of 1. By combining these characteristics, the linear characteristics or
bilinear characteristics can be obtained. Furthermore, the hysteresis curve has an elliptical
shape in the case of linear properties, and a rectangular shape in the case of bilinear
properties. Figures 2.2.29 through 2.2.32 show the performance curves and the hysteresis
curves for the linear and the bilinear dampers.
Linear properties are properties in which the damping force (F) is generated in proportion
to the movement velocity (V), and the slope of this curve represents the viscous damping
coefficient (C). With linear characteristics, the viscous damping coefficient (C) is a constant
value over the whole velocity range. This relation is expressed by the following equation.
F＝C・V (2.2.17)
@Seismicisolation
@Seismicisolation
Bilinear characteristics are obtained by a relief valve that starts to operate above an
actuation velocity (denoted as V1 here) in parallel with the pressure adjustment valve for the
linear characteristics. The primary damping coefficient is C1, the secondary damping
coefficient is C2, the relief velocity is V1, and the relief damping force is F1. The damping
force (F) with these characteristics is expressed by the following equation. In the range for
V<V1;
F＝C1・V (2.2.18)
In this range the equivalent viscous damping coefficient is equal to the primary damping
coefficient (C1).
In the range for V>V1;

F＝C1・V1＋C2・(V－V1) (2.2.19)
As stated above, C2 is the secondary damping coefficient. The equivalent viscous damping
coefficient (Ceq), which is equivalent to the energy absorbed in one cycle, is obtained from
the following equation when the displacement is a.
1    V 

2
V 2 − V1 (2.2.20)
Ceq = ⋅ 2 ⋅ (C1 − C 2 ) ⋅ V1 ⋅ + V ⋅ a sin  1   + π ⋅ C 2 ⋅ V 
π ⋅V   V 2
 V  
  
F
Fmax
減
Damping force
衰
力
C
V
Vmax
Velocity
速度
Fig. 2.2.29 Relation of damping force to velocity (linear type)
Fd
ud
Fig. 2.2.30 Hysteresis loop (linear type)
@Seismicisolation
@Seismicisolation
Ｆ
Fmax
C2
減
Damping force
衰 F1
力
C1
Ceq
Ｖ
V1 Vmax
Velocity
速度
Fig. 2.2.31 Relation of damping force to velocity (bilinear type)

Fd
ud
Fig. 2.2.32 Hysteresis loop (bilinear type)
The limit deformation of oil dampers does not depend on their material. Instead, the limit
deformation is determined by the stroke between the mechanical contacts of both sides.
This is different than the limit deformation found with hysteretic dampers, such steel
dampers, lead dampers, and frictional dampers.
Therefore, if the mid-position of the displacement in the horizontal axis direction is taken to
be the neutral point, the total stroke is the difference between the maximum extension
(Lmax) and the maximum contraction (Lmin), and the limiting deformation is defined as half
this value.
Limiting deformation = (Lmax－Lmin) / 2
The limiting deformation that can be manufactured is defined as the maximum value for
each model (size) for each manufacturer, and the desired model can be selected within this
range. If the oil damper is used under the conditions of limit deformation, a mechanical
stopper should be used to prevent damage to the oil damper.
The limiting velocity is the maximum actuation velocity at which the oil damper can
exhibit stable performance. This value is determined so that the performance at the limiting
velocity is not decreased and the damping force generated does not exceed the limiting
allowable strength. The damping force of oil dampers increases in accordance with the
actuation velocity, so the load due to an excessive actuation velocity can cause deformation,
damage, rod buckling, and breakage in the cylinder and can have an adverse effect on the
strength members. Furthermore, at excessive actuation velocities, the flow rate of hydraulic
oil flowing through the pressure adjustment valve increases. As the flow rate increases, the
internal pressure increases and causes faulty operation, such as vibrations in the pressure
adjustment valve, so the performance of the oil dampers becomes unstable. Hence, the
@Seismicisolation
@Seismicisolation
limiting velocity is determined from both strength and performance, and it is necessary to
use oil dampers below the limiting velocity.
The temperature dependence has been confirmed over a range from -30°C to +80°C relative
to 20°C as standard. For temperatures greater than 0°C, the dependence can be ignored.
Regarding frequency dependence, the stiffness of oil dampers is a result of the
compressibility of hydraulic oil. This stiffness acts in series with the damping in a direction
to reduce the damping coefficient. But, for long period (low frequency) seismic isolation
devices, normally the stiffness can be neglected. However, stiffness should be considered
for cases with linear properties with a large primary damping coefficient, where the initial
period of the seismic isolation devices is 1 second or less. Regarding directionality, oil
dampers only act in the axial direction. By using a rotational pin structure in the installation
and by arranging the same number of devices in the X-Y directions, equal damping can be
obtained within a plane.
Dependence on the number of repetitions is normally a function of the wear of the

mechanical components, but the dependence on the number of repetitions shows virtually
no change in the case of earthquake or wind vibrations. Regarding the changes in oil
dampers with age, actual devices including seal members and oil have been in use for over
40 years. Furthermore, from thermal aging tests on seal members, it has been found that
they can perform for more than 100 years at 30°C, and the estimated life of oil from
oxidation stability is about 103 years, so the durability is 60 years or more.
@Seismicisolation
@Seismicisolation
2.2.8 Damping Devices with an Amplification Mechanism (Rotary Damping Tube)
Figure 2.2.33 shows a schematic diagram of a damping device with an amplification

mechanism (rotary damping tube: RDT)1). This device converts linear motion from the
relative displacement into rotaional motion via a ball screw (referred to as reverse
efficiency). Amplifying the velocity of the linear motion in the rotational direction enables
the amplified rotational velocity to act as a viscous material, and so a large damping force is
generated in the damper.
Inner Fluid Rotary Tube

Ball Nut
Outer Tube
Viscous Recirculating Balls
Shaft Bearing Fluid Shaft Bearing
Support Bearing Ball Nut
Screwed Shaft
Seal Screwed Shaft

Seal
Recirculating Balls
Damping Section Transmission
Section Velocity Amplified Section
Fig. 2.2.33 Schematic diagram of a damping device with an amplification mechanism

(Rotary Damping Tube)
(2) Horizontal properties
Rotary damping tube is available as two types, a small viscosity damper (maximum
capacity up to 299kN) and a large viscosity damper (damping force from 300kN to
1400kN). The rear number of RDT means the maximum damping force (unit: ton).
The damping force due to an amplification mechanism is given by equations (2.2.21)

through (2.2.23) in accordance to the type. The two types for dynamic viscosity are
between 100,000 centistokes (cSt) and 300,000 cSt, and high viscosity types are used when
the damping force is large.
Pn=1.03×(α・Qv＋2.7) (100,000 cSt RDT2～RDT20) （2.2.21）
Pn=1.06×(α・Qv＋8.1) (300,000 cSt RDT30～RDT60) （2.2.22）
Pn=1.15×(α・Qv＋11.5) (300,000 cSt RDT70～RDT140)（2.2.23）

where
Pn: damping force (kN) at the maximum velocity 1.5 m/s
α: repeated deformation dependence coefficient
Qv: viscous damping force (kN) Qv = S･η･Vs･A
S: amplification ratio
η: viscosity of the viscous fluid (cSt)
Vs: shear rate (s-1)
A: effective shear area (m2).
@Seismicisolation
@Seismicisolation
The relation between the standard value of damping force calculated from the these
equations and the axial velocity (0.01 to 1.5 m/s) is shown in Fig. 2.2.34 for an example of
300,000 cSt, and the hysteresis curve is shown in Fig. 2.2.35.
2000
(kN) force (kN)

T=20 ℃
1500
damping
RDT140
軸減衰力 RDT125
1000
RDT100
RDT90
Horizontal axial
RDT80
RDT70
RDT60
500 RDT50
RDT40
RDT30
0
0 0.25 0.5 0.75 1 1.25 1.
軸速度
Horizontal axial (m/sec)
velocity（m/s）
Fig. 2.2.34 Reference value of damping force

Horizontal axial damping force (kN)
Horizontal axial displacement (mm)
Fig. 2.2.35 Hysteresis loop
The limit deformation of a damping device with an amplification mechanism is determined

by the stroke from the length of the shaft of the ball screw. The stroke of the screwed shaft
can be arbitrarily set within the limiting range of the buckling load. Furthermore, the
limiting velocity is given as 1.5 (m/s), within which the properties have been confirmed
with property tests2). This is equivalent to a shear rate of 6020 (s-1).
A Maxwell model or a Voigt model is generally used as the analysis model. The following
two calculation methods can be applied for the properties between the damping force and
the axial velocity: first, the standard equation modeled method and second, the modeled
method with a multi-linear dashpot model that can relatively conveniently represent the
standard equations.
Regarding temperature dependence, the viscosity decreases with increasing temperature.

The apparent viscosity in a damping device with an amplification mechanism depends on
the viscosity of the viscous material and the shear strain rate. By using the following
empirical equation, the damping force obtained from a test at an arbitrary ambient
temperature can be converted to the value at a standard temperature of 20°C for evaluation.
@Seismicisolation
@Seismicisolation
η（Vs, t）=ηt／（1+b・Vsβ）×106
where
η(Vs, t): viscosity b=0.000472・ηt0.276 β=0.308・ηt0.081
ηt: viscosity (cSt) of viscous fluid at temperature of t°C ηt=1.02(25-t)×η25
Vs: shear rate (s-1).
In repeated vibration tests, the damping force is reduced by continuous vibrations. This
reduction in damping force is not produced by a breakdown of the molecular structure of
the viscous fluid material, so the damping performance is restored as the stress relaxation
between molecules returns to its original state after several minutes.
References
1）Fumiaki Arima, Tadashi Arakaki, Ryoji Suzuki, et al.：Development of the Seismic Damper with Amplified Mechanism
Part.1~Part.7, Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan, B-2, pp.825-876,
1998～2000
2）Toru Suzuki, Yuji Mitsusaka, Shigeki Nakaminami, et al.：The Efficiency under High Speed Conditions of Viscous
Damping Devices with Amplification System Part.1~Part.3, Summaries of Technical Papers of Annual Meeting,
Architectural Institute of Japan, B-2, pp.759-764, 2003
@Seismicisolation
@Seismicisolation
2.2.9 Viscoelastic Dampers
Viscoelastic dampers consist of a viscoelastic material, which is an acrylic polymer or a

special blend of rubber sandwiched between steel plates. The dampers absorb vibration
energy by obtaining the viscous resistance force by shearing the viscoelastic material.
Viscoelastic dampers have been widely used in the past as vibration control devices, but in
recent years they also have been developed and applied to seismically isolated buildings.
Viscoelastic material is a type of rubber material that exhibits high damping. The damping
performance is usually generated by the following factors:
(1) Viscous damping elements using viscous material
(2) Frictional damping elements between molecules or between molecules and filling
materials
(3) Plastic deformation elements due to resin filling materials
(4) Phase transitions of molecules and blending agents
By sandwiching this material between steel plates, with the dimensions shown in Table
2.2.1 and the structure shown in Fig. 2.2.36, and by applying shear deformation to the
seismic isolation level, a damping force can be generated.
Table 2.2.1 Dimensions of the viscoelastic damper

Long-side direction dimensions 900mm
Short-side direction dimensions 600mm
Height of product 136mm
Total thickness of viscoelastic material 66mm
First shape factor 5.45
Outer plate
端部鋼板
Rubber
ゴム
Inner plate
中間鋼板
Fig. 2.2.36 Viscoelastic damper
@Seismicisolation
@Seismicisolation
Fig. 2.2.37 shows the hysteresis loop of the damper. The basic performance of viscoelastic
dampers, namely the equivalent stiffness K'd and the equivalent damping coefficient Cd, are
calculated from the restoring force characteristics of dynamic loading tests shown in Fig.
2.2.38.
400 10%
300 50%
100%
200 200%
force (kN)
100 300%
荷重（kN）
0
300%
-100
Horizontal
200%
-200 100%
-300 50%
10%
-400
-300 -200 -100 0 100 200 300
Horizontal 変位（mm）
displacement (mm)
Fig. 2.2.37 Hysteresis loop
Fd (u d max )
K d′ =
u d max
∆W
C d=
π ω u d max 2
ω : Angular natural frequency
Fig. 2.2.38 Calculation method of K'd and Cd
The ultimate performance of viscoelastic dampers is determined by the amount of

deformation, the cumulative amount of energy, and changes with age. The effect of the
resistance to ozone or ultraviolet light is considered to be small because the exposed area is
small.
The ultimate deformation is indicated from vibration test results for various strain ranges,
from which the various standard values are set. Furthermore, from the results of repetition
tests at strains of 200% up to 150 repetitions, it was found that the stiffness and equivalent
viscous damping coefficient gradually decrease, but stable loops even after 150 repetitions
were maintained.
At the conceptual design stage, analysis is carried out using K'd and Cd considering the
various dependencies and variation. The normal manufacturing variation is determined
through discussion between the structural engineer and the manufacturer, but the standard
variation for both K'd and Cd is ±20%.
@Seismicisolation
@Seismicisolation
Viscoelastic dampers have various types of dependencies. As the amplitude, frequency, and
temperature increase, the stiffness and damping both decrease. Regarding directional
dependency, both the stiffness and the damping tend to be somewhat larger in the
longitudinal direction than in the transverse direction.
It is considered that the changes with age of the viscoelastic material can be obtained from
a converted equation that relates time and temperature, as derived from chemical reaction
kinetics. In this case the changes in properties with age can be predicted by using the
Arrhenius equation from heated accelerated aging tests. The predicted change in the
equivalent stiffness and the equivalent viscous damping coefficient at the equivalent of 60
years obtained in this way is 0.9 or higher and 0.8 or higher, respectively.
@Seismicisolation
@Seismicisolation
Section 2.3 Combination and Selection of Seismic Isolation Devices
In the design of seismically isolated buildings, determining the arrangement of seismic

isolation devices in the seismic isolation level is an important factor and is similar to the
design of the superstructure. It is necessary to decide what isolators to select, what dampers
to select, and how to combine them after thoroughly understanding the properties of each of
the seismic isolation devices.
Among isolators, a natural rubber bearing is a laminated rubber bearing that uses mainly
natural rubber, which has excellent tensile strength, deformation capability, and creep
resistance. The area of the hysteresis loop is small so that it cannot be expected to function as
an energy absorption device. Therefore, it is necessary to use separate dampers.
On the other hand, a lead plug rubber bearing or a high-damping rubber bearing has large
hysteresis loops, so their energy absorption capability is high. Therefore, it is not necessary to
install separate dampers. To adjust the period properties and the amount of damping of the
whole seismic isolation level, the diameter of the lead plug is adjusted.
Regarding the total rubber thickness of the laminated rubber bearing, some designs have a
constant rubber thickness in each of the diameters adopted, and some designs have a constant
second shape factor of the laminated rubber bearing. In the case of the constant rubber
thickness, the design ensures that the shear strain in each of the laminated rubber bearings is
constant, and the limiting properties of the seismic isolation level are determined by the
laminated rubber bearing with the smallest diameter. In the case of the constant second shape
factor, the design ensures a constant limiting deformation capability in each laminated rubber
bearing, so the limiting properties of the seismic isolation level are determined by the
laminated rubber bearing with the smallest total thickness of rubber layers.
Sliding bearings and roller bearings are used to adjust the period properties, the amount of
damping, and the eccentricity of the whole seismic isolation level. With high friction type
sliding bearings, a damping effect can be produced by hysteresis, but the effect is small with
a low friction type or with roller bearings. Sliding bearings or roller bearings are normally
placed in locations where the fluctuation in the axial force is small during an earthquake.
Some roller bearings are capable of resisting tensile forces, and examples of their use can be
found in super high-rise seismically isolated buildings.
It is necessary to select dampers after evaluating the damping effect with consideration of
their arrangement in the seismic isolation level after understanding the energy absorption
properties, the limiting properties including the fatigue properties, and the various
dependencies typified by the temperature dependence.
@Seismicisolation
@Seismicisolation
Generally, viscous type oil dampers are used as dampers, because they tend to have small
amplification of the higher modes of the superstructure compared with the use of hysteretic
type dampers. Furthermore, it is difficult for the response acceleration of the superstructure
to be amplified due to the effect of stiff elements and viscous elements in the seismic
isolation level; the phase difference of these elements means that their maximum values are
not generated at the same time. When the effect of wind loading is large, such as in super
high-rise seismically isolated buildings, it is necessary to provide devices with both stiffness
and load resistance with respect to displacement, as typified by hysteretic dampers, so that
the response displacement with respect to the average component of the wind is not too large.
In the same way that a structural engineer determines the structural shape and the size of the
members in the superstructure, the structural engineer should determine the overall suitability
of the seismic isolation devices. After thoroughly understanding the information in the
technical documents by the manufacturers of the seismic isolation devices, the structural
engineer should select the seismic isolation devices. Frequently, economics is a major factor
in the selection of seismic isolation devices in actual design.
When setting the properties of the seismic isolation level, it is essential to predetermine the
response of the seismically isolated building during an earthquake. It is considered that the
seismic response is determined by period properties due to the stiffness of the isolators and
the damping properties of the dampers. It is necessary to evaluate the response considering
the properties of the input seismic motions and the properties of the superstructure.
The seismic isolation period Tf of seismically isolated buildings can be obtained from the
relation of the secondary stiffness Kf of the seismic isolation level, the total rubber thickness
of the laminated rubber bearing, the rubber properties (shear modulus), and the average
compressive stress, as follows.
2π
T f= 　
W , G A
K f = 　r r , W
σ 0= 　
gK f hr Ar
σ 0 hr σ 0D (2.3.1)
2π
T f= 　 = 2π
gGr gGr S 2
where Tf: seismic isolation period, W: total weight of the superstructure, Kf: horizontal
stiffness of the seismic isolation level (secondary stiffness), Gr: shear modulus of the
rubber, Ar: total cross section of rubber bearings, hr: total thickness of rubber layers, σ0:
average compressive stress, D: diameter of rubber bearings, S2: second shape factor.
Figure 2.3.1 shows the relation between the diameter of the devices and the seismic
isolation period Tf. In the figure, the relation is shown for various values of σ0/S2.
@Seismicisolation
@Seismicisolation
Tf(sec)
10
Gr=0.392(N/mm2)
9
8
6.0
7
5.0
6
4.0
5
3.0
4 2.0
3
1.0
2
σ o/S2 (N/mm2)
1
600 800 1000 1200 1400 1600
Diameter　D(mm)
Fig. 2.3.1 Relation of isolation period to diameter of rubber bearing
The shear force coefficient aISO of the seismic isolation level can be represented by the
following equation in the case that the seismic isolation level has bilinear properties, as
shown in Fig. 2.3.2.
Q Qd + K f d K fd 4π 2 d (2.3.2)
α ISO = 　 ISO = = αs + = αs +
W W W g Tf 2
where as = Qd/W.
QISO
Kf
Qd
Fig. 2.3.2 Restoring force characteristics of isolation level
Figure 2.3.3 shows the relation between the deformation of the seismic isolation level and
the shear force coefficient of the seismic isolation level obtained by using equation (2.3.2).
From Fig. 2.3.3, it can be seen that when the yield shear force coefficient of the seismic
isolation level is 0.03 and by assuming the seismic isolation period is 4 seconds and the
deformation of the seismic isolation level is 40 cm, the seismic shear force coefficient of
the seismic isolation lelvel is about 0.13.
From the above, the shear force coefficient of the seismic isolation level can be calculated
from the yield shear force coefficient of the seismic isolation level and the seismic isolation
period for an assumed deformation of the seismic isolation level. The deformation of the
seismic isolation level can be predicted from response prediction methods based on a
simple energy balance or the displacement response spectrum, if the input seismic motion
has been set. Therefore, it is possible to determine the necessary properties of the seismic
isolation level in the initial stages of the seismic isolation design.
@Seismicisolation
@Seismicisolation
aαISOISO
0.30
a
αs=0.03
Tf
Tf=3.0(s)
S
0.25
0.20
0.15 4.0
5.0
0.10
6.0
0.05
0.00
0 10 20 30 40 50
δ(cm)
Fig. 2.3.3 Relation of shear force coefficient to deformation of isolation level
@Seismicisolation
@Seismicisolation
Section 2.4 Seismic Isolation Devices and Systems for Detached Housing
Seismic isolation devices for detached housing can be classified into rubber bearing types,
roller types, and sliding types. However, the building mass of detached housing is very small,
so special measures (such as multi-layered rubber bearings, and doughnut shaped rubber
bearings) are necessary to obtain a sufficient seismic isolation period when using rubber
bearing devices. Therefore, roller type and sliding type devices are frequently used in the
present.
Figures 2.4.1 and 2.4.2 schematically show the seismic isolation systems for housing. The
roller type device has a structure in which a rolling device is supported by a rail or dish, and
the rolling friction is small. To provide the restoring force and damping, some oil dampers,
lead plug rubber bearings, or high-damping rubber bearings are used. In the sliding device, a
sliding device such as PTFE or a similar material is provided on a stainless steel plate or dish,
thereby supporting the mass of the building and at the same time providing damping by
friction.
Lead rubber bearing or High

Slope of dish Wheel or Steel ball damping rubber bearing
Friction of
bearing
Oil damper Inclined rail or dish Roller bearing
Fig. 2.4.1 Example of roller isolation system

Slider Slider Elastic body to support the vertical load
Arc-shaped dish
Stainless plate Elastic body that do not support the vertical load
Fig. 2.4.2 Example of sliding isolation system
Figure 2.4.3 shows the restoring force properties of roller type devices. If the support surface
is a circular conical shape, the resistance force is equivalent to the sum of the slope angle θ of
the circular conical surface and the coefficient of rolling friction µ. In roller bearings that
move on a plane such as a linear guide, the resistance force is equivalent to the coefficient of
friction. The frictional resistance in each case is very small, so it is necessary to separately
provide dampers.
@Seismicisolation
@Seismicisolation
F F
δ δ
: Supported Load
(a) case of cone-like support face (b) case of plane support face
Fig. 2.4.3 Restoring force characteristics of roller isolation system
Figure 2.4.4 shows the restoring force properties of sliding type devices. If the sliding surface
is curved (radius of curvature R), a restoring force is generated with stiffness W/R. If the
sliding surface is a plane, the restoring force properties are the rigid plastic type. A damping
effect due to friction can be exhibited, but in the case of a plane sliding surface it is necessary
to separately provide a device to provide the restoring force.
F F
δ δ
: Supported Load
: Radius of curvature
(a) case of cone-like support face (b) case of plane support face
Fig. 2.4.4 Restoring force characteristics of sliding isolation system
Since the mass of detached housing is small, frequently the wind loading due to weather
events such as typhoons, is similar to the shear force during an earthquake. If the seismic
isolation performance is increased, oscillation of the building during strong winds such as
those associated with typhoons is possible. Therefore, a trigger function for strong winds or
installation of a protection device may be deemed necessary.
When selecting the seismic isolation device, it is necessary to confirm the following issues:
• Whether the vibrations felt are generated by normal winds (winds with a return period of 1
year or less)
• How much is the response acceleration and displacement if an earthquake occurs, when
the fixing protection function for housing is acting during strong winds
• Whether the building is reliably fixed and released. In addition, what is the effect on the
building when fixing and releasing function does not work.
The seismic isolation performance of detached housing is governed by the performance of
@Seismicisolation
@Seismicisolation
the seismic isolation devices as well as by the buildings. It is necessary to appropriately

select seismic isolation devices in accordance with the properties of the superstructure and
the predicted level of seismic motions. For seismic motions that exceed the predicted level,
damage to the seismic isolation devices or reaching the limiting deformation can occur. To
limit deformation, providing a displacement control mechanism (stopper) can be considered.
When the displacement control mechanism operates, a large acceleration will be transmitted
to the superstructure. It is necessary to consider what the effect on the superstructure will be
if this occurs.
So that the various seismic isolation devices exhibit their function and the building exhibits
the set seismic isolation performance, it is necessary that each of seismic isolation devices
deforms integrally. Therefore the seismic isolation platform for fixing the seismic isolation
devices should be requied to be composed of the reinforced concrete or structural steel with
sufficient stiffness.
@Seismicisolation
@Seismicisolation
Section 2.5 Design and Construction for Installation of Devices to Structures

and for Fire Resistant Covering
2.5.1 Design and construction for structures with installed devices
The horizontal displacement of the seismic isolation level during a major earthquake may be
several tens of centimeters. In this case additional stresses are applied around the seismic
isolation devices, so it is necessary that the connections and the surrounding structure have
the appropriate strength and stiffness.
Reaction force distribution

High pressure area
Steel plate
Rubber (a)Compressive loading
Reaction force distribution

Bending moment
(b)Compressive shearing
Resultant of reaction force
(c)at the large deformation
Fig. 2.5.1 Schematic diagram of stress distribution of rubber bearing
Figure 2.5.1 shows the stress distribution associated with horizontal deformation of a rubber
bearing. The rubber bearing itself resists the bending moment when subjected to a horizontal
displacement by moving the position of the reaction. This is because the rubber is in a
hydrostatic state, and the pressure distribution in the rubber varies in accordance with the
horizontal deformation. So, even when subjected to a large deformation in the horizontal
direction, the mechanism that automatically cancels out the moment acting on the rubber
bearing operates. Therefore, the axial force acting on the bolts fixing the rubber bearing does
not become very large over the design range of displacement1).
However, an additional bending moment is produced on the structure (superstructure and

foundation structure) to which the rubber bearing is attached in accordance with the
horizontal deformation δ of the seismic isolation level. When considering the bending
moment, the following two types of bending moment must be recognized: the additional
@Seismicisolation
@Seismicisolation
bending moment produced by the axial force of the columns (P × δ/2) and the bending
moment due to the shear force produced in the rubber bearing (Q × h/2). These bending
moments are processed by the beam, column, or pile above and below the rubber bearing. If
the degree of fixity of a rubber bearing is low, it is known that the buckling stress of the
rubber bearing is reduced, so if the rubber bearing is installed on a column head or pile head,
it is also necessary to pay attention to the rotational stiffness of the structures with the
installed bearings2).
The stress distribution in the rubber bearing is parabolic, that is, higher in the center. The
compressive stress is the average value obtained from the supported axial load divided by the
cross-sectional area, and that stress is about double the compressive stress (average stress)
acting at the center. During an earthquake, an axial load that is higher than the long-term load
should be supported by a smaller cross-sectional area (the area of the overlapping portion of
the upper and lower surfaces of the rubber bearing). When the horizontal deformation is
about half the diameter of the rubber bearing, the effective support area is 40% of the
cross-sectional area, so the generated stress increases by more than a factor of 2.5. If the
structure to which the rubber bearing is attached is a reinforced concrete (RC) structure, it is
necessary to sufficiently consider the extent of stress that can be transmitted with the
structure in a triaxial compression state, the amount of the constraint effect that can be
provided by the reinforcement in the design, and whether the construction of the installation
portions for the rubber bearing can support a high axial load.
In the same way, for rubber bearings as well as sliding bearings and roller bearings, it is
required that the surrounding structure be designed so that it is capable of sufficiently
withstanding stresses due to axial loads, shear forces, and bending moments. In the case of
sliding bearings, it is necessary that the structure with the attached sliding plate can withstand
the additional bending moment (P × δ) due to axial loads. In the case of roller bearings
(linear motion type), the additional moment (P × δ) acts on the structure on both the
foundation side and the superstructure side. When tensile forces act on the isolators due to
horizontal deformation, it is necessary that the connections be designed to sufficiently resist
the tensile forces.
When dampers are provided independently from the isolators, additional stresses are
generated around the dampers during horizontal displacement due to bending moment, shear
force, and axial loads, and so the connections and the surrounding structure should have the
necessary strength and stiffness.
When there are defects in the design or construction for only portions with installed devices,
the structure side will be damaged and the dampers will not function, as seen in actual
seismically isolated buildings after the 2005 West of Fukuoka Prefecture Earthquake3).
@Seismicisolation
@Seismicisolation
Reinforcements, anchor bolts of the base plate and jigs for adjusting the height are provided
in the structure to which seismic isolation devices are attached, so it is also necessary to
sufficiently check the filling and compaction of concrete. Sufficient care is particularly
required regarding the horizontal accuracy of isolators. Furthermore, care should be taken to
ensure that the seismic isolation devices are not damaged during construction. Particularly,
care should be taken regarding dirt or rainwater on the sliding surface of sliding bearings and
the rails of roller bearings.
One advantage of seismically isolated buildings is that displacement and energy absorption is
concentrated in the seismic isolation level, and the seismic isolation devices can be replaced.
Generally, it is easy to replace dampers that are independent of the isolators, and so a
structural system with devices having a low residual performance can be replaced with new
devices.
Isolators support the building mass, so replacement is not as easy compared with that of
independent dampers. Some seismically isolated structures that have already been
constructed in Japan have been reinforced in anticipation of jacking up the whole or a part of
the building to replace the isolators. Replacement of seismic isolation devices is a matter that
should be comprehensively decided between the building owner and the structural engineer,
based on the cost of the initial reinforcement, architectural restrictions, and cost of
replacement.
2.5.2 Fire Resistant Coverings
In the case of base isolation, which a seismic isolation level is on the top of foundation,
basically there is no need for the seismic isolation devices to have a fire resistant covering.
On the other hand, in the case of intermediate level seismic isolation (see Section 4.6.3), it is
necessary that the seismic isolation devices have a fire resistant covering. In particular,
isolators are regarded as columns in normal buildings, so fire resistant covering is needed.
Fire resistance tests were carried out on natural rubber bearing4). The test specimens were a 1
m diameter rubber bearing and a 1.6 m diameter lead rubber bearing. At the start of heating,
cracking occurred on the surface of the rubber, and thereafter ignition occurred at the ignition
temperature (420°C). After further heating, the surface of the rubber started to peel off and
emitted black smoke. From the results of these tests, when subjected to a fire without a fire
resistant covering, the 1 m diameter rubber bearing did not lose its load bearing capacity in
about 2 hours, and the 1.6 m diameter rubber bearing did not lose its load bearing capacity in
more than 3.5 hours (with a compressive stress of 5 MPa). From these test results it is
possible that a bare rubber bearing may be unable to support a load in 2 hours or less.
@Seismicisolation
@Seismicisolation
Fire resistant coverings for rubber bearing have already been developed and applied5),6),7).
Research and development has been carried out on fire resistant coverings for elastic sliding
bearings and roller bearings8),9),10), and approval of the Minister of Land, Infrastructure,
Transport and Tourism has been obtained.
Basically, dampers are not required to have fireproof performance. However, if fluid type
dampers are used in intermediate level seismic isolation, it is necessary to consider the effect
that dampers will have on fire performance of the building in the event that a fire occurs.
References
1）Mineo Takayama, Keiko Morita, et al.：Tensile Force on Fixing Bolts of Elastomeric Isolator Under Shear Deformation,
Journal of structural engineering, Vol.49B, pp.129-142, 2003.3
2）Mineo Takayama, Keiko Morita：Finite Element Analysis of Rubber Bearings including Flange Plates and Bolts,
Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan, pp.511-514, 1998.9
3）Architectural Institute of Japan：Report on the Damage Investigation of the 2005 West off Fukuoka Earthquake, 2005
4）Asao Kato, Masaaki Oba, Shintaro Michikoshi, Yutaka Kobayashi, Teruya Minami：Study on Structural Design for
Seismic Isolated FBR：Part 28. Fire Test Results of Rubber Bearings, Summaries of Technical Papers of Annual Meeting,
Architectural Institute of Japan, B-2, pp.1215-1216, 1999.9
5）Minoru Koyama, Osamu Hasegawa, Ryutaro Yokoyama：Fire Resistance Test of Fire-protected Laminated Rubber
Bearings，Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan, B-2, pp.573-574, 1998.9
6）Kiyoji Izawa, Yoshihiro Onishi, Osamu Tanaka, Masashi Fuji, Norio Kawai, Yasuhiro Miyaguchi：Development of
Multi-Layered Sliding Fireproof Cover of Rubber Bearing for Base Isolation, Summaries of Technical Papers of Annual
7）Koichiro Takahashi, Nagao Hori, Akira Teramura：Study on Fireproof Cover System of Rubber Bearing for Base Isolation
part.1~part.2, Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan, A-2, pp.7-10, 1999.9
8）Hisaya Tanaka, Motohiro Shimizu：The Fire Resistance Test of Steel Ball Bearings (SBB), Summaries of Technical Papers
of Annual Meeting, Architectural Institute of Japan, B-2, pp.651-652, 2001.9
9）Hisaya Tanaka, Ryouji Suzuki, Osamu Tanaka：Experimental proof of Cross Linear Bearing (CLB)：Part.2 The Fire
Resistance Test, Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan, B-2, pp.781-782,
2005.9
10）Keizou Mizuno, Tomonobu Isobe, Shintaro Michikoshi, Shinsuke Kuramoto, Kiyohide Seki：Development of Fire
Protection System for Slide Bearings (Part.1~Part.3), Summaries of Technical Papers of Annual Meeting, Architectural
Institute of Japan, A-2, pp.309-314, 2008.9
@Seismicisolation
@Seismicisolation
Section 2.6 Quality and Durability of Seismic Isolation Devices
2.6.1 Ensuring Quality of Seismic Isolation Devices
To obtain the quality required for seismic isolation devices, mechanical properties such as
stiffness, load resistance, damping performance, and deformation capability, must be in
accordance with the design values, the range of variation of these values must be small, and
the seismic isolation devices must have high durability. The range of variation includes
dependencies on axial stress, temperature, frequency, and number of repetitions. When a
seismically isolated building is subjected to an earthquake or strong winds, it is important
that the fluctuation in the response of the seismically isolated building is as small as possible
with respect to the estimated load fluctuations and repeated deformation of isolation devices.
Structural engineers should recognize that seismic isolation devices are structural members in
the same way as columns and beams. So, seismic isolation devices should be selected,
designed, and specified only after studying and researching the seismic isolation devices.
Based on the specifications, manufacturers should control the method of manufacture of the
seismic isolation devices so that the quality is within the specified range. It is necessary that
engineers are sufficiently aware of the manufacture, performance, and characteristics of
seismic isolation devices.
Therefore, tests are necessary to confirm the mechanical properties of seismic isolation
devices.
Performance inspection on seismic isolation devices includes prequalification tests carried

out before the seismic isolation devices are actually adopted, and product performance tests
are carried out before installation of the seismic isolation devices in a building.
Prequalification tests include basic property tests, ultimate performance tests, durability tests,
and tests for dependencies such as axial stress, velocity, and temperature that are necessary to
provide data for design and modeling. Prequalification tests are mainly carried out by the
manufacturers.
Product performance tests include prototype tests and quality control tests. Prototype tests are
carried out on the actual product to be used to confirm the results of the property evaluation
tests. It is desirable that loading tests up to the design displacement and ultimate performance
tests are carried out. Quality control tests are carried out for variations in the stiffness, and to
check for defects. Basically, 100% of available products are tested, including material tests
and dimensional inspection.
@Seismicisolation
@Seismicisolation
Ideally, the testing is carried out on the actual velocities and actual displacements of full-size
products. However, at the present, no testing machines in Japan are capable of carrying out
these types of tests. In the case of small-size products, it is possible to reproduce the actual
characteristics, but it is difficult to carry out tests in Japan on large diameter rubber bearings
and other products. Therefore, tests are carried out on scale model specimens, or by applying
static forces. The engineer is required to estimate the performance of the actual product from
the results of these types of tests.
In August 2011, "Elastomeric Seismic-protection Isolators for Buildings -- Part 1

Specifications (JIS K6410-1)" and "Elastomeric Seismic-protection Isolators for Buildings --
Part 2 Test Methods (JIS K6410-2)" were adopted for rubber bearings. However, rubber
bearings have not yet been certified by JIS (Japanese Industrial Standard). It is desirable that
the specification and testing methods of seismic isolation devices are formulated by JIS to
enable consistent evaluations to be obtained. At present, however, it is necessary for each
structural engineer to sufficiently confirm the performance of seismic isolation devices.
2.6.2 Durability and Maintenance of Seismic Isolation Devices
It is necessary to appropriately evaluate the durability of seismic isolation devices in

accordance with the type of device and the material used. After appropriately evaluating the
change over time of the materials from which the seismic isolation devices are made and
from the amount of creep due to vertical loads, it is necessary to confirm that the seismic
isolation effects can be maintained over the long term. Two examples of long-term use of
rubber bearings are as follows: one is the elevated railway bridges in Melbourne, Australia1),
and another is the Pelham Bridge in the UK2). The degradation condition of these laminated
rubber bearings has been reported. Virtually no change in the rubber layers occurred in the
center, although the failure strengths and failure in the elongation of the rubber on the surface
decreased and the hardness and the elastic modulus increased.
Long-term change is often predicted by heated accelerated aging to evaluate the long-term
durability. In Japan it is generally estimated that 60 years is a guide to the durability of
laminated rubber bearings. In accelerated aging tests the environmental temperature was
assumed to be 20°C or 15°C, and the accelerated aging temperature was between about 70°C
to 100°C for several tens of days. These accelerated aging tests were carried out under the
condition that the environmental conditions are constant, so it should be noted that the actual
environmental conditions of the seismic isolation levels are different. In addition, in almost
all cases, the test specimens used in accelerated aging are small-scale models or blocks of
rubber, so that the shape or scale effects remain.
@Seismicisolation
@Seismicisolation
Tests on the change in properties depending on the age of a laminated rubber bearing have
been carried out as part of the research into seismic isolation of a nuclear power plant. In
tests carried out using a lead rubber bearing and a high-damping rubber bearing with
diameters of about 800 mm and subjected to the accelerated aging equivalent to 60 years3), 4),
it was reported that the horizontal stiffness increased by a maximum of 15%, and the
displacement at failure was reduced by about 5% compared with test specimens that were not
aged. Furthermore, in tests using a rubber bearing with a diameter of 500 mm that were
subjected to accelerated aging tests equivalent to 40 years5), 6), it was reported that the
horizontal stiffness when subjected to a shear strain of 200% increased by 15% to 22%, the
vertical stiffness increased by 7%, and the failure shear strain was reduced by 20% compared
with that of test specimens that had not been aged. The accelerated aging conditions in each
case were different, so it was not possible to simply compare the results. It is necessary,
however, to perform the design considering the change in property values with time.
Heated accelerated aging tests is a method of predicting the long-term condition in a short
time. In order to show the validity of this method, it is necessary to observe actual
seismically isolated buildings and to continuously evaluate separately placed test specimens.
It was shown by using long-term monitoring results7),8),9) that the change in properties and the
amount of creep in a laminated rubber bearing over about 20 years in an actual building were
within the range of the change in properties predicted in the initial design. The data obtained
from this monitoring is valuable, but the initial design conditions and the current method of
use of seismic isolation devices changed. It is necessary that this type of survey be
continuously carried out on as many buildings as possible. Furthermore, it is necessary that
surveys be carried out on the degradation with age of the seismic isolation devices other than
the laminated rubber bearing.
Formerly, "separately placed test specimens" were placed in the seismic isolation level. It
was difficult to remove the actual seismic isolation devices, so the separately placed test
specimens were removed, and tests on their condition of degradation were periodically
carried out10). Recently, manufacturers periodically have carried out tests on separately
placed test specimens at factories. However, the number and type of test specimens are not
sufficient, so issues such as dealing with cases where the rubber material or the shape is
different, or who should judge the evaluation results for changes with time are not clear.
Besides the changes with time in the constituent materials of the seismic isolation devices, it
is necessary to study the durability when seismic isolation devices are subjected to repeated
deformation during long period long duration seismic motions or strong winds. In seismic
isolation devices in which damage has accumulated as a result of repeated deformation over a
long duration of time, it is necessary to evaluate the damage or to evaluate the residual
performance. If seismic isolation devices have been subjected to large deformation, it is
@Seismicisolation
@Seismicisolation
necessary to consider to what extent the initial performance can be ensured, and the criteria
for judging when to replace the seismic isolation devices should also be considered at the
design stage.
For maintenance of seismically isolated buildings, it is necessary to carry out periodic

inspection, such as checking that obstacles have not been placed around the seismic isolation
clearances, and that the condition of the seismic isolation devices has not changed. As part of
the maintenance management, after completion of the construction inspection, usually a
periodic inspection is carried out, including inspection within the seismic isolation level. It is
desirable that the timing of the periodic inspection and the inspection system be appropriately
set. In emergency inspections carried out after being subject to an earthquake, the seismic
isolation clearances are inspected and an inspection is carried out to determine whether
deformation or damage occurred in the seismic isolation devices.
Although during the 2011 Great East Japan Earthquake, seismically isolated buildings
sufficiently confirmed their effectiveness; it became apparent that the maintenance and
operation management of the seismic isolation devices was insufficient at some buildings.
For example, corrosion was seen in the installation portions of the seismic isolation devices,
the bolts installing steel dampers were loose, and small cracks were seen in the lead dampers.
The hysteretic dampers changed shape plastically by absorbing the seismic input energy. In
this earthquake, although each building exhibited the seismic isolation effect, daily
maintenance and inspection were necessary. To check whether a seismically isolated building
sufficiently exhibited the desired effectiveness during an earthquake and to determine the
extent of damage to the seismic isolation devices, it is essential to know the response during
an earthquake.
Consequently, it is desirable that a seismometer be installed in seismically isolated buildings.

If this is difficult, it is important to install a scribed type displacement meter in the seismic
isolation level to confirm deformation of the seismic isolation level. So that seismically
isolated buildings can sufficiently confirm their function, engineers, building operators, and
building occupants should be required to decide in advance whether a system in a seismically
isolated building needs maintenance.
References
1） Noboru Yamazaki， Mitsuo Miyazaki： Characteristics of Normal Rubber Bearing after using for 100 years，
Recommendation for the Design of Seismically Isolated buildings 2nd Ed., Architectural Institute of Japan, pp.311-312,
1993
2) Asao Kato, Masaaki Ohba, Genji Yoneda, Tsutomu Hirotani：Study on Structural Design for Seismic Isolated FBR
Plant: Part.7 Aging Characteristics of About Forty Year Old Laminated Rubber Bearings, Summaries of Technical Papers
of Annual Meeting, Architectural Institute of Japan, B-2, pp.1135-1136, 1996
3）Taiji Mazda, Masahumi Moteki, Katsuhiko Ishida, Hiroo Shiojiri：Test on Large-Scale Seismic Isolation Elements ( Part.1
@Seismicisolation
@Seismicisolation
Accelerated Aging Test on Lead Rubber Bering), Summaries of Technical Papers of Annual Meeting, Architectural
Institute of Japan, B , pp.611-612, 1991.9
4）Masafumi Moteki, Taiji Mazda：Test on Large-Scale Seismic Isolation Elements：Part.1 Accelerated Aging Test on
High-damping Rubber Bearing, Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan, B,
pp.613-614, 1991.9
5）Taiji Mazda, Yasuki Ohtori, Kazuta Hirata, Katsuhiko Ishida：Evaluation of Long-Term Durability of Natural Rubber
Bearing by Accelerated Aging Test, Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan,
pp.717-718, 1996.9
6）Kazuta Hirata, Yasuki Ohtori, Katsuhiko Ishida,Taiji Mazda：Evaluation of Long-term Durability of Lead Rubber Bearing
by Accelerated Aging Test, Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan, B-2,
pp.719-720, 1996.9
7）Hiroki Hamaguchi, Satoru Aizawa, Yusuke Samejima, et al.：A Study of Aging Effect on a Rubber Bearing After About
Twenty Years in Use, Journal of Technology and Design, AIJ, No.30, pp.393-398, 2009.6
8）Kenji Yasui, Kunio Hayakawa, Satoshi Yamagami：Aged Deterioration of Natural Rubber Bearing Installed in the
Building (Structure), Journal of Technology and Design, AIJ, No.24, pp.167-170, 2006.12
9）Tomotaka Wake, Mitsuru Miyazaki：Survey of Aging for LRB Completed 20 Years Ago, Summaries of Technical Papers
of Annual Meeting, Architectural Institute of Japan, pp.391-392, 2008.9
10）Masayuki Yanagi, Miyuki Kaihotsu, Mineo Takayama, Keiko Morita, Kenji Yasui, Satoshi Yamagami：Study of Aged
Deterioration of Natural Rubber Bearing: Part.1： Aged Deterioration of Natural Rubber Bearing, Part.2 Creep
Characteristics，Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan, pp.399-400, 2010.9
@Seismicisolation
@Seismicisolation
Chapter 3 Design – General
Section 3.1 History of Seismically Isolated Buildings
The history of seismically isolated buildings in Japan commenced about thirty years ago in
1983 with the Yachiyodai housing. Table 3.1.1 classifies time periods according to the
changes in seismically isolated buildings and shows the characteristics in each time period.
With the passage of time, the types of isolators and other seismic isolation devices have
diversified, and their performance has improved greatly. As a result, the seismic safety of
seismically isolated buildings has progressed considerably, and their capabilities have been
demonstrated in several major earthquakes, as described in the next chapter.
Table 3.1.1 Features according to age of a seismically isolated building

Classification Age Characteristics of seismically isolated buildings
First period ～1988 ・Seismically isolated building by pioneer
(Genesis) ・Application to facilities of major construction companies
・Natural laminated rubber bearing & damper
・Rubber material G4.5 & G6 were mainstream
・The consideration by the standard observed wave of 0.25 and 0.5 m/s
Second period 1989 ・Effectively utilize of the seismic isolation level
(Before Kobe ～1994 ・Condominium apartment by seismic isolation system
Earthquake) ・Intermediate level seismically isolated building
・Examination of margin of design.
Third period 1995 ・Rapid increase of seismically isolated building
(After Kobe ～1999 ・Diversification of applications
Earthquake) ・Diversification of seismic isolation system
・Rubber material G4 were mainstream
・Increasing the compressive stress of isolators
・longer the seismic isolation period
・The consideration by the standard simulated earthquake motions.
Fourth period 2000～・Design method by the Minister is unnecessary，because the design
(Revised Building standard was beginning
Standards Law) ・The Minister approved seismic isolation devices as building materials
・Application to high rise seismically isolated building
・The consideration by 3 waves of simulated earthquake motions & 3
waves of standard observed earthquake records.
@Seismicisolation
@Seismicisolation
The number of buildings annual plans Cumulative number of buildings

300 3000
Cumulative number of buildings：2,788（2010）
250 2500
200 2000
150 1500
100 1000
50 500
0 0
year
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Fig. 3.1.1 Number of seismically isolated buildings per year
Next, the number of seismically isolated buildings designed according to year and the
cumulative number of buildings constructed are shown in Fig. 3.1.1 based on the database of
the Japan Society of Seismic Isolation (JSSI). From this graph, it can be seen that during the
first and second periods prior to the Southern Hyogo Prefecture Earthquake (Kobe
Earthquake) in 1995, the number of seismically isolated buildings per year was extremely
few, but after the Kobe Earthquake, the number of buildings rapidly increased to about 150
buildings per year. In recent years, many seismically isolated buildings have been designed
and constructed; these are mainly disaster preparedness facilities that will serve as bases
during disasters. New seismic isolation technologies have recently been developed and
designed, such as super high-rise seismic isolation, pile head seismic isolation,
three-dimensional seismic isolation. In this way, seismic isolation technology has been
evolving day by day, so it is extremely important that the latest knowledge is collected,
researched, and surveyed in order to design better seismically isolated buildings.
@Seismicisolation
@Seismicisolation
Section 3.2 Effectiveness of Seismically Isolated Buildings
Since 1995, many more seismically isolated buildings have been designed. This was simply
because it was the first time that the seismic safety of seismically isolated buildings was
verified on actual buildings in a major earthquake such as the Kobe Earthquake. Figure 3.2.1
shows an overall view of the WEST Building (computer center) with seismic isolation. This
building was located 35 km northeast of the epicenter of the Kobe Earthquake, and the
measured seismic record of the earthquake is shown in Table 3.2.1. In a normal
non-seismically isolated building, it is considered that the horizontal acceleration on the
uppermost floor is amplified by a factor of two to three compared with the foundation. As
can be seen from the results, the horizontal seismic response acceleration was dramatically
reduced in the WEST Building compared with a non-seismically isolated building.
In the case of a seismic-resistant building, the structural members are damaged (plastically
deformed) to absorb energy and prevent toppling and collapse, but it is hard to maintain the
function of the building after a major earthquake. On the other hand, in a seismically isolated
building, the seismic energy is mostly absorbed in the seismic isolation level, the structural
members are not damaged, and collapse and toppling of non-structural members (ceilings,
external walls) and fixtures within the building are prevented. Thus, the building function is
maintained after a major earthquake. In other words, a seismically isolated building not only
protects human life, but can also preserve the property. Because very little repair of
earthquake damage is needed, the lifecycle costs are reduced and economics are improved.
Furthermore, the demand is growing for longer-life buildings from the viewpoint of
conservation of the global environment, and this entails a high-performance structural design
system for an earthquake country such as Japan. In seismically isolated buildings, basically
almost all the seismic energy is absorbed by the seismic isolation level, so the shear force and
inter-story drift for design of the superstructure are small. This enables a reduction in the size
of the structural members compared with a conventional building and simplification of the
attachments of non-structural members.
Based on the benefits of seismically isolated buildings, as described above, seismic isolation
is being adopted mainly to provide disaster prevention facilities that will serve as bases
during disasters, and also to provide a wide range of structural forms and applications
ranging from detached housing to super high-rise buildings. Furthermore, apart from new
buildings, seismic isolation is being adopted as an extremely effective method of seismic
retrofit of existing buildings in order to ensure seismic safety, and the construction can be
carried out while the building is being used.
@Seismicisolation
@Seismicisolation
Fig. 3.2.1 WEST building (computer center)
Table 3.2.1 Maximum acceleration of WEST building (unit: cm/s2)

Direction
Seismic observation point
East and West North and South Vertical
6th floor 103 75 377
1st floor 106 57 193
Foundation 300 263 213
@Seismicisolation
@Seismicisolation
Section 3.3 Points to Note Regarding Planning and Design
A seismically isolated structure is a system in which the natural period is lengthened by the
use of isolators such as rubber bearings, and the seismic energy is absorbed by dampers. As a
result of the lengthening of the natural period, the seismic forces are reduced. On the other
hand, a large displacement may be produced in the seismic isolation level. It is necessary to
deal with this large displacement produced in the seismic isolation level not only in the
architectural planning, but also in the building equipment, and it should also be sufficiently
considered in the construction planning.
(1) Architectural planning
Compared with a seismic-resistant building, the seismic force applied to the superstructure
is small, and the degree of freedom in design planning is much more improved due to the
absence of restrictions from the seismic elements. However, the design should consider the
large displacements produced in the seismic isolation level, so that during an earthquake
occupants are not harmed and building use is not obstructed. Furthermore, it is necessary to
provide workspace in the seismic isolation level and an appropriate access route for
inspection during maintenance, repair in case of a problem, and replacement of seismic
isolation devices.
(2) Structural planning
It is necessary for the structural engineer to set the target seismic performance (criteria) to
satisfy the required targets. The target seismic performance depends on the design input
seismic motions that are set in accordance with the soil properties at the site, such as
whether an active fault exists. The structural engineer should use discretion to set the target
seismic performance, not limited only to regulatory standards, to satisfy the required
seismic performance by considering the various ranges of variation, including:
• the extent to which the actual behavior is reproduced in an analysis model
• the effect of variation in the property values of the seismic isolation devices on the
response results
• the effect of uncertainties that cannot be recognized in the analysis.
In addition, to ensure safety with respect to seismic motions that were not predicted, it is
desirable to set a safety margin in the target seismic performance with respect to the
ultimate limit state.
The performance of seismically isolated buildings is determined by the combination of

isolators that lengthen the natural period and dampers that provide damping. Normally, this
combination is determined to satisfy the set target seismic performance by a time history
response analysis, but, in addition, simple design methods have been proposed, such as the
@Seismicisolation
@Seismicisolation
energy balance method (see Section 4.3.1), the response spectrum method (see Section
4.3.2(1)), and the response prediction method using graphs (see Section 4.3.2(2)).
For the analysis, a survey of the technical properties of seismic isolation devices, including
their various dependencies, is carried out. When the performance is understood, the
appropriate property values are set. If the property values of the seismic isolation devices
that have been set differ from the actual properties, the seismic isolation effect will possibly
be reduced due to unexpected behavior, and so it is necessary to take particular care with
setting the device properties, such as the restoring force characteristics. Furthermore, if long
period components dominate – such as in long period ground motions – the response
displacement becomes large if the amount of damping is small, so it is desirable that the
appropriate amount of damping is set. However, if excessive damping is set, it is necessary
to recognize that the response shear force of the superstructure will increase. Next, because
almost all the seismic energy is absorbed in the seismic isolation level, it is necessary to
carry out the analysis with consideration of the variation in the property values of the
seismic isolation devices. The variation in the property values is an appropriate
combination of the variations due to factors such as environmental temperature, changes
with age, manufacturing variation, and construction tolerances.
Other points to note regarding design are as follows.
(a) Seismic isolation level
The seismic isolation devices are arranged so that the center of mass and the center of
rigidity coincide in order to prevent torsional vibrations. The design of the installation
structures of the seismic isolation devices should consider the additional stresses and
eccentricities generated by the large deformation in an earthquake, and the design of the
connections should ensure rigidity for the reliable transfer of forces. In buildings with a
large aspect ratio such as super high-rise buildings, it is possible that tension forces can
arise in the isolators. In this case, suitable isolators should be selected and an
appropriate model analyzed, or measures must be taken so that the tension force does
not exceed the limiting tensile strength. For the fluctuation in the axial loads of isolators
due to vertical seismic motions, besides performing a time history response analysis,
about ±0.3G should be considered for a static evaluation. If hysteretic dampers are
mainly provided for horizontal deformation, a residual deformation is produced after the
earthquake, and so it is necessary to take measures to restore the building to its original
position. Furthermore, it should be confirmed that the energy absorption capacity is less
than the limiting value in earthquakes that continue for a long period of time.
(b) Superstructure
In seismiclly isolated structures, energy is mainly absorbed by the seismic isolation

level, so it is desirable that the superstructure be designed to remain in the elastic region.
If the horizontal rigidity of the superstructure is low, the floor response acceleration
increases, the structural members become plastic and the inter-story drift angle
increases, and the energy absorption capability of the seismic isolation level decreases.
In this case, it is necessary to take measures to ensure the horizontal stiffness and load
resistance of the superstructure. Furthermore, the viscous damping set in the time
@Seismicisolation
@Seismicisolation
history response analysis should be set to a smaller value than the value set in a
seismic-resistant structure, because the deformation of the superstructure is small.
(c) Foundation structures
The additional stresses generated by the large deformation of the seismic isolation
devices should be considered because the foundations should remain in the elastic
region. If differential settlement or tilting is produced in the foundation, not only will
the vertical load of the seismic isolation devices vary, but also the apparent horizontal
rigidity due to the horizontal component of the vertical load will be reduced and the
response of the seismic isolation level will increase. Consequently, the seismic isolation
devices will be constantly subjected to a horizontal force; during an earthquake, the
deformation will be unbalanced and the required seismic isolation performance will not
be evident. Therefore, in a major earthquake it is desirable that the numerical values of
the upper limit deformation angle of the foundation structure design guide are satisfied.
Particularly in the case of weak soils, the effect of the response displacement of the
foundation is large, so sufficient care in the design of the foundation is necessary.
Furthermore, retaining walls should be designed so that they do not collapse in a major
earthquake. The thickness of a retaining wall is several tens of centimeters; so, if no
margin is set for the site, it is necessary to have a distance equal to the horizontal
clearance plus the thickness of the retaining wall separated from the site boundary line.
This distance has a large effect on planning the layout.
(3) Architectural considerations for the seismic isolation level and the clearances of the
seismic isolation devices
a) Horizontal clearances of the seismic isolation level
In seismically isolated buildings, a horizontal displacement of several tens of

centimeters is produced between the superstructure and the lower structure during an
earthquake. This horizontal displacement is generated in the seismic isolation level,
retaining wall, site boundary, adjacent buildings, building equipment piping, and
expansion joints. Clearances should be set so that the horizontal deformation of the
seismic isolation devices is not obstructed. This also applies to the elevator pit
foundations, and the part of the stairs that extends from the superstructure to the seismic
isolation level. If the seismically isolated building is connected to a non-seismically
isolated building or other structure, similar measures should be taken at the connecting
portion. It is desirable that the clearance should be the maximum response displacement
in consideration of the torsional deformation of the seismic isolation level, the input of
the seismic motions in the two horizontal directions, the effect of unanticipated
earthquakes, plus a margin. Normally, it is desirable that the clearance of the seismic
isolation level is at least 1.5 times the design displacement of the seismic isolation level
in order that the seismic isolation devices can exhibit their ultimate performance.
b) Clearances in the vertical direction in the seismic isolation level
The laminated rubber bearing that is normally used as an isolator is a device with large
vertical stiffness. Nonetheless, when the horizontal deformation is large during a large
earthquake, small settlement is produced. Furthermore, vertical clearances should be
provided that consider the settlement as a result of creep over a period equivalent to the
@Seismicisolation
@Seismicisolation
service life, expansion and contraction due to temperature variations and construction
tolerances. Normally, vertical clearances of several centimeters are provided.
c) Architectural considerations
When installing the seismic isolation devices, it is necessary to consider whether the
desired performance can be shown over a long period of time. Rubber may degrade
early as a result of exposure to ultraviolet light such as daylight, and steel members may
corrode due to rainwater or condensation. So, the layout should be designed to avoid
these. Furthermore, when sliding bearings are used, it is necessary to take measures to
protect the sliding surface from sand and dust. In addition, if contact with harmful
substances such as salt air, and sulfurous acid gas, is possible, it is necessary to take
protective measures.
Base isolated structures, which isolation level is on the top of foundation, have almost
no risk of direct exposure of seismic isolation devices to flames due to a fire, so it is not
necessary that they have a fireproof structure. But, a fireproof structure is necessary in
the case of intermediate level seismically isolated structures. The adopted method of
fire-resistant covering is to coat around the seismic isolation devices with an extensible
fireproof material or to make the seismic isolation level a fireproof compartment.
@Seismicisolation
@Seismicisolation
Section 3.4 Points to Note Regarding Construction Planning
During design the structural engineer should consider the following points regarding
construction.
The building is supported by the seismic isolation devices, and the superstructure and the
foundations are separated by sandwiching the devices in between. In construction planning,
some differences arise from a non-seismically isolated building. Even at the design stage, it is
necessary to consider matters relating to construction planning, such as the method of
bringing in and taking out the seismic isolation devices, and the installation sequence.
(1) Construction of the seismic isolation level and seismic isolation devices
The following two construction procedures are used for installation of seismic isolation
devices, and the structural engineer should understand the characteristics of these methods
during design.
a) Method of installing seismic isolation devices in a new construction
After construction of the foundations and the trenches around the building has been
completed, the isolators are set at their proper height and position, the dampers are
installed, and then construction of the superstructure proceeds. This is shown in Fig.
3.4.1. In some cases the dampers are installed last.
Pitting Foundation Installation of Construction

Installation of
& Trenching in Isolation of
Damper System
Building Perimeter System Superstructure
Fig. 3.4.1 Procedure of installation of isolation system in new construction
If large eccentric loads or horizontal forces that could adversely affect concrete with
insufficient strength are possible during construction, measures should be taken at the
time of construction planning. Such measures include installing horizontal supports or
fixings.
If the scale of the building is large in plan, the seismic isolation devices may be
subjected to forced displacement due to contraction of the concrete. Horizontal
displacements as large as about 100 mm have been reported, and so the sequence of
pouring the concrete should be established to reduce the amount of concrete shrinkage.
If structural steel is to be erected above the devices, measures should be taken during
construction to prevent deformation of the seismic isolation devices; such deformation
could easily occur. After the seismic isolation devices such as isolators and dampers
have been installed, protection should be carried out until the completion of
construction, and it is necessary that appropriate methods are indicated, for example, in
a special specification.
@Seismicisolation
@Seismicisolation
b) Method of installing the seismic isolation devices for retrofit
This method is used in seismic isolation retrofit construction. The piles, columns, or
walls at the portion that is to become the seismic isolation level are cut and the seismic
isolation devices are installed. The design and the construction planning should include
methods of temporarily supporting the building loads, construction sequence,
construction accuracy, and seismic measures to be taken during construction.
In the case of seismic isolation retrofit construction in piled foundations, after

completing the construction of trenches surrounding the building, excavation below the
building is carried out. Next, a pressure-resistant plate or similar item is provided, and a
temporary cradle is constructed to temporarily support the building load. Then, the piles
are cut, base plates are set in the cut portion, and the isolators are inserted and
integrated. Finally, the temporary supports are removed. During this time, seismic
measures are taken during construction by some method such as restraining the
deformation of seismic isolation level. This procedure is shown in Fig. 3.4.2.
Trenching in Installation of Installation of Removal of

Building Perimeter Temporary Isolation Temporary
& Excavation Supports System Supports
Fig. 3.4.2 Procedure of retrofit installation of seismic isolation system
Depending on the scale of the building, the isolators are installed by jacking up the
whole building to temporarily support the building and installing the isolators, or
temporarily supporting each block and installing the isolators. This should be
meticulously planned in advance. In each case, it is necessary to carefully consider how
to properly support the building load with the temporary support.
(2) Construction of the superstructure
When constructing the superstructure, it should be noted that the superstructure can easily
move in the horizontal direction before the dampers are installed. Normally, the
superstructure, which is supported by a rubber bearing as necessary, is temporarily
constrained to the lower structure to prevent movement of the first floor due to horizontal
loads during construction. Likewise, the external wall scaffolding during construction has
support points on the superstructure to prevent adverse relative displacements in the case of
oscillation of the superstructure due to strong winds or small to medium earthquakes.
(3) Building equipment piping and flexible construction to the building surroundings
If the pipe connections for gas pipes and high voltage cables cannot move during an
earthquake because of construction of the building equipment piping, a major accident
could occur. Space should be provided to allow the connections to move without
obstruction, and the pipe supports should be properly connected. Furthermore, the
connections should be checked by color-coding or a similar technique. Vertical and
horizontal space should be provided to enable the flexible construction to properly move. In
this case, it is necessary that the amount of space satisfies the design amount of movement
with consideration of the construction tolerances.
@Seismicisolation
@Seismicisolation
Section 3.5 Quality Control of Seismic Isolation Devices
(1) Product performance testing during manufacture at the factory
At the factory where the seismic isolation devices are manufactured, a manufacturing
guidance document is prepared in accordance with the design drawings and the special
specification. The seismic isolation devices are manufactured based on quality control
standards and the quality system for product manufacture. After passing the specified
testing and inspection, the seismic isolation devices are shipped. At each stage, the
structural engineer or the factory supervisor should evaluate and confirm the following by
quality performance testing. As a rule, the quality performance testing is carried out on
100% of the products. The rubber bearing is prescribed by the Japan Industrial Standard
(JIS), so the performance and quality control of the rubber bearing is carried out in
accordance with JIS.
(1) Material property tests
Confirm that each of the materials of the seismic isolation devices is being used in
accordance with the certificate of material tests and the manufacturing guidance
document.
(2) Dimensional inspection and external visual inspection
The dimensional inspection should confirm that the dimensions are within the
allowable tolerances in accordance with dimensional inspection criteria such as the
manufacturing guidance document. Measuring equipment and jigs that are capable of
accurately measuring the dimensions are used in the inspection. The items inspected
include product height, product inclination and flange misalignment, flange diameter,
film thickness for corrosion prevention on the flange, and other necessary items.
The external visual inspection visually confirms any damage or deformation,

corrosion, peeling of paint, and the device number.
(3) Performance testing
Performance testing confirms that the performance is within the allowable values in
accordance with testing criteria such as the manufacturing guidance document.
The items inspected for isolators include vertical stiffness and horizontal stiffness.
For isolators having a damping function, the yield load (intercept load) and the
equivalent damping coefficient are measured. For sliding bearings and roller
bearings, the friction coefficient is measured.
Measurement items for dampers include the restoring force properties, including the
initial stiffness, the secondary stiffness, and the yield load. For damping performance,
the property values of the damping force and the equivalent damping coefficient are
measured.
@Seismicisolation
@Seismicisolation
(2) Control during construction
Control of the seismic isolation level and the seismic isolation devices during construction
is carried out from the time that the seismic isolation devices are transported onto the site
until handover of the building.
During construction, the installation accuracy or the dimensional accuracy of the building
side corresponding to the dimensions of the seismic isolation devices must be verified to be
within the ranges indicated in the special specification. In particular, care is required to
ensure the horizontal accuracy when installing isolators such as rubber bearings. After
installation, the seismic isolation devices are protected until completion of the building to
ensure that no damage occurs. If necessary, the relation between the deformation of the
seismic isolation devices and the load is measured at each stage of construction until
completion of construction.
(3) Testing and inspection at completion of construction of the building
When the building is completed, the structural engineer carries out comprehensive testing
and inspection.
The inspection includes confirming the condition of the seismic isolation devices
(deformation or slack bolts), clearances of the seismic isolation level, and the building
position. In particular, it is confirmed that no obstacles exist within the range of movement
of the seismic isolation devices so that an earthquake would not impact the service piping.
On rare occasions, dynamic testing such as forced displacement testing may be carried out.
After completion of construction, periodic inspection is carried out on the seismic isolation
level and seismic isolation devices so that the performance at the time of completion of
construction can be maintained properly. For this purpose, the inspection records upon
completion of construction are stored for use as criterion values.
@Seismicisolation
@Seismicisolation
Section 3.6 Maintenance of the Seismic Isolation Level and Seismic Isolation Devices
In seismically isolated buildings, the seismic performance of the building is greatly

determined by the performance of the seismic isolation level, which absorbs almost all of the
seismic energy. If a problem occurs in the seismic isolation level, the seismic isolation
performance expected during design will not be shown, and damage to various parts such as
the superstructure is a possibility. Therefore, maintenance should be carried out so that the
seismic isolation level functions properly during earthquakes that may occur at any moment.
Furthermore, after the occurrence of a disaster, it should be confirmed that the seismic
isolation level is functioning properly. The following characteristics of seismic isolation
levels should be properly recognized to carry out maintenance.
(1) The seismic isolation devices exhibit their required performance during a major
earthquake. A large relative displacement is produced in the seismic isolation level in the
horizontal direction on the order of several tens of centimeters.
(2) Movable parts other than the seismic isolation devices are capable of following the
relative displacement of the seismic isolation level without resistance.
It is necessary that these characteristics are recognized and the basic principles that enable
the seismic isolation performance are understood. The structural engineer has an obligation
to explain the characteristics of the seismic isolation level to all those involved. Furthermore,
if required, the structural engineer should prepare a maintenance guidance document for
persons carrying out building maintenance, and should provide guidance and support.
The structural engineer should carry out the inspection upon completion of construction,
summarize the inspection results in a document, and hand over this document to the person
carrying out building maintenance as criterion values to be used in periodic inspections after
completion of construction. In addition, it is desirable that measures be taken to obtain the
understanding of the actual building users so that objects will not be placed in necessary
clearances such as a dry area. It is also desirable to display panels that explain that the
building is a seismically isolated building. Finally, to confirm the seismic performance of the
seismically isolated building during an earthquake, it is desirable that a measuring device
such as a seismometer or a scribed type displacement meter is installed.
References
1) Japanese Industrial Standard：JIS K 6410-1：2011 Elastomeric Seismic Protection Isolators for Buildings Part.1：
Specifications
2) Japanese Industrial Standard：JIS K 6410-2：2011 Elastomeric Seismic Protection Isolators for Buildings Part.2：Test
methods
@Seismicisolation
@Seismicisolation
Chapter 4 Structural Design
Section 4.1 Design Targets
4.1.1 Fundamental Considerations
In a seismically isolated structure, the structural system has the following characteristics: as
the natural period of the structure becomes longer, the response acceleration decreases, and if
the structural damping is large, the response amplification is small. Evaluation of the input
ground motions is an important part of the design criteria for seismically isolated as well as
conventional structures. However, because seismically isolated structures incorporate a
dynamic effect in the design method, the importance of evaluation of the input ground
motions is even greater. Therefore, appropriate evaluation of the input ground motions at the
proposed site is important in order to set the design criteria and for the subsequent design of
the seismically isolated structure.
To determine whether the design criteria are satisfied and to understand the behavior of the
seismically isolated building during an earthquake, various methods, such as the following,
are used for response prediction:
(1) time history response analysis
(2) energy balance method
(3) response spectrum method
Time history response analysis directly calculates the response values of structures, so it is an
effective method. However, the response values are sensitive to the characteristics of the
input ground motions, so it is effective to verify the results by another calculation method.
Each of the different methods applies the input ground motions in a different form. In the
time history response analysis, the input ground motions are an acceleration waveform. In the
energy balance method, an energy spectrum is used instead of an input acceleration
waveform, and in the response spectrum method, an acceleration response spectrum is
applied.
In recent years, significant progress has been made in the technology for evaluating ground
motions, and methods have been produced for predicting ground motions by setting
parameters such as the properties of the epicenter, the transmission path of the seismic
motions, and the soil properties at the site.
@Seismicisolation
@Seismicisolation
The following are methods to calculate the input ground motions: theoretical and
semi-empirical methods assuming individual fault models, probabilistic methods of settings
based on the frequency of occurrence of ground motions by using the properties of ground
motions obtained empirically based on past earthquake data, and methods of generating
acceleration waveforms numerically so that they are equivalent to a velocity response
spectrum or an energy spectrum that has been defined as the design spectrum.
Each of these methods includes parameters that have various uncertainty factors such as the
properties of the source, the seismic wave path effect and the soils. The input ground motions
vary greatly depending on these values, and the building response also varies. Therefore, in
structural design, it is necessary to set the input ground motions by using engineering
judgment, after thoroughly understanding the information regarding earthquakes and ground
motions obtained to date, and after comparing the results with other calculation methods,
including variations based on these uncertain factors, and studying various ground motions
by varying the parameters.
Furthermore, it is important to use past research results as material for setting the ground
motions. The contents of various documents produced by the Architectural Institute of Japan
(AIJ) can be consulted. Records of past major earthquakes and the properties of their ground
motions can also be examined for making design judgments.
Design for wind loading is carried out in the same way for both seismically isolated as well
as conventional buildings. However, in recent years seismically isolated structures are being
applied to buildings that are greatly affected by wind loading, such as super high-rise
buildings and detached houses, so it is important to design these super high-rise buildings
and detached houses against wind loading.
To increase the seismic isolation effect, it is important to increase the natural period of the
seismically isolated building; but if the wind loading as a percentage of the seismic load is
high, vibrations are excited in the across-wind direction, and the displacement of the seismic
isolation level is greatly increased by the dynamic component of the wind load. Therefore, it
is essential to confirm the degradation in performance of the seismic isolation devices due to
repeated displacement for a long duration time. On the other hand, if the yield shear force of
the seismic isolation level is set high in order to reduce the displacement of the seismic
isolation level due to wind loads, the shear force of the superstructure increases.
4.1.2 Matters to be Considered in Evaluating the Input Ground Motions
The factors affecting the properties of the input ground motions (maximum acceleration,
@Seismicisolation
@Seismicisolation
maximum velocity, spectral properties, phase properties, waveform envelope,

non-steady-state spectral properties) include the following:
a) degree of seismic activity in the surrounding area
b) fault properties
c) properties of the transmission path from the fault
d) structure of the deep soils
e) surface topography and soil properties.
To define the input ground motions for design, it is necessary to appropriately evaluate the
effects of these factors.
In particular, it is necessary to set the properties of the ground motions in the long period
region, which has a great effect on seismically isolated structures, and to consider the effect
of the scale of the earthquake and the amplification properties of the deep soil structure up to
the bedrock. In alluvial plains such as the Kanto Plains or the Osaka Plains, research results
have shown that the input of the long period component is more than 1.5 times that of
components with region periods of 2 to 4 seconds1), so it is necessary to be particularly
careful if the seismic isolation period is extended to 5 seconds or higher.
The acceleration response spectrum used for calculating the seismic loads can be broadly
defined in the free engineering bedrock (about Vs (shear wave velocity) = 400 m/s or higher),
or the position of input ground motions to the building analysis model. If the values (input
ground motions or spectrum) for the free engineering bedrock are provided, it is necessary to
appropriately evaluate the amplification due to the surface soils.
In particular, when constructing on weak soils, the soil period during major earthquakes is
extended, and the seismic isolation effect is reduced, so a careful design is necessary. On the
other hand, when the reduction in stiffness due to the shear deformation of the soils is
over-evaluated, the input ground motions on the short period side are reduced. So, it is
necessary to appropriately determine the various constants for the soils. In particular, the
occurrence of liquefaction is significant. When the design input ground motions are obtained,
it is necessary to consider whether liquefaction is accounted for in the design, or if
liquefaction has been ignored.
In the actual ground motion properties, the short period components of the input ground
motions of a site are attenuated by the seismic motions of the surface soils, and the
high-frequency components are eliminated by the input loss effect due to the dynamic
interaction between the building – piles – soils system. This is different from the reduction
effect due to seismic isolation; but, because of the long period of the seismically isolated
building, the input loss effect on the response displacement in the horizontal direction of the
@Seismicisolation
@Seismicisolation
seismic isolation level is generally small.
4.1.3 Design Criteria
In the structural design of buildings, large-scale earthquakes that may occur in the near future
and have a significant impact on the construction site are estimated, and the safety is
evaluated with respect to the ground motions caused by these earthquakes.
Safety related to the structural system or structural members is not the only performance
requirement of buildings. A wide variety of performances is necessary, such as the integrity
and usability of the assets of the building itself or its interior (functionality, dwelling
comfort). Therefore, the design criteria should just not consider the maximum level of
ground motions; instead, the design criteria should be set from the viewpoint of vibration
level and the frequency of occurrence of ground motions. In other words, it is necessary to
set design criteria for the seismic strength (amplitude level) and the frequency of occurrence
(return period) in accordance with the safety, reparability, and usability of the structure.
Table 4.1.1 shows examples of the design criteria. For example, rarely occurring ground
motions (ground motions that have a large possibility of occurring once or more during the
service life) can be set as level 1, with a target that the acceleration within the building is
reduced, the building function is maintained, and the occupants feel secure. Extremely rarely
occurring ground motions (the strongest ground motions that have occurred in the past at the
site of the building, or the strongest ground motions that are considered to possibly occur in
the future) can be set as level 2, with the target that damage does not occur, and human life
and the assets of the building are protected. Furthermore, design criteria considered include
the seismic probable maximum loss (PML), which expresses the predicted maximum value
of loss due to the maximum scale of earthquake expected during the period of use of the
building as a proportion of the cost of re-procurement of the building.
A seismically isolated structure is a structural system with a low degree of static

indeterminacy governed by the response of the seismic isolation level, so a design with a
margin in the seismic isolation level is important to ensure safety. Therefore, the redundancy
design level is set to investigate the failure mode of the building and to ensure a certain
margin of safety with respect to inputs that exceed the design ground motion level.
Normally, if the seismic isolation level is designed with a large yield shear force with the
priority at level 2 safety, the seismic isolation effect cannot be obtained at level 1, the
response acceleration increases, and other problems occur. Therefore, it is important to set
the criteria with a balance between levels.
@Seismicisolation
@Seismicisolation
The inter-story drift angle criteria can be set in accordance with the type of structure and the
structural shape, but it is necessary to sufficiently take into consideration the prevention of
plastic deformation of the superstructure and the effect on the deformation of the seismic
isolation level during the wind response. In particular, if plasticity of the superstructure
increases, the inter-story drift angle increases greatly, as discussed in Section 4.4.9; so, as a
rule, the superstructure is designed to remain within the elastic region.
Table 4.1.1 Examples of design criteria

Level 1 design Level 2 design Redundancy design
Level of input ground motion
Rarely occurred Extremely rarely
Level 2 × 1.5
earthquake occurred earthquake
Building performance Serviceability limit Reparability limit Safety limit
Story drift
≦ 1/300 ≦ 1/200 ≦ 1/150
ratio
Superstructure ≦ Allowable stress

State of ≦ Ultimate strength
above the for temporary Load ≦ Elastic limit of story
structure of story
isolation level （Each member）
Response
≦ 250 Gal ≦ 300 Gal －
acceleration
Response ≦ Stable limit ≦ Performable limit ≦ Ultimate

deformation deformation deformation deformation
State of ≦ Ultimate tensile
Isolation level No tensile force No excessive uplift
uplifting strength
Energy
≦ 1/10 of maximum ≦ 1/2 of maximum ≦ maximum absorbed
absorption
absorbed energy absorbed energy energy
by dampers
State of ≦ Allowable stress ≦ Allowable stress for
Sub-structure ≦ Ultimate strength
structure for temporary Load temporary Load
below the
isolation level State of ≦ Allowable stress ≦ Allowable stress for
≦ Ultimate strength
foundation for temporary Load temporary Load
As a rule, excessive tensile axial forces may not act on the isolators. Some isolators can resist
tensile axial forces, but usually the axial stiffness of rubber bearings under a tensile axial load
is smaller compared with the vertical stiffness under a compressive load. If a tensile axial
load is generated at the isolators, the effect of rocking vibrations must be appropriately
evaluated, and it is necessary to ensure a margin of safety with respect to toppling and the
vertical load bearing capacity of the isolators on the compressive side.
Figure 4.1.1, obtained from the response prediction based on the energy balance, shows the
relation between the limiting aspect ratio of the superstructure H/B (building height/width of
shortest side direction) and the seismic isolation period Tf to prevent production of tensile
forces in the seismic isolation devices2).
@Seismicisolation
@Seismicisolation
H/B VE(cm/s)
120
150
200
Period of seismic isolation Tf(s)
Fig. 4.1.1 Relation of isolation period (Tf) to aspect ratio (H/B)
All the seismic input energy is ultimately absorbed by the dampers. Therefore, an appropriate
margin of safety is set for the damper repetition durability and energy absorption capability
by considering the frequency of occurrence of earthquakes including aftershocks.
The lower structure below the seismic isolation level is also, as a rule, designed to remain
elastic under level 2, which is the same level for the superstructure, to ensure rigidity and
load bearing capacity to enable the vertical loads to be safely supported corresponding to the
margin of safety of the building.
The effect of ground motions in the horizontal direction on the safety and reparability of
conventional seismic-resistant structures is large, but in the case of seismically isolated
structures, the building response acceleration in the horizontal direction is greatly reduced, so
the effect of the horizontal forces on the stresses of structural members is smaller. However,
in the case of vertical ground motions, the acting input is about the same for seismically
isolated and conventional seismic-resistant structures, and in some cases members are
determined by the vertical motions. So, necessary criteria may be set with respect to the
vertical ground motions so that safety can be evaluated.
4.1.4 Judgement of the Seismic Resistance Performance
The seismic resistance performance of the seismic isolation level, the superstructure, and the
lower structure are determined based on response prediction.
For the seismic isolation level, it is confirmed that the response displacement and the acting
axial forces do not exceed the design criteria. In this case, the effect of the two horizontal
ground motions and the vertical ground motions is appropriately taken into consideration by
simultaneous input into the time history response analysis or by combining the stresses and
@Seismicisolation
@Seismicisolation
deformations in each direction or the amounts of energy absorption in each direction. Also,
the variation in the properties of the seismic isolation devices, such as manufacturing
tolerances, temperature dependence, and degradation with age, is appropriately evaluated.
It is also checked whether each story of the superstructure and the lower structure responds
elastically and the strength of the structure can be resistant against the combination of
vertical forces during an earthquake produced by the horizontal shear forces, overturning
moments, dead load, and vertical seismic motions.
As a result of cracking in a reinforced concrete structure, the natural period can extend to
more than 1.5 times the natural period when elastic, so it is necessary to ensure that
resonance with the seismic isolation period does not occur. Normally, when the equivalent
period of the seismic isolation level (superstructure is assumed to be rigid) is more than
double the primary elastic period of the superstructure, the effectiveness of seismic isolation
can be expected3).
The seismic resistance of the structural members around the seismic isolation devices is
determined appropriately by taking into consideration the effects of additional bending
moment due to axial forces. The vertical load bearing capacity of the lower structure, such as
the foundation and pile, is also checked by considering the response displacement of the
soils.
4.1.5 Investigation of Input Ground Motions Exceeding the Design Level
With current knowledge it is not possible to accurately predict the magnitude or properties of
the largest ground motions that can occur at the site of a building. It is also not possible to
completely eliminate the possibility of occurrence of a future giant earthquake exceeding the
design ground motion level. Therefore, when designing the building, it is important to take
into consideration cases in which the input ground motions exceed the design level.
In an ordinary seismic-resistant design, lower bound values lower than the actual strength are
used for the allowable stresses of materials. Furthermore, the conventional seismic-resistant
structural system absorbs the seismic input energy throughout the whole building, so an
unseen factor of safety exists with respect to input ground motions that exceed the design
level. On the other hand, in a seismically isolated building, the majority of the seismic energy
is absorbed in the seismic isolation level and the structural system has low static
indeterminacy. Therefore, although the response can be accurately predicted if the input
ground motions are defined, it is considered that the safety margin with respect to ground
motions exceeding the design level is slightly lower compared with an ordinary
@Seismicisolation
@Seismicisolation
seismic-resistant structure. Also, the response properties of a seismically isolated building

depend greatly on the properties of the isolators and the dampers, but it is possible for the
properties to differ considerably from the set properties due to variation of the individual
products, degradation with age, repeated deformation, and being subjected to large
deformation or tensile deformation. Therefore, when designing a seismically isolated
building, special consideration is required for cases in which the ground motions exceed the
design level and cases in which the properties of the seismic isolation devices differ
significantly from the design values.
When subjected to long duration long period seismic motions and when the seismic isolation
level is subjected to repeated deformation over a long duration or subjected to multiple
aftershocks with increasing cumulative deformation, the rigidity and energy absorption
capability is gradually reduced, depending on the type of seismic isolation devices. As a
possible result, the maximum deformation of the seismic isolation level could greatly exceed
the estimated maximum. In cases in which it is not possible to obtain a sufficient margin in
seismic isolation clearance, such as on a narrow site, the superstructure could possibly impact
the surrounding retaining wall due to horizontal deformations that exceed the design
deformation, and so it is necessary to confirm the safety during impact. According to several
analytical and experimental studies, the impact acceleration when the superstructure impacts
a retaining wall varies depending on the conditions of rigidity of the retaining wall and
impact velocity4), but it is necessary to take this into consideration in the design. Also, it is
necessary to pay attention to the generation of rocking vibrations in the superstructure due to
the impact5).
The linear range of the laminated rubber bearing is up to shear strains of about 200%–250%,
and above the exhibited nonlinear elastic properties associated with hardening. If the
response exceeds the rubber linear range, it is necessary to be aware that the seismic forces
transmitted to the superstructure increase. The response displacement of the superstructure of
seismically isolated buildings tends to increase due to plasticity6), so it is necessary to
provide a margin in the load resistance of the superstructure. If the response of the seismic
isolation level is further increased, the laminated rubber bearing approaches its ultimate limit
state in shear failure or compression buckling. It is generally said that the shear strain of a
laminated rubber bearing at failure is greater than 400%, so it is considered that the margin is
determined with respect to the linear limit strain. However, shear failure tests provide
insufficient data, in particular for the large-diameter laminated rubber bearing, and it has
been reported that the shear failure strain reduces due to degradation with age7), 8), so caution
is necessary.
In the case of buildings with a large aspect ratio superstructure, caution is necessary
regarding variations in the axial force of isolators with respect to overturning moments
@Seismicisolation
@Seismicisolation
associated with horizontal motions. When the aspect ratio is large, the rocking rigidity is
reduced relative to the shear rigidity, and rocking vibrations dominate. Input that excites
rocking vibrations is not normally taken into consideration, and the seismic isolation level
has no damper effect against rocking vibrations. Therefore, it is desirable that a sufficient
factor of safety be incorporated so that excessive tensile displacement of the laminated
rubber bearing on the tensile side or buckling of the laminated rubber bearing on the
compression side does not lead to a form of failure with no redundancy.
However, the tensile stiffness of a rubber bearing is extremely low, except in comparison
with some roller type isolators having a mechanism to resist tension. In the case of the rubber
bearing, the tensile stiffness is only 1/10 to one in several tens the stiffness on the
compression side. Therefore, if compression forces act on the surrounding isolators, it is
possible to redistribute the tensile reaction.
The laminated rubber bearing has a sufficiently high tensile deformation capability when
tensile forces act on them, and after voids are generated within the rubber, the reduction in
properties in the horizontal direction is not large. However, multiple rubber sheets and steel
plates are bonded in series, so it can be considered to be the most fragile link model. In this
model, the damage due to tension is concentrated at the most fragile single location, and thus
it is desirable that tension strain is as small as possible.
4.1.6 Trade-off Relation between Seismic Load and Wind Load in Structural Design
In recent years, the application of seismic isolation to super high-rise buildings with aspect
ratios (building height/width of shortest side direction) exceeding 4.0 has been increasing.
The specific design method is to reduce the response shear force of the superstructure by
increasing the natural period by the seismic isolation level. In most cases, the technical
problems have been overcome. Among these is the comprehensive analysis of the relation
between seismic loading and wind loading in the design of seismically isolated structures.
As shown in Fig. 4.1.2, if the yield shear coefficient of the seismic isolation level is reduced
to reduce the response shear force on the superstructure, the seismic isolation effect on the
superstructure can be easily obtained in severe earthquakes. However, the cumulative
damage in the seismic isolation level and the residual deformation has now become a
problem with respect to maximum external wind loads such as storms (in addition, it is
necessary to examine the temperature dependence of the seismic isolation devices and the
mechanical properties of the devices in two input directions by considering the effects
normal to the wind direction). Also, as recognition of the problem of long period long
duration seismic motions increases9), it is also necessary to sufficiently examine the
@Seismicisolation
@Seismicisolation
resonance phenomenon. However, if the yield shear force coefficient of the seismic isolation
level in this type of seismically isolated structure is set high, the response shear force of the
seismic isolation level during a severe earthquake is easily transmitted to the floors above.
The larger response shear force can cause plasticity in the superstructure, and in some cases
an excessive uplift force that exceeds the allowable values of the seismic isolation devices
can occur10).
On the other hand, to reduce and resolve the degree of difficulty in the design of this type of
super high-rise seismically isolated building, various developments are being actively carried
out, such as providing a mechanism to reduce the uplift forces on the bottom flange of the
rubber bearings11),12),13), or wind resistance shear pins that increase the yield shear force
coefficient of the seismic isolation level during strong winds14). To satisfy the various design
conditions for seismic loads and wind loads, the actual safety limit on the aspect ratio of
super high-rise seismically isolated buildings that can benefit from the seismic isolation
effect is between 5.0 and 6.0.
Various aspects of consideration in the seismic isolation system design.
Larger ・Larger design shear force in the superstructure

・Larger tensile force in the isolators
To be applicable.
e.g. Uplift-resistant
Yield shear force coefficient
mechanism etc.
(Isolation level)
Applicable zone for

the seismic isolation system
To be applicable.
e.g. “Shear pin” for
response control toＺ the
wind load etc.
・Need for consideration of resonance to the long period earthquakes.
Smaller ・Need for consideration of the wind load.
(e.q. Consideration of accumulated deformation in the devices)
Larger (longer)
Aspect ratio（natural period）
Fig. 4.1.2 Comprehensive examination in seismic isolation design
To understand the mechanics of this phenomenon, the trade-off relation between the seismic
load and the wind load for super high-rise seismically isolated buildings is shown in Figs.
4.1.3 and 4.1.4. Also, to specifically grasp this trade-off relation, time history analyses using
seismic loads (ground motions) and wind loads (external wind forces) were carried out. In
these analyses, the design problems for super high-rise seismically isolated buildings are
comprehensively examined by carrying out a 3-dimensional analysis simulation for a super
high-rise seismically isolated building of building height 150 m and aspect ratio of 5.0 by
using an analysis model with a yield shear force coefficient of the seismic isolation level of
0.02 and 0.03. The results show that during the seismic response, when the model with the
@Seismicisolation
@Seismicisolation
smaller yield shear force coefficient of the seismic isolation level of 0.02 is compared with
the model with the larger yield shear force coefficient of 0.03, the seismic isolation affect of
the superstructure is greater in the model with the smaller yield shear force coefficient of the
seismic isolation level, and a safety advantage is confirmed for the tensile stress on the
seismic isolation devices arranged in the corners of the building. On the other hand, during
the wind response, the cumulative deformation and the residual deformation of the seismic
isolation level (devices) is larger in the model with the smaller yield shear force coefficient of
the seismic isolation level, so a trade-off relation between seismic loads and wind loads is
confirmed.
Response shear force is small Wind load
Deformation of the
isolation level is large
Deformation of the
Tensile force is small isolation level is large
Accumulated deformations in the devices are large

Overturning
Earthquake motions
(horizontal+vertical)
moment is small *Need for check the performances of isolator against the wind load
Fig. 4.1.3 Trade-off relation between the seismic load and the wind load acting
on the super high-rise seismically isolated building
(Case of small yield shear force coefficient in the isolation level)
@Seismicisolation
@Seismicisolation
Response shear force is large

Wind
風圧力load
Deformation of the
免震層の変形
isolation level is small
が小さい場合
Tensile force
引抜き力is large 大 Deformation of the
免震層の変形
isolationが小さい場合
level is small
Overturning 大 Accumulated deformations

累積変形小 in the devices are small
転倒モーメント
※地震応答時の免震効果が小さい
moment is large
Earthquake motions *Less isolation effect against earthquake motions
地震動(水平・上下)
(horizontal+vertical)
Fig. 4.1.4 Trade-off relation between the seismic load and the wind load acting
on the super high-rise seismically isolated building
(Case of large yield shear force coefficient in the isolation level)
References
1) Hiroshi Akiyama, Haruyuki Kitamura：Relation Between Energy Spectra and Velocity Response Spectra, Journal of
Structural and Construction Engineering, AIJ, No. 608, pp.37-43, 2006.10
2) Architectural Institute of Japan：Recommendation for the Design of Seismically Isolated buildings (2nd Ed.), 1993
3) R. I. Skinner, W. H. Robinson, G. H. McVery："An Introduction to Seismic Isolation", John Wiley & Sons, 1993
4) Toshimoto Maeno, Shunji Yamamoto, Hiroyuki Masuda, Naohiro Nakamura, Takuya Kinoshita, Hiroshi Kawase,
Shin-ichi Matsushima：Experiment on Imapct of Base-Isolated Building Model Against Retaining Wall：Part.1：
Experiment, Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan, 2011.8
5) Jumpei Komaki, Goro Miwada, Hideo Katsumata, Kotaro Sato, Takeshi Sano, Noriko Takiyama, Yasuhiro Hayashi：
Experiments of Collision to Retaining Wall with Real Scale Base-Isolated Building：Part.4：A Simulation Analysis
Using Multi-Degree of Freedom System, Summaries of Technical Papers of Annual Meeting, Architectural Institute of
Japan, pp.433-434, 2010.7
6) Kikuchi M, Black J.C. and Aiken I.D.：On The Response of Yielding Seismically Isolated Structures, Earthquake
Engineering and Structural Dynamics, 37, 2008
7) Taiji Mazda, Yasuki Ohtori, Kazuta Hirata, Katsuhiko Ishida：Evaluation of Long-term Durability of Natural Rubber
pp.717-718, 1996.9
8) Kazuta Hirata, Yasuki Ohtori, Katsuhiko Ishida, Taiji Mazda：Evaluation of Long-term Durability of Lead Rubber
1996.9
9) Architectural Institute of Japan：Structual Response and Performance for Long Period Seismic Ground Motions,
@Seismicisolation
@Seismicisolation
2007.12
10) Tadamichi Yamashita, Yoji Ooki, Tetsushi Inubushi, Jun Kawabata, Masayuki Ninomiya, Mitsuhiro Saito：A Study on
Tensile Force by Isolation Devise with Vertical-Direction Passive Control System for Base-Isolated High-Rise Building
(Part.1：The Validation of Maximum Tensile Force to Base-Isolation Devices)，Summaries of Technical Papers of
Annual Meeting, Architectural Institute of Japan, pp.203-204, 2010.9
11) Tetsushi Inubushi, Yoji Ooki, Tadamichi Yamashita, Jun Kawabata, Masayuki Ninomiya, Mitsuhiro Saito：A Study on
Reduction of Tensile Force by Isolation Device with Vertical-Direction Passive Control System for Base-Isolated
High-Rise Building (Part.2: The Reduction of Maximum Tensile Force to Base-Isolation Devices), Summaries of
Technical Papers of Annual Meeting, Architectural Institute of Japan, pp.205-206, 2010.9
12) Yasuo Takenaka, Toshiyuki Yoshimatsu, Jun Tagami, Arinori Nimura, Yasuhiro Kasahara, Masami Mochimaru：New
Setting Method of Laminated Rubber Bearings using Wingshape Steel Plates for Tensile Counter Measure (Part.1)
Outline of Setting Method and Testing Plan, Summaries of Technical Papers of Annual Meeting, Architectural Institute
of Japan, pp467-468, 2004.8
13) Nobuyuki Ogino, Naoki Kato, Masayuki Yanagi, Miyuki Kaihotsu, Yusuke Hotta, Akira Wada：Study of Isolation
System with IBT: Part.1: Case Study. IBT, Summaries of Technical Papers of Annual Meeting, Architectural Institute of
Japan, pp.837-838, 2009.8
14) Yuuichi Iwai, Hideaki Tsubaki, Kiyoaki Hirakawa, Kouhei Kishimoto：Structural Design of Base-Isolated High-Rise
Housing Using a Shear-Pin Resisting Wind Load with a Gap, Summaries of Technical Papers of Annual Meeting,
Architectural Institute of Japan, pp.583-584, 2006.9
@Seismicisolation
@Seismicisolation
Section 4.2 Input Ground Motions
4.2.1 Establishing the Input Ground Motions
(1) Evaluation of the frequency of occurrence of earthquake ground motion strength
The ground and buildings oscillate when an earthquake occurs. An earthquake is a

phenomenon in which forces and strains that have gradually built up within the rocks
deeply underground reach their limit, thereby producing a slip failure. When an earthquake
occurs, seismic waves are generated and transmitted through the ground or along the
ground surface. The oscillations on the ground surface or underground at a point to which
seismic waves have been transmitted are referred to as "earthquake ground motions." In
everyday terminology these motions are frequently referred to as “earthquakes”, but here
the words “earthquake”, “seismic wave”, and “earthquake ground motions” are
distinguished.
To evaluate the frequency or probability of occurrence of earthquake ground motions, it is

necessary to calculate the "probability of exceedance of ground motions" obtained by
multiplying the "probability of occurrence of a certain earthquake" by the "probability that
the oscillations at a certain point when that earthquake occurs will exceed a certain
amount," summed over all the earthquakes.
In Japan, the Headquarters for Earthquake Research Promotion (hereafter referred to as

HERP) is carrying out long-term evaluations of the major active fault zone and subduction
zone earthquakes throughout Japan1). In the long-term evaluation by HERP, the location
and shape of faults that are considered to cause earthquakes in the future, the magnitude of
those earthquakes, and the slip amount on the fault planes accompanying the earthquakes,
are being estimated, and the probabilities of occurrence of earthquakes in the future are also
being calculated.
Actually, multiple and repeated earthquakes occur at all locations of Japan. The
characteristics of earthquakes such as the major active fault zones and subduction zone
earthquakes, have been identified and modeled according to position, magnitude,
probability of occurrence, and failure characteristics1). For earthquakes whose history of
occurrence is clear, the repeated occurrence of similar earthquakes is modeled by
considering their activity interval as being in accordance with a Brownian Passage Time
distribution (BPT distribution). On the other hand, in some cases, the most recent active
period in the past is unclear. In these cases, it is assumed that the occurrence of these
earthquakes follows the "Poisson process," and the probability of occurrence of the
earthquake is calculated by using only the information that "the earthquake occurs on
average at intervals of so many years." Actually, frequently a difference exists in the
evaluation results between the "average activity interval" and the "most recent activity
period." When preparing probabilistic seismic motion prediction maps, described later,
consideration is given to both the "average case" and the “maximum case.” The average
case uses the central value of the abovementioned interval and period as the representative
value of the probability of occurrence of an earthquake, and the "maximum case" uses the
maximum value of the evaluated probabilities. Also, many earthquakes actually occur that
are not major active fault zone and subduction zone earthquakes for which the seismic fault
has been identified in the long-term evaluation by HERP. For example, shallow inland
@Seismicisolation
@Seismicisolation
earthquakes have occurred at an unknown active fault, and small and medium sized
inter-plate earthquakes, have occurred. HERP refers to these earthquakes as "earthquakes
for which it is difficult to identify the seismic fault in advance," and their effect is taken into
consideration when preparing probabilistic seismic hazard maps. Because it is difficult to
determine in advance the location, the magnitude, and the probability of occurrence for
each earthquake lacking an easily identifiable seismic fault in advance, their characteristics
as an earthquake group are represented by a probabilistic model.
Based on their long-term evaluation, HERP has produced and published two types of
seismic hazard maps, named "probabilistic seismic hazard maps" and "seismic hazard map
for specified seismic source faults"2), 3). Of these, "probabilistic seismic hazard maps" are
maps based on the positions, magnitude, and probabilities of all earthquakes. The
probabilities and intensities of ground motion at each location are calculated, and their
distributions are shown on maps. In other words, the probability of exceedance of ground
motions is calculated by multiplying the "probability of occurrence of a certain earthquake"
by the "probability of exceedance of the amplitude of ground motions at a location when
that earthquake has occurred," summed over all earthquakes. The results are shown on the
maps as the probability of ground motions at each point and the distribution of the
amplitude of the motions. In other words, the maps summarize information on the relation
between the amplitude, period, and probability of ground motions at each point based on
hazard curves that probabilistically process the occurrence of multiple repeated earthquakes
and the amplitude of the ground motions of these earthquakes.
In the maps, two of either the amplitude, period, and probability of the ground motions are
fixed and the spatial distribution of the remaining parameter is shown on the map. The
types of maps that are currently published by HERP are maps showing the distribution of
probability of exceedance and maps showing the distribution of seismic intensities2), 3).
Therefore, the distribution of ground motions occurring at the same time due to a single
earthquake is not shown. All the earthquakes that can occur in the surrounding area are
modeled probabilistically according to type, and at present the ground motions are basically
evaluated empirically (by using ground motion prediction equations).
Specifically, after modeling the occurrence of multiple repeated earthquakes, the maximum
velocity of the ground motion at the engineering bedrock is obtained by using the ground
motion prediction equations. Then, the maximum velocity of the ground motions at the
ground surface is obtained by multiplying by the surface soil amplification factor, and the
seismic intensity at the ground surface is obtained by conversion with an empirical
equation. A similar process is carried out for predicting the ground motions at depths
shallower than engineering bedrock for the seismic hazard maps for a specified seismic
source.
The weaker the ground motions at a given location (amplitude and seismic intensity are
small), the greater the probability of exceedances of that value. The lower the probability of
exceedance at a given location, the larger the motions (amplitude and seismic intensity are
large), and this trend is particularly significant on nearby active faults. Even considering the
same location, the same period, and the same probability, the earthquakes and their ground
motions (seismic intensity) that should be considered are diverse, and it is desirable that
specific preparations are considered based on their characteristics.
@Seismicisolation
@Seismicisolation
(2) Setting the specified (scenario) earthquake
Based on the results of the long-term evaluation by HERP, seismic fault models for
predicting strong motions have been set for earthquakes in which the seismic fault has been
identified in advance, such as earthquakes along the major active fault zone and subduction
zone earthquakes,2),3). If detailed information on the location can be obtained, that
earthquake should be considered to be a priority, and parameters that are not clear should be
set in accordance with a procedural "recipe"4) that is uniform for the whole country of
Japan.
The method of setting source parameters summarized in the "recipe" is fundamental for
constructing characteristic fault models for earthquakes occurring at postulated seismic
faults. The aim is that the standard values of fault parameters for predicting strong motions
have reproducibility. The characteristic source model treats the plane fault in a binary
manner that includes asperities, which are the main rupture regions where large slip and
stress drops occur during an earthquake, and the background region other than asperities.
Procedures are actually provided for progressively setting parameters for properties such as
the macroscopic source properties that indicate the shape and scale of the overall fault, the
microscopic source properties that mainly indicate the non-uniformity of the fault, and
other properties that indicate the rupture processes4).
The macroscopic fault parameters (parameters related to the macroscopic source properties)
describe the overall view of the seismic fault. Specifically, the following parameters are
included:
• location and structure of the seismic fault model (location, strike, segment)
• size of the seismic fault model (length, width), depth, dip angle
• earthquake magnitude
• average slip amount of the seismic fault model.
The microscopic fault parameters (parameters related to microscopic source properties)

describe a detailed view of the seismic fault. Specifically, the following parameters are
included:
• location and number of asperities

• area of asperities
• average slip amounts of the asperity and of the background region
• effective stresses of the asperity and background regions
• high frequency cutoff frequency (fmax)
• average rupture propagation velocity
• slip velocity time function
• rake angle.
Other parameters related to the seismic properties specifically include the following:
@Seismicisolation
@Seismicisolation
• rupture start point

• rupture mode.
Figure 4.2.1 shows the main fault parameters. This example schematically shows one
macroscopic fault plane with one asperity. However, actual cases can include multiple fault
planes and multiple asperities, with microscopic parameters individually allocated to each
asperity and background region.
North
Ground surface
Dip direction
Origin of surface trace φ
Strike θ Dip
Surface trace
Origin of fault model angle δ
Fault Plane
Top depth t
Area S
Seismogenic Slip amount D
zone Asperity Seismic moment Mo
Bottom depth b Hypocenter

Width W Background
Length L
Fig. 4.2.1 Fault parameters
One of the two types of seismic hazard maps prepared and published by HERP based on
their long-term evaluation is the "seismic hazard maps for specified seismic faults", which
are maps that show the amplitude of seismic motion produced at the same time in a certain
region. In the failure scenario of an identified postulated earthquake, calculations by a
method based on the hybrid method (described later) obtains the amplitude of motion at
each point2), 3). Advanced seismic motion evaluation of a region is possible by using details
of the various properties of the source, path, and site, and so time histories on engineering
bedrock have been evaluated. Identified earthquakes or identified rupture scenarios are
postulated in advance, so different evaluation results can be obtained for multiple
earthquakes and multiple rupture scenarios. In addition, the results have been verified by a
simplified method based on the ground motion prediction equations.
The setting results for various postulated earthquakes are presently being published, for
example, in the investigations of the Central Disaster Management Council of the Cabinet
Office5), earthquake disaster prevention investigations of each local government, and
research reports from various projects. Most investigations adopt the HERP method of
setting the postulated earthquakes, but many investigations add special measures in
accordance with the objectives of the study or the individual information of the region,
when necessary.
This concept of treating the epicenter fault surface, including asperities and the background
region apart from the asperities, in a binary manner has been deemed to be a generally
appropriate fault model for magnitude 7 class earthquakes. However, this is being reviewed
@Seismicisolation
@Seismicisolation
as a result of magnitude 9 class earthquakes that have occurred throughout the world in
recent years. The results of analysis6) of the fault rupture process of the 2004 Sumatra
Earthquake (December 26, 2004, Mw 9.2), the 2010 Chile Earthquake (February 27, 2010,
Mw 8.8), and the 2011 Great East Japan (off the Pacific Coast of Tohoku) Earthquake
(March 11, 2011, Mw 9.0) showed that large slip occurs in the shallow regions near the
trench axis of a giant plate boundary and generates mainly long period seismic waves and
high tsunamis. Many short period seismic waves are generated in the deep regions close to
the coast and greatly contribute to the strong motions. In the future, it could be necessary, at
least for magnitude 8 and larger giant inter-plate earthquakes, to separately consider
• regions that mainly govern long period seismic waves and tsunamis by generating large
slip
• regions that mainly govern short period seismic waves by generating a large stress drop.
Clarifying the relation between the two regions and appropriate modeling will be an
important task in the future.
The results for the method of setting postulated earthquakes as described above will
provide a great perspective when setting the fault models for the postulated earthquakes at
the time of designing seismically isolated buildings.
(3) Generation method of earthquake ground motions
(a) Response spectrum fitting method
The response spectrum fitting method is a method of producing artificial ground

motions to fit a target response spectrum. Many empirical equations have been proposed
for response spectra, but this section explains response spectra7) based on "Notification
No. 1461, 2000: Stipulation of Criteria for Structural Calculations Performed to
Confirm the Safety Concerning Structural Strength of High-rise Buildings," Ministry of
Construction (currently the Ministry of Land, Infrastructure, Transport and Tourism).
This is the most widely used method for fitting a response spectrum (hereafter, the
Notification Spectrum), and the method of preparation of ground motions fitting this
response spectrum (hereafter, the notification wave) and some important points are
presented.
The following is an outline of the method of preparing the notification wave. First, the
acceleration response spectrum (damping coefficient 5%) of Table 4.2.1 is used as the
target response spectrum at free engineering bedrock (a stratum having sufficient
thickness and rigidity, with a shear wave velocity of about 400 m/s or higher). It is
verified that no damage occurred to the main parts of the building in terms of structural
resistance due to "rarely occurring ground motions," and that no toppling or collapse of
the building occurred due to "extremely rarely occurring ground motions." Other
conditions include the following: if the duration is 60 seconds or longer, the numerical
values of the ground motions (accelerations, velocities, or displacements) must be
clearly determined at appropriate intervals of time; multiple of seismic waves must be
formulated; amplification due to the surface strata must be appropriately considered;
horizontal motions in the direction normal to the ground motions, and the effect on the
@Seismicisolation
@Seismicisolation
deformation in the horizontal direction of the phase differences in the ground motions
and the loads in the vertical direction, must be appropriately considered.
Table 4.2.1 Japanese seismic code response spectra, free rock site (h=5%)
Acceleration Response Spectrum（m/s/s）
Period T（s）
Rare Earthquake（Level 1） Very Rare Earthquake（Level 2）
T<0.16 （0.64+6T）Z （3.2+30T）Z
0.16≦T<0.64 1.6Z 8Z
0.64≦T （1.024/T）Z （5.12/T）Z
Notes; T: Fundamental Period of the 1st Mode, Z: Seismic Zone Factor
The following sine wave synthesis method is normally used for waveform synthesis to
produce standard artificial ground motions.
=a (t ) E (t )∑ Ai cos(ωit + ϕi ) (4.2.1)
i
where a(t) is the time history waveform of the artificial ground motions, E(t) is a wave
function (envelope function of the waveform), and Ai, ωi, and φi are the Fourier
amplitude, circular frequency, and phase angle of the ith vibration, respectively.
The following is the method of synthesizing the notification spectrum. First, a

waveform, such as a random wave or a measured earthquake, is prepared as the starting
point. The initial waveform is set by inverse Fourier transform by using the notification
response spectrum as the Fourier amplitude spectrum of the acceleration waveform.
Next, the response spectrum of this waveform is compared with the notification
spectrum, and a new corrected waveform is obtained from the amplitude spectrum in
which each period is corrected. This operation is repeated until the required degree of
matching is obtained, to determine the final artificial ground motions.
The following four conditions are normally used as the conditions for fitting a response
spectrum (Sai) from an artificial ground motion to the notification spectrum (DSai):
(1) minimum response spectrum ratio: εmin= (Sai / DSai) min ≧ 0.85
(2) spectrum strength ratio: SIratio = ∑pSvi /∑DpSvi ≧ 1.0
(3) coefficient of variation of the spectral ratios: εCV = {∑(εi −1.0)2/ N}1/2 ≦ 0.05
(4) mean error value of the spectral ratio: εave=∑εi / N ≧ 0.98
In the above, pSvi and DpSvi are the pseudo-velocity spectrum of the Sai and DSai,
respectively, N is the number of period points. As a rule, the target period region is from
0.02 seconds to 5 seconds, but in sedimentary plains having deep sedimentary strata,
such as the Kanto Plain or Osaka Plain, it is considered that long period ground motions
dominate, so up to 10 seconds is considered. Also, the setting for spectral strength in (2)
is set to 1 to 5 seconds, or a similar value in accordance with the period of the building
under consideration, i.e., not 0.1 to 2.5 seconds, which is the original definition.
@Seismicisolation
@Seismicisolation
The Jennings function8) is normally used as the envelope function E(t) as follows:
 (t / tb ) 2 　　　　　　　　　　　 0 < t ≤ tb

 1.0 　　　　　　　　　　　　tb < t ≤ tc
E (t ) =  (4.2.2)
 { e fac
exp log ( A ) ⋅ (t − tc ) (td − tc )} 　tc < t ≤ td

 Afac {(te − t ) (te − td )} 　　　　td < t ≤ te
2
where tb, tc, td, and te are the rise time to the maximum amplitude, the time at the end of
the maximum amplitude, the duration, and the finish time, and the amplitude after te is 0
(see Fig. 4.2.2). Also, Afac is the amplitude ratio at time td with respect to amplitude 1 of
the main motion, and normally a value of about 0.1 is used. The specific proposal of
Jennings et al.8) for each time is tb = 4 seconds, tc = 35 seconds (15 seconds for a
medium sized earthquake), and td is 60 seconds or more in accordance with the
notification. With this method, long duration ground motions of more than several
minutes can be produced. Various empirical equations have been proposed for the
magnitude and epicenter distance, as parameters assigned for each time.
Next, an example of formulation of the notification wave and points to note are
described. First, Fig. 4.2.2 shows an acceleration waveform and response spectrum by
using random phases, as an example of producing an artificial ground motion. By
carrying out the convergence calculation 10 times, all the fitting conditions were
satisfied (εmin ≒ 0.91, SIratio ≒ 1.0, εCV ≒ 0.032, εave ≒ 1.0). However, in the wave in
Fig. 4.2.2, the amplitude of the rising portion and the end portion are significantly small
because the envelope function E was multiplied 10 times due to the convergence
calculation, and so the condition for 60 seconds duration was not satisfied. Methods to
improve this include the method of increasing the duration (td or te), and the method of
using an envelope function that converges more slowly. As an example, Fig. 4.2.3 shows
the results of using an envelope in which E is raised to the power of 0.2. Comparing the
waves of Figs. 4.2.2 and 4.2.3 shows virtually no difference in the main motion, but a
more gradual convergence is exhibited close to the original envelope function, and the
60 second duration is satisfied.
Figure 4.2.4 shows an example of calculation using the NS component of the Japan
Meteorology Agency (JMA) Kobe wave as an example of using the phase of an
observed wave. Apart from the phase, all the conditions were the same as those for Fig.
4.2.3. The matching conditions after convergence 10 times were (1) εmin ≈ 0.84, (2)
SIratio ≈ 1.0, (3) εCV ≈ 0.052, and (4) εave ≈ 0.98, which are considered to not be a
problem for practical purposes, but conditions (1) and (3) are not satisfied. Normally,
with an observed waveform with a short duration it is not easy to completely satisfy the
fitting conditions. Even by using the same notification spectrum, with shock waveforms
with short durations such as the Kobe wave, it is known that the building response
increases. Therefore, it is necessary for the structural engineer to appropriately judge the
phase used in the notification wave from the given fault and the various site conditions.
@Seismicisolation
@Seismicisolation
400 Simulated Acceleration 1000 (notification)

Target Spectrum (Code)
300 Target Envelope
Spectral Acceleration
Simulated Spectrum
acceleration (cm/s/s)
200 (10 Iterations)
(cm/s/s)
100
100
0
-100 0 20 40 60 80
time (s)
-200
-300 10
0.01 0.1 1 10 period (s) 100
-400
Fig. 4.2.2 Simulated accelerations and response spectra by using random phases
(tb=4, tc=35, td=60, te=80 s)
400 Simulated Acceleration 400 Simulated Acceleration

Target Envelope Target Envelope
300 Envelope for Each Iteration 300
Envelope for Each Iteration
200 200
100 100
0 0
-100 0 20 40 60 80 -100 0 20 40 60 80
time (s) time (s)
-200 -200
-300 -300
-400 -400
Fig. 4.2.3 Simulated accelerations of Fig. 4.2.4 Simulated accelerations of the

random phases by using the modified recorded phases at the JMA Kobe
envelope function station during the 1995 Kobe
Earthquake
The artificial ground motions that fit the notification spectrum have small components
on the long period side. In particular, when a random phase is assumed, these small
components do not contribute to the destructive power that produces large deformations
in a building. Therefore, the artificial seismic motions fitting the notification spectrum
are considered to be the minimum level ground motions. If an active fault is near the
site, or if the site is on a deep sedimentary stratum or weak soil, it is possible to produce
ground motions that can produce more destructive forces. If the building site is under
the conditions described above, in addition to the ground motions that fit the spectrum,
it is desirable to use the ground motions derived from a fault model as the site wave.
For the method of formulating the vertical motions by using the response spectrum
method, refer to Section 4.2.2(3).
(b) Method based on fault model
Methods of producing earthquake ground motions based on fault models are broadly
classified into the following methods:
@Seismicisolation
@Seismicisolation
- theoretical methods
- semi-empirical methods (empirical Green's function method and statistical Green's
function method)
- broadband hybrid method.
The following explanation summarizes the characteristics of each method9).
1) Theoretical methods
The ground motions at a building site are evaluated by obtaining the properties of the
seismic waves released from the seismic fault of a postulated earthquake based on fault
model theory, and calculating the path effect of the seismic waves and the site
amplification effects of the near-surface soils based on elastic wave theory. At present
obtaining the detailed information necessary for modeling the various properties is
difficult, and the evaluation accuracy in the short period region is relatively poor.
Methods include those for irregular media such as the "finite difference method," and
methods for horizontally stratified media such as the "wave number integration
method."
2) Empirical Green's function method
Figure 4.2.5 schematically shows the semi-empirical methods (empirical Green's

function method and statistical Green's function method). The empirical Green's
function method, which is one of the semi-empirical methods, considers the measured
record at a location during a small- or medium-size earthquake to be an "empirically
obtained Green's function" for evaluating the ground motions at a location in the case
that a large earthquake occurs.
Actually, a waveform is obtained by appropriately removing the effect of the

near-surface strata and the long period noise from the observed record for a small- or
medium-size earthquake at the location where the ground motions are to be evaluated.
Or, the measurement record for a small- or medium-size earthquake obtained near that
location is taken to be an element waveform (empirically obtained Green's function).
From these, the ground motions for a large earthquake are evaluated by superimposing
many of these waveforms based on the concept of the fault model. It is desirable that the
hypocenter of the small- or medium-size earthquake (element earthquake) is close to the
fault plane of the postulated large earthquake.
The frequency of occurrence of small- and medium-sized earthquakes is higher than that
of large earthquakes, so the possibility of obtaining a measurement record near the
evaluation site or near the site is greater in the case of a small- or medium-size
earthquake. Also, the measurement record of a small- or medium-size earthquake
includes the path effects of the seismic waves from the hypocenter to the site under
consideration and the soil properties near the site. So, if the properties of the ground
motions of small- or medium-size earthquakes and large earthquakes are linear, it is not
necessary to carry out a separate detailed investigation of the underground structure to
estimate the ground motions for the large earthquake. Because this method takes into
consideration the effect of the planar spread of the fault (fault model), it is possible to
@Seismicisolation
@Seismicisolation
reflect comparatively well the various properties of the individual source, path, and the
site.
Site
Ground surface Earthquake ground motion
Local site conditions

Seismic bedrock
Seismic Source Propagation path of seismic waves

Mesh elements of
Hypocenter Rupture front fault model
Seismic wave
Fault plane of large earthquake Small earthquake
Small earthquake
ground motion
Recorded motion is used for the empirical Green’s function method
Simulated motion is used for the stochastic Green’s function method
The slip amount and the effective stress of the large earthquake are estimated
Ground motion corresponding to one mesh element
Ground motion corresponding to each mesh element is estimated
The ground motions are

synthesized considering the
rupture process of the
large earthquake fault.
Evaluated ground motion

of the large earthquake
Time
Fig. 4.2.5 Conceptual explanation of semi-empirical methods

(empirical and stochastic Green’s function methods)
3) Statistical Green's function method
The major advantage of the empirical Green's function method is that the various
properties of the source, path, and the near-surface soils at the site are incorporated by
using the measurement record for a small- or medium-size earthquake obtained near the
site where the ground motions are to be evaluated. At the present time, however, it is not
always possible to obtain an appropriate measurement record for a small- or
medium-size earthquake near the site under consideration. In these cases, to apply the
empirical Green's function method, a statistically evaluated artificial ground motion is
used as the Green's function instead of the measurement record of the small- or
medium-size earthquake. This semi-empirical method is referred to as the statistical
Green's function method.
In the statistical Green's function method, an artificial ground motion corresponding to a

small- or medium-size earthquake is produced by using the spectra and time history
@Seismicisolation
@Seismicisolation
properties obtained by statistically processing several earthquake motion records

obtained in advance at many different sites. This is taken to be the element motion
(Green's function); by superimposing the records according to empirical Green's
function method, the planar effect of the fault in a large earthquake is reproduced, and
the ground motions in a large earthquake are evaluated. Therefore, the source properties
are taken into account by using the combination of theory and statistical analysis of the
record, the path properties are taken into account by using the statistical analysis of
several earthquake ground records, and the near-surface soil properties are taken into
account by applying theory or using the statistical analysis of several earthquake motion
records. If the soil amplification between the ground where the spectrum and the time
history properties are defined (for example, engineering bedrock) and the position where
the ground motions are to be evaluated cannot be ignored, it is necessary to separately
consider the soil amplification properties. The empirical Green's function method is a
method that enables a highly reliable evaluation of the ground motions that reflect the
properties of the source, path, and the near-surface soils at the site, if sufficient
qualitative and quantitative information and a measurement record of a small- or
medium-size earthquake can be obtained. In contrast, the statistical Green's function
method, which does not use any measurement record of a small- or medium-size
earthquake, has a relatively low evaluation accuracy for the particular ground motion
properties of the site under consideration. But, on the other hand, it is a method that
enables comparatively stable ground motions to be evaluated regardless of the quality or
quantity of information obtained for the site. Therefore, it is a very practical method
when used for purposes such as formulating ground motions for the design of a structure
or evaluating the ground motions for disaster prevention over a wide area.
The minimum information necessary to evaluate ground motions by the statistical

Green's function method is the ground structure directly below the site at depths
shallower than the depths for setting the element motions (configuration of strata, P
wave velocity, S wave velocity, and damping coefficient for each stratum), and the fault
parameters of the earthquake.
4) Broadband hybrid method
In this method, the ground motion in the long period range is calculated theoretically,
and the seismic motion in the short period range is calculated by a semi-empirical
method (empirical Green's function method or statistical Green's function method), and
the two are superimposed. In this way, it is possible to incorporate the advantages of the
theoretical method in the long period range and the advantages of the semi-empirical
method in the short period range. Therefore, if it is possible to obtain the necessary
information, the hybrid method is the best method in practice. The hybrid Green's
function method and the hybrid synthesis method depend on whether superposition of
the long period ground motion and the short period ground motion is carried out at the
stage of the element motion, or at the stage of the postulated ground motion.
5) Points to note regarding selection of seismic ground evaluation method
The ground motions for a postulated earthquake at the site under consideration are
evaluated by employing the source properties of the earthquake, the propagation
properties of the seismic waves, and the near-surface soil properties near the site.
Various information for the site and its surroundings is normally used to carry out a
@Seismicisolation
@Seismicisolation
higher-quality evaluation of the ground motions. The evaluation of strong motions,

which are low-frequency natural phenomena, will include variation, however small, so it
is considered that the factor that has the greatest effect on improving the overall quality
of the seismic motion evaluation is maximizing the information that can be obtained for
the site and adopting an appropriate ground motion evaluation method in accordance
with the quality and quantity of information.
Each of the various types of ground motion evaluation methods that have been
researched and developed to date require certain information; if that information cannot
be obtained, the ground motions cannot be evaluated. Also, even if the information can
be obtained, the accuracy of the evaluation of the ground motions varies depending on
the quality and the quantity of the information. Therefore, it is not practical to use just
one type of evaluation method for ground motions. Based on the fact that methods that
require a small amount of information have a lower evaluation accuracy of the ground
motions of the site under consideration, normally several evaluation methods that can be
adopted are prepared, and it is necessary to adopt the most appropriate method based on
rational judgment in accordance with the quality and quantity of specific information
obtained for the site under consideration.
As a practical problem, certain restrictions are imposed on the quality and quantity of
information that can be obtained and the combination of all the ground motion
evaluation methods using this information. The usefulness of methods differs, and
depends on the main target periods in the evaluation of the ground motions. At present,
direct methods are very useful for the evaluation of motions in the long period range,
and semi-empirical methods are very useful for the evaluation of motions in the short
period range. The broadband hybrid method, which combines the advantages of both, is
very useful for the evaluation of broadband motions. It is therefore considered that the
decision on what method to adopt should depend on the target periods of the motion to
be evaluated. Furthermore, to evaluate the motions uniformly at several locations in a
wide area by using a common method within the area (for example, for earthquake
disaster prevention over a wide area), a method in which a considerable amount of
high-quality information is reflected in the evaluation of the ground motions is not
necessarily appropriate. Methods that require a small amount of information have the
advantage that they are easy to use, so the decision regarding what method to adopt
should be made in accordance with the objective of the evaluation of the ground
motions.
Under these circumstances, the Earthquake Research Committee of HERP published the
"Strong Ground Motions Prediction Method (“Recipe”) for Earthquakes with Specified
Source Faults"4). This "recipe" summarizes the concept and standard procedures for the
four processes that are constituent elements of the strong ground motion prediction
method used by the Earthquake Research Committee when evaluating strong motions of
earthquakes occurring at active fault zones or subduction zone earthquakes. The
“recipe” consists of the following:
(1) setting each of the parameters of the specified source model

(2) preparing an underground structure model
(3) predictive calculation of strong ground motions
(4) verification of the results of the strong motion prediction.
@Seismicisolation
@Seismicisolation
An objective of this "recipe" is to establish a "standard methodology from which anyone

can obtain the same answer" for predicting strong ground motions with high accuracy
for scenario earthquakes (earthquakes for which the seismic fault has been identified).
The standard methodology will be modified or improved in the future as necessary.
Also, the "Generation Guide for Seismic Input Motions Based on the Recent
Advancement of Ground Motion Studies"10) has been published based on investigations
of the Research Subcommittee on Earthquake Ground Motion of the Steering
Committee for Structural Dynamics and the Research Committee on Structures of the
Architectural Institute of Japan. This guidebook has standardized the parts of the
prediction methods where possible. For the parts of prediction methods that cannot be
standardized, the guidebook provides assistance on preparation with suitable
information, with the aim that high-accuracy predicted ground motions over a wide
frequency range for specified sources and sites can be used in the design of buildings.
It is expected that the above methods of producing strong ground motions and the
predicted results obtained by using these methods will be widely used for evaluation of
the input ground motions for seismically isolated buildings.
(4) Evaluation of the near-surface soil properties
To apply ground motions produced at seismic bedrock or engineering bedrock to the input
ground motions affecting seismically isolated buildings, it is necessary to appropriately
model the ground shallower than the bedrock, where the ground motions were produced. It
is also necessary to evaluate the ground motions at the input location to the building by
taking into consideration the amplification properties of the ground by earthquake response
analysis.
Normally, when considering the soil properties of ground motions, it is convenient to

consider the problem in two stages: the deep ground from seismic bedrock to engineering
bedrock, and the near-surface soils that are shallower than engineering bedrock. In seismic
bedrock, which is the uppermost layer of the earth's crust, the S wave velocity reaches
about 3 km/s and the P wave velocity reaches nearly 6 km/s. Engineering bedrock has an S
wave velocity of about 400 m/s or higher and can support high-rise buildings, and a
common earthquake having a certain amount of spread is considered to occur in this area.
The soils from seismic bedrock to engineering bedrock are modeled with reference to
underground structural sounding results and investigative research results for that area. In
recent years, information, including that from the Japan Seismic Hazard Information
Station (J-SHIS)11), has become available on the Internet and can be actively used. If
necessary, it may be desirable to separately carry out an underground structural survey that
is close to the construction site. The P wave velocity and the S wave velocity are estimated
from existing public data or from the results of underground structural surveys, and the Q
value is frequently estimated by the analysis of earthquake measurement records on the
same plain or from knowledge based on investigations using the simulation of ground
motions. Normally, the behavior of deep soils is assumed to be linear, but it is necessary to
note that the Q value is greatly affected by the vibrations after the main motions.
Nonlinear phenomena such as reduction of rigidity and liquefaction can occur as a result of
strong ground motions in the near-surface soils that are shallower than engineering bedrock.
@Seismicisolation
@Seismicisolation
Therefore, it is necessary to appropriately model the nonlinear properties. Detailed

information of the ground is necessary to construct an appropriate soil structure model. In
addition to existing survey results, properties such as soil density, P wave velocity, S wave
velocity, stratum thicknesses, damping coefficients, dynamic deformation properties (G-γ
properties, h-γ properties, G: shear modulus, h: damping factor, γ: shear strain), should be
obtained from standard penetration tests carried out on site, seismic velocity logging (PS
well logging), steady state micro-tremor measurements, and laboratory tests from borehole
cores. If the information is insufficient, empirical equations are used to estimate the
necessary values for constructing the soil structure model; but, because the variation of
known information and equations is large, great accuracy cannot be expected.
Fundamentally, the near-surface soils are assumed to be horizontally stratified, and an
equivalent linear or nonlinear earthquake response analysis is carried out by taking into
consideration the nonlinear properties of the soil based on 1-dimensional wave transmission
theory. If liquefaction is a risk, the analysis should also take into consideration the
nonlinear properties of the soil due to liquefaction by using an effective stress analysis that
inputs the pore water pressure.
References
1) The Headquarters for Earthquake Research Promotion：Evaluation of Earthquakes
（ http://www.jishin.go.jp/main/p_hyoka.htm ）
2) The Headquaters for Earthquake Research Promotion, Earthquake Reseearch Committee：National Seismic Hazard
Maps for Japan (2009)（ http://www.jishin.go.jp/main/chousa/09_yosokuchizu/index.htm ）
Maps for Japan (2010) （ http://www.jishin.go.jp/main/chousa/10_yosokuchizu/index.htm ）
4) The Headquaters for Earthquake Research Promotion, Earthquake Reseearch Committee：Strong Ground Motion
Prediction Method (“Recipe”) for Earthquake with Specified Source Faults
（ http://www.jishin.go.jp/main/chousa/09_yosokuchizu/g_furoku3.pdf ）
5) Cabinet Office：Disaster Management（ http://www.bousai.go.jp/5jishin/index.html ）
6) T.Lay, H.Kanamori, C.J.Ammon, K.D.Koper, A.R.Hutko, L.Ye, H.Yue and T.M.Rushing: Depth-varying Rupture
Properties of Subduction Zone Megathrust Faults, Journal of Geophysical Research, Vol.117, B04311, 2012
7) Izuru Okawa, Kazuo Dan, Masanobu Tohdo：Seismic Evaluation of Buildings by Acceleration Response Spectrum
at Engineering Bedrock.：Part.4 Evaluation Spectrum Considering Local Seismicity, Summaries of Technical
Papers of Annual Meeting, Architectural Institute of Japan, pp.1131-1132, 1999
8) P.C. Jennings, G.W. Housner, N.C. Tsai: Simulated Earthquake Motions for Design Purposes, A1.pp.145-160,
4WCEE, 1969
9) Architectural Institute of Japan：Recommendations for Loads on Buildings（2004）
10) Architectural Institute of Japan：Generation Guide for Seismic Input Motions Based on the Recent Advancement of
Ground Motion Studies ( 2009)
11) Japan Seismic Hazard Information Station（ http://www.j-shis.bosai.go.jp/ ）
@Seismicisolation
@Seismicisolation
4.2.2 Points to Note Regarding Evaluation of Input Ground Motions
(1) Long period ground motions
Long period ground motions are an important issue requiring care when designing long
period structures, such as seismically isolated structures and high-rise buildings. The range
of periods normally referred to as "long period" in connection with these structures is
actually the short period region in the field of seismology. So, in the field of seismic
engineering, a range of periods from about 1–2 seconds to several tens of seconds is often
referred to as "relatively long period." Historically, attention has focused on the causes of
several characteristic incidents of damage, such as the following: the 1983 Nihonkai-Chubu
Earthquake (M7.7), where oil leaked from large oil tanks in Niigata; the 1985 Michoacan
Earthquake (M8.1), where major damage was caused in Mexico City (about 400 km from
the epicenter); the 2003 Tokachi Oki Earthquake (M8.0), where fires occurred at oil tanks at
Tomakomai (about 200 km from the epicenter). Also, the fact that when an earthquake
occurs near the Izu Peninsula, long period ground motions (fairly long period ground
motions) are predominant, as has been confirmed from seismic measurements.
In recent years, the Earthquake Research Committee of the HERP has been studying
methods of predicting long period ground motions and methods of publicizing the
prediction results. They prepared a provisional long period seismic motion prediction map
from the Kanto region to the Kinki region for the postulated Tokai Earthquake and
Tonankai Earthquake (1944 Tonankai Earthquake type) and a provisional long period
ground motion prediction map from the center of the Tohoku region to the Kanto region for
the Miyagi-ken Oki Earthquake (1978 Miyagi-ken Oki Earthquake type) as interim results.
These predictions were published in the provisional publication "Probabilistic Seismic
Hazard Maps" in 20091). Next, studies were undertaken to improve the source models and
the underground structure models, to further improve the long period ground motion
prediction maps. In 2012 they published the provisional publication "Probabilistic Seismic
Hazard Maps" from the west of Chubu region to the Kyushu region for the Nankai
Earthquake (1946 Nankai Earthquake type)2).
These provisional publications are a first step towards future prediction of broadband
ground motions, including long period ground motions for various postulated earthquakes.
Postulated earthquakes need to include giant subduction zone earthquakes that can cause
destruction over a wide area. Because the emphasis has been placed on reproducing
earthquakes that have occurred comparatively recently, the largest earthquakes in history or
the postulated largest earthquakes have not been the target. Technical issues remain in the
construction of characterized source models for future prediction, and the areas over which
underground structure models have been constructed do not cover the whole country.
Verification and improvement of the model accuracy are also issues. In the 2009
provisional publication, the calculated periods were limited to 3.5 seconds or higher due to
numerical calculation problems, and it is very possible that the ground motions are
under-evaluated for periods less than 5 seconds. Furthermore, in the 2012 provisional
publication, the calculated periods were limited to 2 seconds or higher, and so it also is
possible that the ground motions are under-evaluated for periods less than 3 seconds.
Therefore, issues still remain regarding the use of these predictions as input ground motions
for design, and studies are currently underway to rectify these issues, but caution is
necessary.
@Seismicisolation
@Seismicisolation
However, the Ministry of Land, Infrastructure, Transport and Tourism is performing

surveys and studies into ground motions to take into consideration long period motions for
the design of buildings. They have prepared and published draft countermeasures against
long period ground motions for seismically isolated buildings and super high-rise buildings,
and have invited opinions on implementing these measures in the future. The operation of
the Ministerial approval system is being reviewed. Moreover, structural calculations using
design ground motions that take into consideration the three earthquakes in the 2009
provisional publication "Probabilistic Seismic Hazard Maps"1) by the Earthquake Research
Committee referred to above (the postulated Tokai Earthquake, the Tonankai Earthquake,
and the Miyagi-ken Oki Earthquake) will be required, explanation of design measures to
prevent toppling of furniture will be required, and measures such as retrofit that are deemed
necessary will be required. Efforts are being made to provide reference information for
carrying out a design margin against earthquakes other than the above three earthquakes or
multiple coupled earthquakes.
The ground motion evaluation method described in the previouse paragraph has a broad
period range between 0.1 to 10 seconds, and is based on many seismic records from all over
Japan. The absolute acceleration response spectrum (damping coefficients 5%, 1%) and the
energy spectrum (damping coefficient 10%) that represent the amplitude properties, and the
group delay time and standard deviation that represent the phase properties are empirical
equations3),4). For the postulated earthquake, if the magnitude, position, shape, rupture start
position of the fault surface, and the target site have been determined, the amplification
properties of the site properties of the target site and the phase properties can be
conveniently calculated, so the advantage is that this evaluation method could possibly
calculate the time history of the ground motions.
As a result of the 2003 Tokachi Oki Earthquake (M8.0), interest in long period ground
motions associated with the occurrence of giant subduction zone earthquakes has increased.
In 2004, AIJ organized the "Special Research Committee into the Response to Major
Damage in the Tokai Earthquake." The research results over two years were compiled into
the publication “Structural Response and Performance for Long Period Seismic Ground
Motions” in which ground motions for the Tokyo area, Nagoya, and Osaka were predicted
in the case of the postulated occurrence of the Kanto Earthquake, the Tokai Earthquake, the
Tokai/Tonankai Earthquake, and the Nankai Earthquake, by using the ground motion
preparation method described in item (3) Generation Method of Earthquake Ground
Motions, and provided as the common wave in Table 4.2.2.
When designing a seismically isolated building, it is necessary to refer to and study the
varied knowledge about the characteristics of long period ground motions that could occur
around the site. It is particularly necessary to sufficiently take into account how much
excitation of ground motions in a specific period range could occur and for what reason.
Sufficient consideration must be given to the predominant period of the ground motions
produced by the seismic waves emitted from the seismic fault and its asperities, the
predominant periods of the seismic motions produced in the deep underground structure
around the site, and the natural periods of the seismically isolated structure; these
conditions will indicate careful decisions to be made, including whether to adopt a
seismically isolated structure.
@Seismicisolation
@Seismicisolation
Table 4.2.2 List of predicted earthquake ground motions5)

Type Time
Earthquake Developer Location Component Code Title Note
Acc. Vel. increment (s)
Japan Meteorological Kanto Earthquake, Sato, Japan Meteorogical

NS TS-TOK-NS ○ ○ 0.02 1st release
Agency, Tokyo Agency, Toky, NS
Kanto Sato, Dan
Minato Mirai NS TS-YKL-NS Kanto Earthquake, Sato, Minato Mirai, NS ○ ○ 0.02 1st release
Shinjuku NS KH-SNJ-NS Tokai Earthquake, Hijikata, Shinjuku, NS ○ ○ 0.01 2nd release
Tokai Hijikata et al. Shin-Toyosu NS KH-STY-NS Tokai Earthquake, Hijikata, Shin-Toyosu, NS ○ ○ 0.01 2nd release
Yokohama NS KH-YKH-NS Tokai Earthquake, Hijikata, Yokohama, NS ○ ○ 0.01 2nd release
Chubu Regional Tokai-Tonankai Earthquake, Chubu,

Sannomaru EW C-SAN-EW ○ － 0.01 1st release
Bureau Sannomaru, EW
Tokai- Tokai-Tonankai Earthquake, Aichi-ken, Nagoya

Nagoya Station EW A-NST-EW ○ － 0.01 2nd release
Tonankai Aichi Building St., EW
Housing Center
Tokai-Tonankai Earthquake, Aichiken,
Minakami Branch EW A-SJB-EW ○ － 0.01 2nd release
Minakami Branch, EW
Osaka District Nankai Earthquake, Kamae, Oshaka District

NS KK-OSA-NS ○ － 0.01 1st release
Meteorological Observatory Meteorogical Observatory, NS
Nankai Kamae, Irikura
Nishi-Ohsaka EW KK-WOS-EW Nankai Earthquake, Kamae, Nishi-Oshaka, EW ○ － 0.01 1st release
Nankai Earthquake, Sekiguchi, Average level,

Fukushima NS HS6-FKS-NS ○ － 0.02 2nd release
Fukushima, NS
HS6-OSK005- Nankai Earthquake, Sekiguchi, Average level,

Nankai Sekiguchi et al. K-NET Ohsaka EW ○ － 0.02 2nd release
EW K-NET Oshaka, EW
HS6- Nankai Earthquake, Sekiguchi, Average level,

KiK-net Konohana NS ○ － 0.02 2nd release
OSKH02-NS KiK-net Konohana, NS
○: released, －: not released
(2) Duration time
The properties of earthquake ground motions can be explained by three factors: amplitude
(strength of motions), period properties, and time properties (duration of motions).
Considerable research has been carried out into the amplitude and period properties of
broadband ground motions including long period components, and their quantitative
evaluation and treatment in seismic design have permeated down to structural engineers.
However, the discussion regarding the quantitative evaluation and treatment in the seismic
design of the time properties has not advanced sufficiently. However, in the case of the
2011 Great East Japan Earthquake (off the Pacific Coast of the Tohoku Earthquake; March
11, 2011, 14:46, MW 9.0), for example, the duration time of the motions in Tokyo, which
was distant from the seismic fault, was extremely long. Foreshocks, aftershocks, and
induced earthquakes continued, and it is considered that the repeated motions over an
extremely long time due to this series of earthquakes caused various damages to buildings
and their surrounding environment to expand. Therefore, it is desirable that the time
properties of the input ground motions be taken into consideration when designing
seismically isolated buildings in the future.
In the following, an example of a quantitative analysis of the period properties of the

earthquake ground motion duration is introduced. This example uses a response duration
time spectrum of the ground motions of the 2011 Great East Japan Earthquake recorded in
Tokyo6).
If a certain ground motion is input to a single-degree-of-freedom system (single mass

system) with damping coefficient h and natural period T [s], and if the value of the
@Seismicisolation
@Seismicisolation
cumulative squared amplitude of the response time history is obtained at each instant, the
values normalized by the value of the cumulative squared amplitude for the whole
waveform is referred to as the response duration time spectrum TS(h, T) for the normalized
cumulative squared amplitude from p1 to p2. In this example, p1 = 0.03, p2 = 0.95, and
damping coefficient h = 0.05. The method of obtaining the response duration time spectrum
is shown in Fig. 4.2.6. Examples of acceleration time histories of ground motions recorded
in Tokyo for the Great East Japan Earthquake are shown in Fig. 4.2.7. Examples of the
corresponding acceleration response spectra SA [cm/s/s], acceleration response duration
time spectra TSA [s], velocity response spectra SV [cm/s], velocity response duration time
spectra TSV [s], displacement response spectra SD [cm], and the displacement response
duration time spectra TSD [s] are shown in Fig. 4.2.8. In each case, the response duration
spectra are between 100 to 200 seconds for the short period range of periods less than 1
second, but the response duration spectra for periods 1 second and longer increase as the
period increases, and the response duration spectra for periods of about 2 seconds or higher
exceed 200 seconds (more than about 3 minutes). In particular, the duration of the
horizontal component is longer than the vertical component, and peaks occur at periods of
about 10 seconds and at periods of slightly less than 20 seconds of about 600 seconds (10
minutes) or more of durations.
Taking into consideration that the damping coefficient of a seismic isolation system is in the
region of 0.1 to 0.2, calculations carried out for greater damping coefficients would greatly
reduce the value of each response spectrum. On the other hand, even if the values of each
response duration time spectrum were reduced somewhat, the values on the longer period
range in particular would still be large.
Fig. 4.2.6 Procedure to calculate response duration time spectrum TS 6)
@Seismicisolation
@Seismicisolation
Fig. 4.2.7 Examples of ground motion acceleration time histories recorded in Tokyo during
the 2011 off the Pacific Coast of Tohoku Earthquake 6)
Fig. 4.2.8 Examples of acceleration response spectra SA [cm/s2], acceleration response

duration time spectra TSA [s], velocity response spectra SV [cm/s], velocity response duration
time spectra TSV [s], displacement response spectra SD [cm], and displacement response
duration time spectra TSD [s], calculated by using ground motions recorded in Tokyo during
the 2011 off the Pacific Coast of Tohoku Earthquake 6)
In the design of seismically isolated buildings, it is desirable that not only the amplitude of
the motions be considered but also the duration time of the motions by taking into
consideration the various performances required of the building and the necessary measures
that should be taken.
@Seismicisolation
@Seismicisolation
(3) Vertical motions
In this section, the generations of vertical motions are explained, and the response spectrum
fitting method and the methods based on fault models as ways of determining the vertical
motions for practical input ground motions are explained. Normally, the ratio of the
amplitude of vertical motions relative to horizontal motions is about 1/2 to 1/3, but the
closer the seismic source is, or the greater the amplitude of the short period components is,
the greater the vertical motions tend to be. As a result of the enhancement of the seismic
observation network in recent years, many large acceleration records with vertical motions
exceeding 1G have been obtained. For example, near the epicenter of the 2008
Iwate-Miyagi Nairiku Earthquake, an upward acceleration of 3866 Gal was measured. This
was a special case explained as an effect of the near-surface soils acting like a trampoline7).
The reason that the vertical motions became relatively larger during large acceleration near
to the seismic source is because the effect of nonlinearities of the near-surface soils on the
horizontal motions was great.
Figure 4.2.9 shows the concept of the generation of vertical motions in the case of a
measurement site on a sedimentary basin or plain. The initial portion of the seismic motion
is the P wave emitted from the seismic bedrock. The P wave is measured after the
amplification effect within the sedimentary strata. However, the main motion is considered
to be the SV wave that is emitted at an inclination from the seismic bedrock and converted
into a P wave amplified within the sedimentary strata to become the seismic motions.
Normally, the main component of the succeeding wave is predominantly a surface wave
(Rayleigh wave) of the sedimentary stratum that is emitted from the seismic motion from
the ends or bottom of the basin into the sedimentary strata.
Site
Surface layers
Engineering
Surface (Rayleigh) waves Surface (Rayleigh) waves
bedrock
P wave SV wave Surface (Rayleigh) waves
Fig. 4.2.9 Generation of vertical motions on a sedimentary basin or plain
The response spectrum fitting method normally uses an empirical vertical motion response
spectrum. For example, the Building Research Institute, Japan (BRI) provides a standard
response spectrum for horizontal motions and vertical motions. The ratio of the vertical
motion to the horizontal motion is about 0.6 on the short period side, becoming 0.5 for
periods of about 0.2 seconds or more. On the other hand, in Reference 8, the response
@Seismicisolation
@Seismicisolation
spectrum of the vertical motion is set by multiplying the amplitudes of the horizontal
response spectrum by 1/2, and the periods by 2/3. Another method uses the many distance
attenuation equations proposed in recent years for response spectra including vertical
motions7). Therefore, by appropriately selecting a distance attenuation relation based on the
magnitude of the target earthquake, the epicenter distance, and the site ground conditions, it
is possible to formulate the response spectrum for the vertical motions from the amplitude
ratio of the vertical and horizontal motions. For the phase spectrum, a random phase that is
independent of the horizontal motion is normally used, but another method uses the phase
of the waveform measured at the same time as the horizontal motions.
Methods based on fault models are handled differently in theoretical methods and empirical
methods. In theoretical methods used for calculating ground motions on the long period
side, the strong ground motions near the fault and the long period ground motions are
calculated theoretically, so they can be used as they are. On the other hand, in the empirical
methods used for calculating the ground motions on the short period side, the vertical
motions of the measurement records are used, as is the same for horizontal motions in the
empirical Green's function method (semi-empirical Green's function method). In the
statistical Green's function method, normally the horizontal motions are calculated by using
the SH wave, so various devices are necessary to calculate the vertical motions. For
example, it is considered that the vertical motion of the main motion is an S wave, as
shown in Fig. 4.2.9. So, one method obtains the amplification factor of the sedimentary
strata due to the SV wave that is emitted at an inclination in order to calculate the vertical
motions. In this case, the sedimentary strata are normally assumed to be horizontal strata,
and the amplification factor is calculated theoretically.
In the case of a site nearly directly above the epicenter, the ground motions are emitted
almost vertically into the sedimentary strata, so almost no conversion of the SV wave into a
P wave is needed, and the vertical motions are also very small. Therefore, as long as the
ground is simply horizontally stratified, it is difficult for vertical motions to be produced in
the main motion. The actual structure of the soil is complex, so it is considered that the
ground motions are scattered, reflected, refracted, and emitted at an inclination into the
sedimentary strata in many directions. In practice, if a measurement record close to the site
can be used, the empirical site amplification ratio can be used9), and if not, a distance
attenuation relation can be used.
(4) Ground motions near faults
Very characteristic ground waves are sometimes measured near active faults. These ground
waves include ground motions (directivity pulses) with the power to collapse a building by
overturning, large amplitude permanent displacement (fling step) due to the emergence of a
surface fault, etc10). In this section, the characteristics and points to note regarding ground
motions near faults are described.
First, the factors generating random waves and directivity pulses are explained as the
characteristics of ground motions near faults. Figure 4.2.10 shows a lateral fault viewed
from above, and the rupture propagation in the fault surface indicated by the bold line is
assumed to be transmitted from left to right. If fault rupture is assumed to be at a point, its
force system is equivalent to a double couple (two sets of couples), so the epicenter having
S wave releasing properties moves on the fault surface as shown in the figure. In particular,
at a measurement point near the fault surface, the component parallel to the fault surface, as
@Seismicisolation
@Seismicisolation
indicated by the bold arrow symbol at the epicenter, is virtually 0, but the normal
component has the maximum amplitude, so the component normal to the fault surface is
very large. Therefore, the fault rupture progresses to a measurement point, such as
measurement point B near the fault surface. If the velocity of the propagation of the rupture
and the S wave propagation velocity are almost equal, the pulse-shaped ground motions
generated at each point of the fault are constructively superimposed to form a large
amplitude ground motion with destructive force. This ground motion is referred to as a
directivity pulse having directivity in the normal component exhibited at measurement
points near the fault surface, and this phenomenon is referred to as Near Fault Rupture
Directivity (NFRD). Particularly, in the 1995 Southern Hyogo Prefecture Earthquake, the
seismic motions caused destructive damage to Kobe City. On the other hand, the rupture
propagation was away from measurement point A. So, the ground motions generated from
each point had short duration and were superimposed at random. Also, the coherence was
lost, and a relatively small amplitude random wave was formed. The El Centro wave
measured near to the active fault in the 1940 Imperial Valley Earthquake is a representative
example of this type of wave.
Fault normal components of seismic waves
Site A Site B
Right lateral fault
Radiation pattern of Direction of rupture

S waves
Fig. 4.2.10 Causes of forward directivity pulses near lateral seismic fault
Figure 4.2.11 is a diagram explaining the cause of directivity pulses in reverse faults and
lateral faults. In the case of the lateral fault (a), at a measurement point B near the
propagation of the fault rupture, a large amplitude directivity pulse is generated normal to
the fault surface. In the case of the reverse fault, if the fault surface has a high dip angle, as
shown in (b), rupture starts from the bottom of the fault, and the necessary condition is
propagation upwards. At measurement point C near the extension of the fault surface, a
directivity pulse is generated in a direction normal to the fault surface. On the other hand, it
is difficult for a directivity pulse to be generated, so normally a random wave is
predominant in the following cases: at measurement point A, from which the rupture is
propagating away; at measurement point D that is distant from the extension of the fault
surface, or when the reverse fault has a low dip angle as shown in (c).
Normally at about 10 to 20 km away from the fault surface, depending on the predominant
periods in the pulse wave, the coherence of the pulse wave collapses due to heterogeneity
of the ground, and random waves are predominant.
Calculation of directivity pulses is incorporated into methods using seismic fault models. In
this case it is necessary to be aware that directivity pulses are generated from asperities
(areas where slip rupture is particularly large), not from the fault surface as a whole. For the
method of setting asperities and source parameters, refer to the strong ground motion
prediction “recipe” proposed by HERP.
@Seismicisolation
@Seismicisolation
Ground surface Site B Site C Site D Site E
Site A
Rupture
Low dip angle
Hypocenter High dip angle Hypocenter
Hypocenter
(a) Lateral fault (b) Reverse fault with (c) Reverse fault with
high dip angle low dip angle
Fig. 4.2.11 Causes of forward directivity pulses for various shallow seismic faults
If small asperities with a large stress drop (virtually equivalent to the amount of slip) are
arranged as shown in Fig. 4.2.12, a directivity pulse can be measured at measurement point
A, but in some cases, pulse-shaped strong motions are also measured at measurement point
B. It is considered that the strong ground motion pulse measured at Kashiwazaki City in the
2007 Niigata-kin Oki Earthquake is equivalent to this case.
Site A Site B
Asperities
Hypocenter
Fig. 4.2.12 Causes of strong motion pulses from small asperities with high stress drop
When the epicenter is shallow and the scale of the fault is large, a surface fault appears, and
the nearby area is referred to as a fling step. The ground motions measured are
accompanied by a large permanent displacement in the form of a step function. As shown in
Fig. 4.2.13, in the case of a lateral fault in (a), a fling step is produced by slip in the
opposite direction and sandwiches the fault. In the reverse fault in (b), a downward
settlement of the lower ground and an upward rise of the upper ground are produced, and
the fault is sandwiched in between. In a reverse fault, normally the lower ground is on the
plain side and the upper ground is on the mountain side, but the displacement due to the
fling step is very large on the upper ground side. Permanent displacement due to a fling step
exhibits a large attenuation with the distance in inverse proportion to the square of the
distance; in this phenomenon, a large amplitude appears very close to the surface fault11).
Normally, long periods are predominant in fling steps. Particularly, in reverse faults, rupture
of the ground is on the upper ground side and may be associated with large tilting of the
ground. Photo 4.2.1 shows examples of damage to buildings directly above surface faults.
@Seismicisolation
@Seismicisolation
As shown, a building is torn at an expansion joint and produces large tilts. Therefore, if an
active fault that is evaluated to be highly hazardous is located very close to a construction
site, options such as reviewing the construction site should be considered. In places such as
California and New Zealand, active fault laws restricting construction near active faults
have been implemented. On the other hand, in Japan it is considered that no alternative to
construction in urban areas exists, so this risk is accepted, but it is necessary to take
measures in preparation of ground deformation such as fling steps or tilting. In this case,
not seismically isolating a building is an option10).
Site B Site D
Site A
Site C Hanging wall
C
Foot wall
(a) Lateral fault (b) Reverse fault

Fig. 4.2.13 Causes of fling steps above seismic surface faulting
(a) Separated building (b) Collapsed building (c) Tilted building

Picture 4.2.1 Building damage at surface faults in the 1999 Chi-Chi, Taiwan earthquake
(5) Phase difference input
In buildings with length exceeding 100 m and a long plan shape, it may not be possible to
ignore the phase difference of the ground motions produced at the two ends of the building.
Figure 4.2.14 shows the various causes that produce the phase difference input.
Site A Site B Site C

Different
Soil condition
Body waves
Surface wave Scattered
with inclined
seismic waves
（ incident angles
Bedrock
Fig. 4.2.14 Causes of input ground motions with different phases
@Seismicisolation
@Seismicisolation
At site A, the ground structure differs greatly, so a difference in amplitude as well as in

phase may be produced. In this case, a numerical analysis method such as the finite element
method can be used, or an input bedrock surface can be provided at a suitable depth, and
ground motions that take into consideration the amplification factor of the ground structure
directly below each point can be used.
In site B, a phase difference is produced by the main wave being emitted at an inclination
or by the coherence properties of the surface wave. In this case, methods have been
proposed that consider surface waves for long periods and body waves input at an
inclination for short periods12).
In site C, a phase difference is produced by incoherent scattering due to the heterogeneity

of the ground. The coherence function yjk at frequency f between point j and point k is given
by the following equation.
S (f) Im(γ jk )
γ jk ( f ) = jk γ jk ( f ) exp ( i ⋅ θ jk ( f ) ) 　、θ jk ( f ) =
= tan −1 (4.2.3)
S jj ( f ) S kk ( f ) Re(γ jk )
where Sjk is the cross-spectral density of the waveform at the two points, Sjj and Skk are the
power spectral densities, i is an imaginary number. The absolute value of the coherence
function is close to 1 when the coherence is high, such as at low frequencies when the
distance between the two points is short, and close to 0 when the coherence is low. On the
other hand, the phase angle θij is 1 for vertical plane wave input. But, when the incoming
wave is inclined, the apparent velocity Vapp of propagation is given by the following
equation.
θ jk ( f ) = 2p ⋅ f ⋅ rjk Vapp (4.2.4)
where rjk is the horizontal distance between the two points. Various equations have been
proposed for empirical coherence functions using array measurement records. A typical
example from Reference 13 is as follows.
tanh {5.39 − 0.622ln(rjk )} e( −0.252 f ) + 0.35

γ jk ( f ) = (4.2.5)
Various examples of incoherent phase spectra and mock seismic motion setting methods
apply this equation13),14).
References
Maps for Japan (2009)（ http://www.jishin.go.jp/main/chousa/09_choshuki/index.htm ）
Maps for Japan (2012)（ http://www.jishin.go.jp/main/chousa/12_choshuki/index.htm ）
3) Toshimi Sato, Izuru Okawa, Takao Nishikawa, Toshiaki Sato, Matsutaro Seki：Prediction of Waveforms of
Long-period Ground Motions for Hypothetical Earthquakes Using Empirical Regression Relations of Response
Spectra and Phase Spectra, Journal of Structural and Construction Engineering, AIJ, No.649, pp.521-530, 2010.3
@Seismicisolation
@Seismicisolation
4) Building Research Institute：Study on the Safety Countermeasures for Super-High-Rise Buildings etc. Against the
Long-Period Earthquake Ground Motions, Building Research Data, No.127, 2010.12
（ http://www.kenken.go.jp/japanese/contents/publications/data/127/index.html ）
5) Architectural Institute of Japan：Structural Response and Performance for Long Period Seismic Ground Motions，
2007
6) Toru Ishi：Response Duration Time Spectra of Earthquake Ground Motions: Propasal of the Evaluation Method and
Its Application to the Ground Motions in Tokyo During the 2011 Off the Pacific Coast of Tohoku Earthquake, its
Foreshock, Aftershocks, and Induced Earthquakes, Journal of Structural and Construction Engineering, AIJ, No.676,
pp.843-850, 2012.6
7) Shin Aoi：Strong Motions of the 2008 Iwate-Miyagi Nairiku Earthquake-Asymmetric Ground Motion and
Trampoline Effect, Natural Disaster Research Report, No.43, National Research Institute for Earth Science and
Disaster Preveution, 2010
8) Architectural Institute of Japan：Structural Performance under Multi-dimensional Earthquake Ground Motions,
1998
9) Akira Kowada, Masaru Tai, Yoshinori Iwasaki, Kojiro Irikura：Evaluation of Horizontal and Vertical Strong Ground
Motions Using Empirical Site-Specific Amplification and Phase Characteristics, Journal of Structural and
Construction Engineering, AIJ, No.514, pp.97-104, 1998
10) Y. Hisada：Active Fault and Mitigation of Building Damage, Active Fault Research, No. 28, pp.77-87, Japan Society
for Active Fault Studies, 2008
11) Y. Hisada, J. Bielak: A Theoretical Method for Computing Near-Fault Strong Motions in Layered Half-Space
Considering Static Offset due to Surface Faulting, with a Physical Interpretation of Fling Step and Rupture
Directivity, Bull. Seism. Soc. America, Vol.93, No.3, pp.1154-1168, 2003
12) Toru Ishii, Mika Kaneko, Takeshi Sugiyama ： Input Earthquake Motions for Long-sized Structures with
Consideration of Propagation Directions and Velocities of Body Waves and Surface Waves, Proceedings of the
Japan Earthquake Engineering Synposium, pp.355-360, 1994
13) Abrahamson, N.A.: Generation of Spatially Incoherent Strong Motion Time Histories, Proc. 10th World Conf.
Earthquake Eng., pp.845-850, 1992
14) Hirotoshi Uebayashi, Yoshihiro Takeuchi：Input Motion of an Actual Rigid Embedded Foundation Due to Spatially
Variable Seismic Ground Motions, Journal of Structural and Construction Engineering, AIJ, No.529, pp.105-112，
2000
4.2.3 Earthquake Response Spectra
The effect of a load on a building can be represented by various types of spectra such as a
response spectrum or an energy spectrum.
A response spectrum shows the relation between the maximum response value of a
single-degree-of-freedom system structure having various natural periods and damping
coefficients, and the natural period. Therefore, a response spectrum is used for representing
@Seismicisolation
@Seismicisolation
the response properties (maximum response value) of elastic bodies.
On the other hand, an energy spectrum represents the total amount of input energy to a
single-degree-of-freedom system structure having various natural periods and damping
coefficients.
The equations of motion of a single-degree-of-freedom system subject to horizontal seismic

motions can be represented by the following equation.
My + Cy + F ( y ) = − Mz0 (4.2.6)
where M: mass of the system

C: damping coefficient
F(y): restoring force property
y: relative displacement of the system
z0: horizontal ground motion
By multiplying both sides of equation (4.2.6) by the increment of displacement during an

earthquake y = dy , and integrating with respect to time t, the following energy equilibrium
dt
balance can be obtained.
∫
t t t t
Myy dt + ∫ Cy 2 dt + ∫ F ( y ) y dt = − ∫ Mz0 y dt (4.2.7)
0 0 0 0
Equation (4.2.7) shows the energy balance at time t. If the duration of the ground motion is t0,
and if t=t0 in equation (4.2.7), the following equation for energy balance at the end of the
earthquake is obtained.
We + W p + Wh = E
t
E = − ∫ Mz0 y dt : total energy input
0
t
Wh = ∫ Cy 2 dt : energy absorbed by damping
0
t t
We + W p = ∫ Myy dt + ∫ F ( y ) y dt
0 0
We: elastic vibration energy

Wp: cumulative plastic strain energy
The total energy input corresponds to the total cumulative plastic strain energy absorbed by
the structure. For designing structures, the response spectrum and the energy spectrum are
indispensable.
@Seismicisolation
@Seismicisolation
In energy spectra, the total input energy E produced in the structure by the ground motions is
represented by the equivalent velocity VE defined in equation (4.2.8). The relation between
the natural period and the equivalent velocity VE of the total input energy to a single structure
by a single ground motion is referred to as the energy spectrum.
2E (4.2.8)
VE =
M
where M: total mass
The energy spectrum of an undamped elastic single mass system is the same as the Fourier
velocity amplitude spectrum of the ground motion. Also, the energy spectra for nonlinear
systems such as systems with damping and elasto-plasticity correspond to the Fourier
velocity amplitudes smoothed. The energy spectrum of a normal structure is represented by
the energy spectrum of an elastic single mass system with 10% damping1). Therefore, the
total energy input per unit mass of a structure due to a single ground motion is a stable value
that depends solely on the first natural period and is not dependent on the strength
distribution or stiffness distribution of the building. In short, the load effect of a ground
motion on a structure can be taken to be the input energy.
In the response prediction method based on the energy balance described in Section 4.3.1, the
input energy becomes the basic quantity of the design, and if the design energy spectrum is
determined, the response can be predicted. Also, correspondence between the energy
spectrum and the velocity response spectrum and the acceleration response spectrum is given
by equations (4.2.9) and (4.2.10), respectively.
VE (T ) (4.2.9)
VD (T ) =
1 + 3h + 1.2 h
2π (4.2.10)
A(T ) = VD (T )
T
where VD(T): velocity response spectrum

VE(T): energy spectrum
h: damping factor
A(T): acceleration response spectrum
(1) Design response spectra
Equation (4.2.11) represents the energy spectrum of a horizontal ground motion as the sum
of the loading affects input from two directions.
2(E NS (T ) + E EW (T )) (4.2.11)
VE (T ) =
M
@Seismicisolation
@Seismicisolation
where ENS(T): total input energy due to the NS component of the horizontal motion
EEW(T): total input energy due to the EW component of the horizontal motion
Bilinear type design energy spectra, as shown in Fig. 4.2.15, for the horizontal motions and
the vertical motions have been proposed. In a short period region, the equivalent velocity of
the design energy spectrum varies in proportion to the period, But, above a certain value of
period, the velocity becomes a constant value VEm1). The value VEm shown on the design
energy spectrum is a value determined by taking into consideration the soil conditions, and
the distance to the postulated epicenter (fault). In addition, it is necessary to take into
consideration the points indicated in Section 4.2.2 for setting the design energy spectrum.
VE
VEm
0 T
Fig. 4.2.15 Shape of an energy spectrum
(2) Correspondence to design velocity response spectrum
Here, the correspondence between the maximum velocity evaluation used in the input
ground motion level indicated in the Building Standard Law and the input level of super
high-risebuildings and seismically isolated buildings, and the energy spectrum are
described.
First, in the Building Standard Law, the seismic force is given as story shear force
coefficient Ci.
Ci = Z ⋅ Rt ⋅ Ai ⋅ C0 (4.2.12)
where Z: seismic hazard zoning factor

Rt: design spectral coefficient
Ai: lateral shear distribution factor
C0: standard shear coefficient
Z is taken to be 1.0, and if C0 = 1.0 is used for calculating the required horizontal load
resistance as the level against the strongest seismic input, A1 = 1.0, then the shear force
coefficient of the first story is given by
C1 = Rt (4.2.13)
Equation (4.2.13) can be considered to be the acceleration response spectrum of the ground
motions divided by the acceleration due to gravity. The relation between the acceleration
@Seismicisolation
@Seismicisolation
response spectrum and the velocity response spectrum of the ground motions is given by
equation (4.2.10), and the relation between the velocity response spectrum and the energy
spectrum is given by equation (4.2.9). So, the level of input ground motions indicated by
the Building Standard Law represented as an energy spectrum (VE, BCJ) is given by
(
VE，BCJ = 1 + 3h + 1.2 h ) g ⋅2Tπ⋅ R
t
(4.2.14)
The conversion value of the energy spectrum contributing to damage VD (= g.T.Rt/2π)

coincides with the velocity response spectrum. Figure 4.2.16 shows the output of the level
of input ground motions indicated in the Building Standard Law as the VD-T relation.
On the other hand, in the performance evaluation, the ground motions as the input level to
super high-rise buildings and seismically isolated buildings are standardized by the
maximum velocity, with 50 cm/s being an extremely rarely occurring ground motion.
Figure 4.2.17 shows the energy spectrum for an elastic single mass system with damping
10% due to an input wave that is a measurement wave standardized at a maximum velocity
of 50 cm/s.
Soil type1 JMA-Kobe(NS)

Soil type 2 El Centro(NS)
Soil type 3 Hachinohe(EW)
Taft(NS)
Fig. 4.2.17 Energy spectrum by the

Fig. 4.2.16 Energy spectrum
input wave, in which the normalized
contributing to the damage specified by
maximal velocity is 50 cm/s
Building Standard Law
(3) Response spectra for measured ground motions
Figure 4.2.18 shows the velocity response spectrum and the energy spectrum of a ground
motion measured near an epicenter. Many ground motions have been measured that exceed
the level of the Building Standard Law, as shown in Fig. 4.2.16. It is necessary to determine
as much as possible the ground motions that could occur in the area that is the target of the
design and to verify the response of seismically isolated buildings. Also, it has been
suggested that a spectrum for verification of long period ground motions has a maximum
value of about 1 to 1.5 times, and an input energy of two to five times compared with the
standard wave or notification wave2).
@Seismicisolation
@Seismicisolation
1994Northridge(Sylmar) 2004Niigata-Chuetsu (Kawaguchi)

1995Kobe (Takatori) 2007Niigata-Chuetsu-Oki (Kariwa)
1995Kobe (Fukiai) 2008Iwate-Miyagi (Naruko)
500 700
h=0.05 h=0.1
600
400
500
pSv (cm/s)
300
VE (cm/s)
400
300
200
200
100
100
0 0
0.1 1 10 0.1 1 10
PERIOD(sec) PERIOD(sec)
Fig. 4.2.18 Pseudo-velocity response spectrum (left) and energy spectrum (right) of an
earthquake record
References
1）H. Akiyama：Earthquake-Resistant Limit-state Design for Buildings, University of Tokyo Press, 1985
2）Architectural Institute of Japan：Structural Response and Performance for Long Period Seismic Ground Motions, 2007
@Seismicisolation
@Seismicisolation
Section 4.3 Seismic Response Prediction by Simplified Analysis Methods
Typical simplified response prediction methods include a response prediction method based
on energy balance (hereafter referred to as the energy balance method), and a response
spectrum method using the equivalent linearization method (hereafter referred to as the
response spectrum method).
In the energy balance method, the seismic input energy to buildings due to ground motion
and the energy absorbed by seismic isolation devices can be directly evaluated. From the
energy response, predictions for the maximum displacement and the cumulative plastic
deformation can be carried out.
The response spectrum method is a methodology based on the equivalent linearization

method, which assumes that the building maximum response is uniquely determined by the
equivalent linear stiffness and equivalent viscous damping factor corresponding to the
maximum response value, irrespective of the hysteresis loop up to the maximum response.
Therefore, extensions of this methodology and research into response prediction other than
this methodology have been reported, so in this section the following representative research
into highly accurate prediction methods are explained:
(1) response prediction of the seismic isolation level based on energy balance
(2) simplified response prediction method for the maximum response using an
equivalent single-degree-of-freedom system
(3) response prediction using response performance design charts.
(1) is a response prediction method1), in which the energy balance method2) proposed by
Akiyama is extended to systems using combined hysteretic and fluid type dampers.
(2) is a response prediction method3) proposed by Kasai et al. It is an extension of the

equivalent linearization method. This method assumes a maximum response value, evaluates
an equivalent period and equivalent viscous damping factor that represents the reduction in
stiffness of the building and the hysteretic energy absorption, and uses the results to predict
the maximum deformation and the maximum forces from the elastic response spectra.
(3) is a response prediction method4) proposed by Ishimaru and Hata et al., in which the
response spectrum of the ground motions and the seismic isolation level are provided as
parameters. The response value is predicted from charts that represent the viscous damping
factor, the ductility factor, and the bilinear coefficient. By using the energy spectrum, the
energy absorbed by the seismic isolation devices can be predicted. In addition, it is possible
@Seismicisolation
@Seismicisolation
to obtain the seismic isolation device parameters to satisfy the target response by reverse
design.
References
1) Sayaka Higashino, Haruyuki Kitamura：Energy balance Based Seismic Response Prediction Methods for Seismic
Isolated Buildings with Rubber Bearings and Viscous Dampers，Journal of Structural and Construction Engineering，
AIJ, No.588，pp.79-86, 2005.2
2) H. Akiyama：Earthquake-Resistant Limit-state Design for Buildings, University of Tokyo Press, 1985
3） Kazuhiko Kasai, Yoshinobu Kawanabe：Equivalent Linearization to Predict Dynamic Properties and Seismic Peak
Responses of a Structural System with High Viscous Damping and Hysteretic Damping, Journal of Structural and
Construction Engineering, AIJ, No.591, pp.43-51, 2005.5
4） Ippei Hata, Shinji Ishimaru, Jun Hasegawa：On Performance Based Design Method for Systems with Nonlinear Viscous
Damper and Elasto-Plastic Damper, Journal of Structural and Construction Engineering, AIJ, No.617, pp.47-54,
2007.7
@Seismicisolation
@Seismicisolation
4.3.1 Energy balance method
(1) Prediction of the response of the seismic isolation level based on energy balance
The energy balance method is an analysis method for evaluating the seismic resistance
based on the balance of the seismic input energy to buildings due to ground motions and the
energy absorbed by the building. The response prediction equation that enables the
combined effect of hysteretic and fluid type dampers to be evaluated is described here.
The seismic isolation level consists of isolators and hysteretic and fluid type dampers, and it
is assumed that these seismic isolation devices absorb all the seismic input energy, and the
superstructure does not absorb the seismic energy. In this case, the energy balance equation
can be expressed as follows.
f We + hWd + sWd = E (4.3.1)
where fWe: elastic strain energy of the isolators, hWd: energy absorbed by the fluid type
dampers, sWd: energy absorbed by the hysteretic type dampers, E: seismic input energy due
to the ground motions.
The seismic input energy E is often normalized by the building mass M and converted into
an equivalent velocity VE (= 2 E / M ) of the energy. The relation between the equivalent
velocity VE and the natural period of the building is the energy spectrum. In proposed
design energy spectra, VE has a constant value in the period range of seismically isolated
buildings, whose value is defined for each soil classification1).
fQ hQ sQ
f Qm sWd
hQ m s Qy
hWd
fWe
ks
kf
δm δ δm δ s δy δm δ
a) Isolators b) Fluid type dampers c) Hysteretic dampers

Fig. 4.3.1 Restoring force characteristics of seismic isolation devices
Each of the terms on the left side of equation (4.3.1) were formulated by the method of
Higashino and Kitamura2) as restoring force characteristics of the seismic isolation level to
derive the equation to predict the response of the seismic isolation level. In the case of
isolators with a damping function, such as high-damping rubber bearings or lead rubber
bearings, their restoring force characteristics are appropriately separated into an elastic
element and a perfectly elasto-plastic element, which can be modeled as shown in Fig.
4.3.1. Here, the various parameters of the seismic isolation devices are set as follows. The
stiffness of the isolators is kf, and the maximum shear force is fQm. Fluid type dampers are
directly proportional to velocity, so their damping coefficient is C, their damping factor is
(= C / 2 Mk f ) , and the maximum shear force is hQm. Hysteretic type dampers have perfectly
h　
@Seismicisolation
@Seismicisolation
elasto-plastic restoring force characteristics, with an elastic stiffness of ks, a yield

displacement of sδy, and a maximum shear force (yield strength) of sQy. Also, the maximum
shear force of the whole seismic isolation level is Qiso and the maximum displacement is
δm.
The maximum shear force coefficient (base shear coefficient) of the seismic isolation level
a1 and the maximum shear force coefficients af, ah, and as of the isolators, fluid type
dampers, and hysteretic type dampers are given by the following equations, where g is the
acceleration due to gravity.
α1 = Qiso / Mg , α f = f Qm / Mg , α h = h Qm / Mg , α s = s Qy / Mg (4.3.2)
The maximum shear force coefficient a0 and the maximum displacement δ0, if no dampers
are installed, are defined as the standard response using the seismic isolation period Tf and
the energy spectrum VE.
2πVE T f VE M
α0 = , δ0 = , T f = 2π (4.3.3)
Tf g 2π kf
The elastic strain energy fWe of the isolators is represented by the following equation.
2 2
1  α f  MVE
f We = f Qm ⋅ δ m =   ⋅ (4.3.4)
2  α0  2
The energy absorbed by fluid type dampers and hysteretic type dampers hWd, sWd can be
represented by the following equations as the equivalent numbers hn and sn of repetitions of
the steady state loop at the maximum displacement δm.
2
 α h   α f  MVE
hWd = h n ⋅ π ⋅h Qm ⋅ d m = 2π ⋅h n  ⋅  ⋅ (4.3.5)
 α0   α0  2
2
 α s   α f  MVE
sWd =4 s n⋅ s Q y ⋅ d m =8 s n  ⋅  ⋅ (4.3.6)
 α0   α0  2
By substituting (4.3.4) through (4.3.6) into the energy balance equation (4.3.1) and solving
for af/a0, the following equation is obtained.
2
αf   αh   α s    αh   α s 
= − π ⋅h n  + 4 s n  + π ⋅h n  + 4 s n   + 1 (4.3.7)
α0   α0   α 0    α0   α 0 
The maximum shear force coefficient a1 of the seismic isolation level, with consideration
of the phase difference of the shear force between the isolators and the fluid type dampers,
is expressed by the following equation.
α1 = α f 2 + α h 2 + α s (4.3.8)
@Seismicisolation
@Seismicisolation
If this is taken into consideration in equation (4.3.7), the response prediction equation for
the maximum shear force coefficient a1 of the seismic isolation level is obtained as follows.
2
 2  2
α1   αh   α s    αh   α s   αh  αs
= − π ⋅h n  + 4 s n  + π ⋅h n  + 4 s n  + 1  +   + (4.3.9)
α0    α0   α 0    α 0   α 0     α 0  α 0
   
VE 2.0
0.35 0.35 T f (s)
( / )
300 =4s
TTff=4sec as VEV=150cm/s
E=150cm/s
The maximum shear coefficient

of the seismic isolation level α1
0.15
0.30 0.30
200 0.01 3.0
0.02
0.25 0.10 0.25
150 0.08 0.03
0.06
0.20 0.04 0.20 0.04
100 0.06
0.03
0.02 4.0
0.15 0.01 0.15
5.0
0.10 0.10
6.0
0.08
0.05 0.05
as 0.10
0.15
0.00 0.00
0 10 20 30 40 50 60 0 10 20 30 40 50 60
The maximum displacement of The maximum displacement of
the seismic isolation level δm (cm) the seismic isolation level δm (cm)
(a) In the case of Tf = 4(s) (b) In the case of VE = 150(cm/s)

Fig. 4.3.2 Relation between maximum shear force coefficient α1 and
maximum displacement δm of the seismic isolation level (Hysteretic dampers only)
VE 0.6 0.46 Damping factor h

0.35 0.35
( / ) T f (s)
300
0.02
0.30 0.30 2.0 0.03

200 0.05 3.0
0.25 Damping factor h 0.25
0.6 0.1
150
0.20 0.46 0.4 0.20
0.3 0.2
100 0.2 0.3
0.15 0.1 0.15 4.0
0.4
0.01
0.10 0.10 5.0
6.0
0.05 0.05
T f =4s
Tf=4sec V =150cm/s
VEE=150cm/s
0.00 0.00
0 10 20 30 40 50 60 0 10 20 30 40 50 60
The maximum displacement of The maximum displacement of
the seismic isolation level δm (cm) the seismic isolation level δm (cm)
(a) In the case of Tf = 4(s) (b) In the case of VE = 150(cm/s)

Fig. 4.3.3 Relation between maximum shear force coefficient α1 and
maximum displacement δm of the seismic isolation level (Fluid type dampers only)
The damping factor h corresponding to this response is represented by the following

equation.
@Seismicisolation
@Seismicisolation
 2 
1  αh    αh   α s    αh   α s 
h=   π ⋅h n

 + 4 s n  + π ⋅h n  + 4 s n  + 1 

(4.3.10)
2  α0    α0   α 0    α0   α 0 
 
Also, if the maximum shear force coefficient a1 and the maximum displacement δm of the
seismic isolation level are obtained from the seismic isolation period Tf and the energy
spectrum VE, the following equations are obtained.
 2
  
α1 =
1 + 4h 2 
− 4 s n ⋅ α s + (4s n ⋅ α s )2 + (1 + 4π ⋅ h⋅h n ) 2πVE 
 + αs (4.3.11)
1 + 4πh⋅h n   Tf g  
 
 2 4 2
 Tf 
δm =
1 
− s n ⋅ α s   g + ( s n ⋅ α s )2  T f
  T f VE  
 g 2 + (1 + 4πh⋅h n )   (4.3.12)
1 + 4πh⋅h n  π  π   2π  
 
(2) Hysteretic dampers only (ah = 0)
If h = 0 in equations (4.3.11) and (4.3.12), then the relation between the maximum shear
force coefficient a1 and the maximum displacement δm of the seismic isolation level is
shown in Fig. 4.3.2 with the seismic isolation period Tf, the energy spectrum VE, and the
yield shear force coefficient of the hysteretic dampers as as parameters. However, it is
assumed that the equivalent number of repetitions sn = 2. The maximum shear force
coefficient a1 of the seismic isolation level has the minimum value when the yield shear
force coefficient as = 1.42VE/Tfg.
(3) Fluid type dampers only (as = 0)
If as = 0 in equations (4.3.11) and (4.3.12), then the relation between the maximum shear
force coefficient a1 and the maximum displacement δm of the seismic isolation level is
shown in Fig. 4.3.3 with the seismic isolation period Tf, the energy spectrum VE, and the
damping factor h as parameters. However, it is assumed that the equivalent number of
repetitions hn = 2. The maximum shear force coefficient a1 of the seismic isolation level
has the minimum value when the damping factor h = 0.46, independent of the seismic
isolation period Tf and the energy spectrum VE.
(4) Effect of combined use of hysteretic and fluid type dampers
The effect of combined use of hysteretic and fluid type dampers is described on the basis of
the response prediction equation (4.3.9). Figure 4.3.4(a) shows the relation between the
shear force coefficient of the fluid type dampers ah/a0 and the damping factor h when the
yield shear force coefficient of the hysteretic dampers as/a0 is varied between 0.0 to 0.4.
Figure 4.3.4(b) shows the relation between the yield shear force coefficient of the hysteretic
dampers as/a0 and the damping factor h when the shear force coefficient of the fluid type
dampers ah/a0 is varied between 0.0 to 0.3. In the figures, the solid lines denote the
maximum shear force coefficient of the seismic isolation level a1/a0, the broken lines
denote the shear force coefficient af/a0 of the isolators or the maximum displacement δm/δ0,
@Seismicisolation
@Seismicisolation
and the chain dotted line denotes the shear force coefficient ah/a0 for the fluid type
dampers in Fig. 4.3.4(a) (and the yield shear force coefficient as/a0 of the hysteretic
dampers in Fig. 4.3.4(b)). Also, the damping factor h, which is determined by the
combination of the damper shear force coefficients ah/a0 and as/a0, is shown. It is shown
that the maximum shear force coefficient a1/a0 of the seismic isolation level of seismically
isolated buildings is generally 0.4 or higher, and by adjusting the dampers, the shear force
response reduction effect of the seismic isolation level is a maximum of about 60%.
αS /α0 αh /α0
0.4 0.3 0.2 0.1 0.0 0.3 0.2 0.1 0.0
1.0 1.0
α1 /α0
0.8 0.8 α1 /α0
α1/α0,αf/α0,αh/α0
α1/α0,αf/α0,αh/α0
α1 /α0, αf /α0, αs /α0

α1 /α0, αf /α0, αh /α0
0.6 0.6
0.4 0.4
αh /α0 αS /α0
0.2 0.2
αf /α0,δm /δ0 αf /α0,δm /δ0
0.0 0.0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
ααhh/α00 ααss/α0
0.2 0.2
0.4 0.4
hh
hh
0.6 0.6
0.8 0.8
0.1
1.0 1.0
αS /α0 0.4 0.2 0.0 αh /α0 0.3 0.2
0.3 0.1
(a)The relation between α1/α0 and αh/α0, h (b) The relation between α1/α0 and αs/α0, h
Fig. 4.3.4 Relation between response prediction equation and damping factor h
（In the case of combination of hysteretic dampers and fluid type dampers）
When hysteretic dampers are added to a seismic isolation level consisting of isolators and
fluid type dampers, the response prediction equation changes as shown in Fig. 4.3.4(a).
When the yield shear force coefficient of the hysteretic dampers is 0.0 ≤ as/a0 ≤ 0.3, the
maximum shear force coefficient of the seismic isolation level a1/a0 has the minimum
value. However, when 0.4 ≤ as/a0, the maximum shear force coefficient a1/a0
monotonically increases, so the effect of combined use of fluid type dampers and hysteretic
dampers is small.
On the other hand, when fluid type dampers are added to a seismic isolation level
consisting of isolators and hysteretic dampers, the response prediction equation varies as
shown in Fig. 4.3.4(b). When the shear force coefficient of the fluid type dampers is 0.0 ≤
ah/a0 ≤ 0.2, the maximum shear force coefficient of the seismic isolation level a1/a0 has
the minimum value. In addition, by increasing the shear force coefficient of the fluid type
dampers ah/a0 within this range, the minimum value is reduced, so a high effect of
combined use can be obtained.
@Seismicisolation
@Seismicisolation
The energy balance method is a response prediction method for approximately determining
the seismic response of seismically isolated buildings. This method is effective for
determining the relation between the ground motion level, the seismic isolation period, and
the response reduction effect of dampers. It is considered that a rational design is to set the
parameters that are effective for controlling the maximum shear force and the maximum
displacement of the seismic isolation level by using the energy balance method, and based
on this, to carry out a detailed time history response analysis.
References
1) H. Akiyama：Earthquake-Resistant Limit-state Design for Buildings, University of Tokyo Press, 1985
2) Sayaka Higashino, Haruyuki Kitamura：Energy balance Based Seismic Response Prediction Methods for Seismic
Isolated Buildings with Rubber Bearings and Viscous Dampers, Journal of Structural and Construction Engineering,
AIJ, No.588, pp.79-86, 2005.2
@Seismicisolation
@Seismicisolation
4.3.2 Response Spectrum Method
(1) Simple prediction method for the maximum response values of equivalent
single-degree-of-freedom (SDOF) system
The simple prediction for the maximum response values of an equivalent SDOF system
involves the evaluation of the dynamic characteristics of the structure by using two different
types of damping, namely, viscous damping and hysteretic damping, followed by the
prediction of the maximum response by using an elastic response spectrum. The theory is
described in detail in references 1), 2) and 3). In this section, the procedure of the response
prediction method is explained.
It is assumed that the seismic isolation level is composed of elasto-plastic elements

combining isolators and hysteretic dampers, linear viscous elements including fluid-type
dampers and other elements. The seismic isolation devices absorb the energy of the input
ground motions, and the superstructure is treated as a rigid body. Then, the SDOF system is
developed (Fig. 4.3.5). Mass M is assumed to be the sum of the masses of the
superstructure and the seismic isolation level. The elastic stiffness of elasto-plastic elements
is represented by K . The elastic period T0 is 2π M K . The viscous element is assumed
to be a linear viscous damper. The procedure for estimating the response is shown below
based on the assumption that the damping coefficient C , the viscous damping factor
h0 = C ⋅ T0 (4Mπ ) and the yield load S Q y of the yield displacement δ y = S Q y K , and
the ratio of secondary stiffness p d are known.
Q Q
h s
1 + pd µ d − pd
M K K eq = K
C ⋅ ω ⋅ δm = h Qm Qy µd
pd ⋅ K s
粘性要素
δ δ
δm δy δm
C K , pd , K eq
粘性要素弾塑性要素
Viscous element Elasto-plastic element

Fig. 4.3.5 Restoring force of seismic isolation device
The response prediction method uses the elastic response spectrum at damping factor h0 for
predicting the maximum elasto-plastic response. Two variations are presented here1). In
response prediction method 1, the maximum response is obtained through a convergence
calculation considering the changes in the equivalent period and in the equivalent damping
factor due to yielding of the SDOF system in the spectrum response. For response prediction
method 2, method 1 is simplified on the assumption that the response spectrum curve is
@Seismicisolation
@Seismicisolation
constant regardless of the period. In method 2, the maximum response at varying values of
T0 , h0 , S Q y , δ y and p d is expressed by a continuous function, and the maximum
response is estimated by a graph. No convergence calculation is therefore necessary. The
procedures of method 1 and 2 are described below in ascending order.
[Procedure of response prediction method 1]
(i) First, the seismic isolation level is assumed to be elastic, and the maximum displacement
δ m is obtained by using the seismic response spectrum at h0 (Fig. 4.3.6). Then, the
ductility factor µ d at that point is obtained.
d m = (T0 2π ) ⋅ S PV (T0 , h0 )
(4.3.13a, b)
md = d m d y
Viscous damping factor h=0.10

粘性減衰定数0.1
1.2
S PV (m s )
0.8 (
S PV Teq , h0 )
(ii)
②
0.4
⑤
(v)
(
S PV Teq , h0 )
S PV (T0 , h0 )
S PV (T
eq ,h )
eq T (s )
0
T0 Teq
0 1 2
Fig. 4.3.6 Pseudo-velocity spectrum
(ii) Equivalent period Teq is obtained.
µd
Teq = T0 (µ d > 1) (4.3.14)
1 − pd + pd ⋅ µ d
(iii) Pseudo-velocity spectrum S PV at damping factor h0 is averaged in the T0 , Teq [ ]

range in equation (4.3.15) to obtain the equivalent velocity spectrum by considering the
changes in the period (Fig. 4.3.6).
1
S pv (Teq , h0 ) =
Teq
∫ S PV (T , h0 )dT (4.3.15)
Teq − T0 T0
@Seismicisolation
@Seismicisolation
(iv) Equivalent damping factor heq is obtained.

 pd −1

h0  λµ d + p d − 1  pd µ d + λ  pd pd
 2  λ 
heq = + λn    (4.3.16a)
  + pµ p λn pd 
µd

pd 
 p d + 1   d d  µd 
 
λ = 1 + p d (µ d − 1) (4.3.16b)
The first term with [ ] and second term of equation (4.3.16a) indicate the contributions of
viscous damping and hysteretic damping, respectively. The hysteretic damping was averaged
based on the assumption that amplitudes occur at the same frequency because the ductility
factor in each cycle during an earthquake ranges from 0 to the maximum value µ d . The
contribution of the damping constant due to the viscous damping factor is also averaged for
each amplitude as the damping factor varies with the period. For theoretical details, refer to
references 1) and 2).
(v) The coefficient of the damping effect2) Dh is obtained by using heq . The response value
S PV (Teq , heq ) at heq is obtained by using Dh and the equivalent velocity spectrum at the
initial damping constant S PV (Teq , h0 ). In an actual observed seismic wave, α = 25, and in
artificial seismic waves, α = 97 and α = 75.
1 + α ⋅ h0
Dh = (4.3.17)
1 + α ⋅ heq
S PV (Teq , heq ) = Dh ⋅ S PV (Teq , h0 ) (4.3.18)
(vi) Based on the above, the maximum displacement δ m is represented by (4.3.19).
T 
δ m = 
eq
 ⋅ S PV (Teq , heq ) (4.3.19)
 2π 
(vii) Steps (i) through (vi) are repeated. When δ m is stabilized, the following calculations
are conducted.
Velocity correction coefficient β is obtained in equation (4.3.20) and the maximum shear
force Qm in equations (4.3.21) and (4.3.20). The maximum acceleration δ + g m is ( )
obtained in equation (4.3.23) and the maximum velocity in equation (4.3.24). Equations
@Seismicisolation
@Seismicisolation
(4.3.21) and (4.3.22) provide simple methods. For details, refer to the attachment to reference
1).
β = (0.6 ⋅ h0 + 0.15)Teq + 0.8 (4.3.20)
µ A =  1 + (Cω K )2 + p d 2 + (Cω K )2  (1 + p d ) (4.3.21)

 
Qm = Fy ⋅ m d 1 + (βCω K ) (1 ≦ m d ≦ m A )
2
(4.3.22a)
Qµ = F y  µ d

p d + (βCω K ) + 1 − p d 
2 2

(µ A ≦µ ) (4.3.22b)
(δ + g)
m = Qm M (4.3.23)
δm = β ⋅ S PV (Teq , heq ) (4.3.24)
[Procedure of response estimation method 2]
In the case that pseudo-acceleration spectrum S pa , pseudo-velocity spectrum S pv and

displacement response spectrum S d , which are often set for short-, medium- and
long-period structures, are constant regardless of the period, equation (4.3.15) can be easily
averaged1), 2). The curves of Fig. 4.3.7 can therefore be developed1) and used for simplified
estimation. The procedure is described below.
(i) The seismic isolation level is assumed to be elastic, and δ el and Qel are obtained from
the spectrum at T0 and h0 .
(ii) One of the combinations of constant S pa , S pv or S d and the ratio of secondary

stiffness p d = 0, 0.2 or 0.5 are selected. A diagonal line is drawn from the breakpoint
(δ y δ el , Q y Qel ) and the point of intersection with the curve at h0 =0.02, 0.1, 0.2, or 0.3 is
used to obtain the maximum displacement δ m , ductility factor m d = d m d y and shear at
the time of maximum displacement Q ′ .
(iii) Velocity correction coefficient β is obtained in equation (4.3.20) and maximum shear
force Qm in equations (4.3.21) and (4.3.22). With the response curves in Fig. 4.3.7, it is
possible to grasp visually and comprehensively, without a convergence calculation, the
tendency of the response reduction in structures with viscous and hysteretic damping due to
the adjustment of elastic viscous damping factor h0 and ductility factor µ d or yield force
Fy . The curves are thus very useful.
@Seismicisolation
@Seismicisolation
pd=0 pd=0.2 pd=0.5
1 Q’ / Qel Q’ / Qel Q’ / Qel
0.8
Spa is
0.6
constant
0.4
0.2
δm /δel δm /δel δm /δel
0
1 Q’ / Qel Q’ / Qel h0=0.3 Q’ / Qel
0.2
0.8
0.1
0.6 0.02
Spv is
constant
0.4
0.2
pd
δm /δel δm /δel
0
δm /δel
1 Q’ / Qel Q’ / Qel Q’ / Qel
0.8
0.6
Sd is
constant
0.4
0.2
δm /δel δm /δel δm /δel
0
0 0.5 1 1.5 0 0.5 1 1.5 0 0.5 1 1.5
Fig. 4.3.7 Pseudo-velocity spectra and response reduction ratios in cases of various yield
forces and viscous damping
( δ el and Qel are response values at elastic damping factor h0 )
(2) Response prediction using response performance design diagrams
To predict the response of the system combining hysteretic damping and viscous damping
(Fig. 4.3.8), Ishimaru4) proposed a theory based on the concept of the “separation of
variables” of energy spectra.
As shown in Fig. 4.3.9, this theory establishes the relation shown in equation (4.3.25) by
shifting the period of the energy spectrum SBi obtained from the system shown in Fig. 4.3.7
to an effective period component T'. The shifting is a result of lengthening the period due to
plasticity of the elasto-plastic element for the input ground motions, and the ratio is
constant. Therefore, the function f(h0, µv, pv, µd, pd), which does not include the period
function, can be given by an empirical equation based on the elasto-plastic response
@Seismicisolation
@Seismicisolation
analysis using 50 seismic waves. Parameters h0, µv, pv, µd, pd represent the performance of
the system. If these parameters are known in advance, it is possible to estimate the response
values by multiplying the value obtained from the empirical equation f(h0, µv, pv, µd, pd) by
the value at the effective period T' of the system from the pseudo-velocity response
spectrum of the seismic motion. Theoretically, if three of the parameters h0, µv, pv, µd, pd are
constant values, the diagrams in Fig. 4.3.12 can be produced, as described later. Therefore,
in this prediction method, the response values can be easily estimated from the diagrams.
Q Q
h s
1 + pd µ d − pd
C K K eq = K
M h Qm Qy µd
h Qy
pd ⋅ K s
粘性要素
δ δ
δm δy δm
C K , pd , K eq
粘性要素
弾塑性要素
Viscous element Elasto-plastic element
Fig. 4.3.8 System using both viscous damping and hysteretic damping
1000 1000
Period shift owing to the elongation of

S Bi (T0 ) period due to plasticization S Bi (T ′)
Pseudo Velocity(cm/sec)
Pseudo Velocity(cm/sec)
100 100
S Bi (T ′)
S Bi (T0 )
10 pSV , 40 10
pSV , 40 pSV , 40
pSV , 40
1 1
0.1 1 10 0.1 1 10
Period (sec) Period (sec)
(a) (b)
Fig. 4.3.9 (a) Comparison between pSV , 40 and S Bi (T0 ) . Both spectra are plotted along the
initial natural period T0 . (b) Comparisons between pSV , 40 and S Bi (T ′) plotted along the
effective natural period T ′
S Bi (T ′)
f (h0 , µV , pv , µ d , pd ) = ≒ constant （4.3.25）
pSv(TE )
The theoretical concept described above is based on reference 4), and the procedure for
estimating the response prediction is explained here. The applicable model treats the
superstructure as a rigid body, and as shown in Fig. 4.3.8, the model of the seismic isolation
level consists of isolators and hysteretic type elasto-plastic elements and viscous elements
such as viscous type dampers. For the viscous elements, nonlinear viscous dampers are
assumed to have bilinear properties with respect to velocity (relief velocity, see Fig.2.2.31)
and the following parameters: the initial viscous damping factor, the relief velocity ratio
with respect to the maximum velocity (maximum velocity/relief velocity), and the
secondary damping coefficient ratio, denoted as h0, µv, and pv, respectively. For the
@Seismicisolation
@Seismicisolation
elasto-plastic element, the parameters are the elastic stiffness, the secondary stiffness ratio,
and the yield displacement, denoted as K, pd, and δy, respectively.
The following is an explanation of the procedure of the prediction method for the case that
the elastic period T0, the viscous damping factor h0, the relief load hQy, the relief velocity
δ , the secondary damping coefficient ratio pv, the yield load sQy, the yield displacement δy,
y
and the secondary stiffness ratio pd of the single-degree-of-freedom system shown in Fig.
4.3.8 are known in advance.
[Response prediction procedure]
(i) The ductility factor µd and the relief velocity ratio µv are assumed from the maximum
displacement δˆm and the maximum velocity δ . Also, the effective period T' is m
obtained from equation (4.3.26). Figure 4.3.10, which plots equation (4.3.26) in
graphical form, shows the relation between the ductility factor µd, the bilinear coefficient
pd, and the ratio T'/T0 (effective period/elastic period).
100
0.02
0.05
Ductility factor
0.075
10 0.10
0.15
0.20
0.30
0.40
0.50
1
1 1.5 2 2.5 3 3.5 4 4.5 5 5.5
T′
The ratio of effective period / elastic period
T0
Fig. 4.3.10 Relation between ductility factor and ratio of effective period/elastic period
 µd 2 
T ′ = T0 1 + −  µd ≧ 3
 1 + p d (µ d − 1) 1 + pd 
（4.3.26）
  µd  
T ′ = T0 0.155  + 0.845 µd < 3
  1 + pd (µ d − 1)  
(ii) As shown in Fig. 4.3.11, from the pseudo-velocity response spectrum for the input
ground motion pSv40 (=ω0･SD40) at a viscous damping factor of 40% and from the
velocity response spectrum SV40, the values of pSV40, SV40 can be read for the effective
period T' obtained in (i). A40 and D40 can be calculated by equation (4.3.27) and equation
(4.3.28) respectively. Here, the pseudo-velocity response spectrum pSv40 (=ω0･SD40) at a
viscous damping factor of 40% is used because it matches very well the "peak parameter
spectrum" of the maximum acceleration, the maximum velocity, and the maximum
displacement for the input ground motion. So, the response can be estimated in the form
of a response factor with respect to the magnitude of the input ground motions.
@Seismicisolation
@Seismicisolation
A40 = ω ′ ⋅ pSV 40 (4.3.27)

D40 = pSV 40 ω′
2p (4.3.28)
ω′ =
T′
10
SV ,40
Pseudo Velocity(m/s)
1
pSV ,40
0.1
0.01
0.1 1 T′ 10
Period(s)
Fig. 4.3.11 Example of reading the values of pSV40, SV40 from the response spectrum
(iii) Parameters h0, pv, and pd are known from the response performance diagram of Fig.
4.3.12(a). Therefore, the response factor of the maximum response acceleration
(absolute acceleration) (δ + g ) A on the vertical axis and the response factor of the
m 40
maximum response displacement δm/D40 on the horizontal axis can be read from the
point of intersection of the assumed µv, µd. By multiplying A40, D40 obtained in
procedure (ii), a new response acceleration (δ + g) and response displacement δm can be m
obtained. Also, from the diagram in Fig. 4.3.12(b), the maximum response velocity δ m
can be obtained. From δm, δ and from the elastic limit deformation δy and the relief
m
velocity δ , a new ductility factor µd and relief velocity ratio µv are obtained.
y
μd=45
1.6
μd=35 μd=40
μd=30
μd=25
1.5 μd=20
μd=15
The response factor μd=10
μd=12.5
of maximum absolute 1.4 μd=7.5
acceleration

1.3
( (δ + g) m ／A40 )
1.2
pd=0.45
pd=0.1
1.1 pd=0.15 pd=0.4
pd=0.2 pd=0.35
1 pd=0.25 pd=0.3
0.7 0.8 0.9 1 1.1 1.2 1.3 1.4
The response factor of maximum displacement δ m D40
(a) Diagram of the response factor of maximum absolute acceleration and maximum displacement
@Seismicisolation
@Seismicisolation
1.4
1.3
1.2
pd=0.3
The response factor 1.1 pd=0.2 pd=0.3
of maximum velocity
pd=0.15
1 pd=0.2
( δm SV , 40 ) pd=0.15
pd=0.1 pd=0.1
0.9 μd=5.0
μd=50.0 μd=7.5
μd=10.0
0.8
μd=15.0
μd=25.0
0.7
0.7 0.8 0.9 1 1.1 1.2 1.3 1.4
The response factor of maximum displacement δ m D40
(b) Diagram of the response factor of maximum velocity and maximum displacement
Fig. 4.3.12 Response performance design diagrams
(4) The calculations in procedures (i) through (iii) are repeated until the assumed µ̂ d and
µ̂V are virtually the same as µd and µv obtained in (iii) and the effective period T' from
equation (4.3.26) by substituting µd → µ̂ d , µv → µ̂V until convergence. As a result, it is
possible to obtain the maximum response acceleration (δ + g) , the maximum velocity m
δ , and the maximum displacement δm.

m
References
1） Kazuhiko Kasai, Yoshinobu Kawanabe：Equivalent Linearization to Predict Dynamic Properties and Seismic Peak
Responses of a Structural System with High Viscous Damping and Hysteretic Damping, Journal of Structural and
2） Kazuhiko Kasai, Hiroshi Ito, Atsushi Watanabe：Peak Response Prediction Rule for a SDOF Elasto-Plastic System
Based on Equivalent Linearization Technique, Journal of Structural and Construction Engineering, AIJ, No.571,
pp.53-62, Sep, 2003
3） Kazuhiko Kasai, Akira Suzuki, Kazuyuki Oohara：Equivalent Linearization of a Passive Control System Having
Viscous Dampers Dependent on Fractional Power of Velocity, Journal of Structural and Construction Engineering, AIJ,
No.583, pp.47-54, Sep, 2004
4） Ippei Hata, Shinji Ishimaru, Jun Hasegawa：On Performance Based Design Method for Systems with Nonlinear
Viscous Damper and Elasto-Plastic Damper, Journal of Structural and Construction Engineering, AIJ, No.617,
pp.47-54, Jul, 2007
@Seismicisolation
@Seismicisolation
4.3.3 Response Estimation Taking the Superstructure into Consideration
The response prediction in the response spectrum method described above basically assumes
a single-degree-of-freedom system, so it can be applied when the superstructure mass ms and
the mass of the seismic isolation level mb can be taken to be a single mass, as shown on the
right of Fig. 4.3.13, and the superstructure can be replaced with a rigid body.
Superstructure
ms
cs ks ms + mb Rigid body
mb
Seismic isolation level cb kb
cb kb
Two degree of freedom Single degree of freedom

(Rigid body model)
Fig. 4.3.13 Rigid body model
In Fig. 4.3.13, the notations ms, cs, ks are the superstructure mass, damping coefficient, and
stiffness, respectively, and mb, cb, kb are the mass, damping coefficient, and stiffness of
seismic isolation level, respectively. Also, the natural period of the superstructure is
T = 2π m k , the period of the seismic isolation is T = 2π (m + m ) k , the mass ratio of the
S S S b b s b
superstructure as a proportion of the total mass is mˆ = ms (mb + ms ) . The damping coefficient of

the seismic isolation level is c = 2h (m + m )⋅ k , and the damping coefficient of the
b b b s b
superstructure is c = 2h m ⋅ k .
s s s s
Here, the vibration equations without damping are as follows.
Mx + Kx = − Mig (4.3.29)
m  k −k s  1
M = s  K = s  i= 
 mb  − k s kb + k s  1 (4.3.30)
From the eigenvalue problem,
 ks − ks  m
2 s  r 
  −ω  r=0 r =  2
− k s kb + k s   mb   r1  (4.3.31)
r2 ks
=
r1 k s − ω 2 ms
 ( 2
 k s − ω m s r2 − k s r1 = 0 ) ∴
r2 k s + k b − ω 2 mb

( 2
− k s r2 + k s + k b − ω mb = 0
 ) r1
=
ks
(4.3.32)
The participation vectors in the primary mode are assumed to be β1r2 and β1r1. The ratio of
@Seismicisolation
@Seismicisolation
the participation vectors is given in (4.3.33), and the relation between m̂ and Tb/Ts is given
in (4.3.34).
b1r2 r2 ks k + k − ω 2 mb
= = = s b (4.3.33)
b1r1 r1 k s − ω ms
2
ks
Tb b1r2 b1r2 mˆ
= − (4.3.34)
Ts b1r2 − b1r1 b1r2 − b1r1 )mˆ + b1r1
(
The range, over which a rigid body can be substituted for the superstructure and the seismic
isolation level, is defined here by the range of the modal participation vectors 1.0 ≤ β1r2 ≤ 1.1
and 0.9 ≤ β1r1 < 1.0. The relation between m̂ and Tb/Ts that satisfies these conditions is
shown in Fig. 4.3.14, and it can be judged whether a rigid body substitution can be made by
using the solid line drawn as the boundary. Also, from this figure, from the relation Tb/Ts
between the superstructure natural period and the seismic isolation period, it can be seen that
it is desirable to design so that the natural period is a factor of about 2.5 or higher, although
this depends on the mass of the seismic isolation level.
4 The range in which model

can be treated as a rigid body
3
Tb/Ts
2
The range in which model
1 cannot be treated as a rigid
body
0
0.5 0.6 0.7 0.8 0.9 1
Fig. 4.3.14 Relation of mass ratio and period ratio that can be applied as a rigid body
Therefore, if an equivalent stiffness of the seismic isolation level is assumed, the response
displacement of the superstructure can be estimated by multiplying the participation vector
obtained from an eigenvalue analysis, which includes the seismic isolation level, by the
maximum deformation δm of a single mass system obtained by a simple response prediction
method.
Outside the range where rigid body substitution can be applied, the effect of higher modes is
significant, so it is necessary to estimate the response by a modal analysis using a complex
eigenvalue analysis. Also, buildings that can be converted into a single mass system include
high-rise buildings. As an example, if the primary natural period of the building under
consideration is 4.0 (s), the equivalent natural period of the seismic isolation level is 10 (s);
@Seismicisolation
@Seismicisolation
however, these long period buildings are substantially removed from this response prediction
method of seismically isolated structures, and factors other than seismic response become
more important. Therefore, it is considered that rigid body substitution can be made if the
target building is about 2.0 (s) or less.
Next, a method of estimating the maximum response displacement, the maximum response
velocity, and the maximum response acceleration (maximum absolute acceleration) is
described for a model in which the superstructure can be treated as a rigid body according to
the response values obtained from the simple response prediction method. In the case of the
range where the superstructure can be treated as a rigid body, the range of the participation
factor of the seismic isolation level is 0.9 ≤ β1r1 < 1.0 and the range of the participation factor
of the superstructure is 1.0 ≤ β1r2 ≤ 1.1, and so the modal participation function of the
primary mode is (β1r1, β1r2) = (0.9, 1.1). Also, the input vector (seismic isolation level,
superstructure) = (1.0, 1.0), so the participation factor of the secondary mode is (β2r1, β2r2) =
(0.1, -0.1).
If the maximum displacement and the maximum velocity response when the seismic
isolation level and the superstructure can be converted into a single-degree-of-freedom
system are δm and δ respectively, then the maximum response displacement δm, i and the
m
maximum response velocity δ can be virtually expressed by equation (4.3.35), because the
m, i
first mode is dominant.
δm
δ m,i = β1ri ⋅
β1r1
δm
δm,i = β1ri ⋅
β1r1
(4.3.35)
Next, the response accelerations of the ith story in the first mode and the second mode are
approximated by (4.3.36). The first equation is the first mode (δ + g ) , and the second m ,1st ,i
equation is the second mode (d + g ) . The second mode is obtained from the response
m , 2 nd ,i
spectrum of the first mode period of the superstructure.
(d + g ) m,1st ,i = 2h1ω1 β1ri ⋅ dm + ω12 β1ri ⋅ d m

( )
(d + g ) m, 2 nd ,i = 2h2ω 2 β 2 ri ⋅ pSV , 40 T2 ⋅ Fh + ω 22 β 2 ri ⋅ D40 (T2 ) ⋅ Fh
(4.3.36)
1
Fh =
2h2
(4.3.37)
Also, ω , ω , h , h are approximate values of the eigenvalues and viscous damping

1 2 1 2
factors for each mode and are obtained from the following equations.
@Seismicisolation
@Seismicisolation
β1r2
γ =
β1r1
γ 2π
T1 = Ts ,1 → ω1 =

γ −1 T1
γ 2π
T2 = Ts , 2 → ω 2 =

γ −1 T2 (4.3.38)
h1 = hb h2 = (hs + hb m
ˆ ωb ω s ) ˆ
1− m
(4.3.39)
With these, the response value of the absolute acceleration of the ith story can be calculated
by using the square root of the sum of squares (SRSS) method.
(d + g ) m ,i = ((d + g ) ) + ((d + g )

m ,1st ,i
2
m , 2 nd ,i )
2
(4.3.40)
References
1) James Kelly: Earthquake-Resistant Design with Rubber, 2nd Edition, Springer, 1997
@Seismicisolation
@Seismicisolation
4.3.4 Method of Calculating the Response Spectrum of Ground Motions Including the
Soil Properties
The seismic response prediction of structures by using the response spectrum method is
convenient, and stable response values can be obtained. Therefore, if the input ground
motions are given as the engineering bedrock response spectrum, and if it is possible to
simply obtain the nonlinear amplification properties of the soils up to the ground surface, it is
possible to simply predict the seismic response of the ground and the building.
Equivalent-linear earthquake site response analysis (SHAKE)1) is commonly used as a

method of analyzing the ground amplification properties of the surface strata due to its
analytical and theoretical convenience. However, it requires the input ground motions in the
form of a time history waveform. So, if the input ground motions are set in the form of a
response spectrum only, it is necessary to analyze the nonlinear amplification properties of
the near-surface soils by using SHAKE after converting the response spectrum into the
corresponding time history waveform. Various analytical methods have been proposed as
methods for simply evaluating the nonlinear amplification properties of the surface strata
without carrying out a time history response analysis. However, the method of Miura et al.2),
which is based on the notification wave, is the only specific evaluation method that takes into
consideration both 1) the effect on wave motions being transmitted through the surface strata
with changes in rigidity and damping due to the progression of nonlinearity in the nonlinear
amplification properties of the surface strata, and 2) the effect of the impedance ratio between
the engineering bedrock and the near-surface soils. In this method, the surface strata and
engineering bedrock are replaced with a two-layer ground. By carrying out a convergence
calculation for the strain with an equivalent linear model, the acceleration amplification ratio
of the surface strata is calculated. This method is clear and convenient, and considers the
interaction between the building and the soil; however, because the convergence calculation
is necessary, it is difficult to carry out manually. Sakai et al.3) have proposed a convenient
evaluation method that does not require the convergence calculation.
Refer to references 2) and 3) for details of the methods of obtaining the soil amplification
coefficient.
References
1) Schnabel, P. B. , Lysmer, J. and Seed, H. B. (1972), “SHAKE – A Computer Program for Earthquake Response
Analysis of Horizontally Layered Sites”, Report No. EERC72-12, EERC
2) Kenji Miura, Kohji Koyamada, Masanori Iiba：Response Spectrum Method for Evaluating Nonlinear Amplification of
Surface Strata, Journal of Structural and Construction Engineering, AIJ, No.539, pp.57-62, Jan.,2001
3) Yuki Sakai, Seiji Tsuno, Kazuyoshi Kudo, Toshimi Kabeyasawa：Simplified Method to Evaluate Ground Surface
Amplification Assuming the Input Motions on the Engineering Bedrock in the Revised Enforcement Order of the
Building Standard Law, Journal of Structural and Construction Engineering, AIJ, No.565, pp.73-78, Mar.,2003
@Seismicisolation
@Seismicisolation
4.3.5 Design Seismic Load for the Superstructure
Normally, the design seismic load for the structure above the seismic isolation level is set on
the basis of a preliminary analysis. However, various equations for calculating the design
seismic load have been proposed instead of the preliminary analysis.
The ideal behavior of a seismically isolated building during an earthquake is a rigid-body

translation response, the superstructure of which is considered to be rigid relative to the
seismic isolation level. In this case, the acceleration response of each story is uniform, so the
seismic load of each floor is proportional to the mass distribution (uniform distribution of the
seismic coefficient, UBC-911)).
wx
Fx = Qiso ⋅ (4.3.41)
∑ wi
where Fx, wx: design seismic load and mass at level x, Σwi: total mass of the superstructure,
Qiso: shear force of seismic isolation level.
If an increase in the response of the superstructure is likely due to the rigidity of the
superstructure and seismic isolation devices, design equations that describe the amplification
of the seismic load with height as an inverted triangular distribution can be used (inverted
triangular distribution of the seismic coefficient). These are mainly adopted in the design
standards (ASCE/SEI 7-102)).
w x ⋅ hx
Fx = Qiso ⋅ (4.3.42)
∑ ( wi ⋅ hx )
where hx: height of level x from the seismic isolation level.
In addition, a modified distribution that takes into consideration the higher mode response
has been proposed based on equation (4.3.42)3)-5).
In Japan, a design shear force coefficient distribution has been proposed based on the
knowledge6) that when the stiffness of the elastic isolators is sufficiently small, the response
of the isolators is not amplified in the superstructure, and amplification of the response of the
superstructure is totally governed by the dampers. In this design recommendation, the
following equation, which takes into consideration the higher mode response and the
nonlinear restoring force characteristics (bilinear model) of the seismic isolation level, has
been proposed6).
@Seismicisolation
@Seismicisolation
a i = a f + ai ⋅ a i ⋅ a s (4.3.43)
where ai: shear force coefficient of the ith floor, af, as: shear force coefficient of elastic
isolators and hysteretic dampers, α i : ideal yield shear force coefficient distribution or Ai
distribution, ai: amplification factor in the height direction of α i of the shear force
coefficient due to the dampers.
The ai values are represented by equation (4.3.44) as a linear distribution in the height
direction. Also, as the value of ai for the uppermost floor (N floor), a is given by equation
(4.3.45) in terms of the ratio of the stiffness of the superstructure relative to the hysteretic
dampers, bs (where bs = k0/ks, k0: shear stiffness of the first floor of the superstructure, ks:
horizontal stiffness of the dampers). In this way, the amplification of the shear force
coefficient due to the higher mode response is represented.
 a −1  N − a
ai =  i + (4.3.44)
 N −1  N −1
3.1238 − 0.1238bs 1 < bs < 10


a = 2.0127 − 0.0127bs 10 ≦ bs < 80 (4.3.45)
 1.0 80 ≦ bs

In equation (4.3.43), α i , which is the Ai distribution, is expressed by using the shear force in
the following equation.
Qiso ai ⋅ Ai ⋅ s Q y + f Qm
ai = ⋅ (4.3.46)
Mg s Q y + f Qm
where fQm, sQy: shear force of elastic isolators and hysteretic dampers, M: total mass of the
superstructure, g: acceleration due to gravity.
Also, the notification regarding seismically isolated structures published in 2000 (Ministry of
Construction Notification No. 2009) defined technical standards for the structural methods of
seismically isolated structures, among which the following equation was used for the design
shear force coefficient.
Qiso Ai ( s Q y + h Qm )+ f Qm
αi = ⋅ (4.3.47)
Mg s Q y + h Qm + f Qm
where hQm: shear force of fluid type dampers.
Equation (4.3.47) separates the shear force of elastic isolators and dampers (including fluid
type dampers) in accordance with equation (4.3.46) and takes into consideration that dampers
@Seismicisolation
@Seismicisolation
only amplify the shear force coefficient; the amplification is expressed by the Ai distribution
only. Therefore, it has been pointed out that the seismic response could be under-evaluated7).
Kikuchi et al.8) proposed the following equation for the Ai' distribution as an alternative to the
Ai distribution in equation (4.3.47). The first, second, and third terms in equation (4.3.48)
represent the uniform distribution of the seismic coefficient, the inverted triangular
distribution of the seismic coefficient, and the α distribution, respectively, and the fourth
term takes into consideration the effect of the amplification properties of the near-surface
soils in the design seismic response spectrum.
 1 
Ai′ = 1 + k1 (1 − α ) + k 2  − α  + k 3 (1 − α ) 2 (4.3.48)
 α 
where a: standardized weight, k1, k2: parameters relating to the natural period of the
superstructure, k3: parameter relating to the ground conditions.
Nishimura et al.9) and Kobayashi et al.10) proposed a shear force coefficient distribution that
incorporates Skinner's11) isolation factor I and nonlinear factor NL. Isolation factor I is an
index that represents the degree of isolation of the seismic isolation level in the
superstructure, and nonlinear factor NL is an index that represents the degree of nonlinearity
of the restoring force characteristics of the seismic isolation level. Isolation factor I is defined
by the following equation for the restoring force characteristics of the seismic isolation level
shown in Fig. 4.3.15 (bilinear model). Furthermore, from the geometric relations, the
nonlinear factor NL can be expressed with the value of the yield strength of the seismic
isolation level and the maximum response, and can also be expressed by the equivalent
viscous damping factor heq.
Qiso
δm δm
Fig. 4.3.15 Restoring force characteristics of the seismic isolation level
Tb1
I= (4.3.49)
T1
@Seismicisolation
@Seismicisolation
P1 Ah π
NL = = Q y / Qiso − δ y / δ m = = heq (4.3.50)
P2 4Qiso ⋅ δ m 2
where Tb1: natural period when the superstructure is considered to be a rigid body relative to
the primary stiffness Kb1 of the seismic isolation level, T1: primary natural period obtained
with base fixity of the superstructure, Ah: area of the bilinear hysteresis loop, heq: equivalent
viscous damping factor.
Nishimura et al.9) proposed two methods for a model in which the seismic isolation level has
bilinear restoring force characteristics: one for the shear force coefficient response
amplification ratio a of the uppermost floor of the superstructure relative to the seismic
isolation level, and the other for the rate of increase b of the shear force coefficient of the
uppermost floor relative to the notification equation (4.3.47). In each case, the values for the
intermediate floors are obtained by linear interpolation that takes the seismic isolation level
to be 1. The response amplification ratio a and the rate of increase b are given by the
following equations, based on the regression analysis for the shear force coefficient from the
time history response analysis.
2.19 + 3.95 NL　　 0.0 < I ≦ 0.5


 2.31 + 3.34 NL　　 0.5 < I ≦ 1.5
a= (4.3.51)
 1 . 66 + 2 . 58 NL 1
　　 .5 < I ≦ 3.0
1.04 + 1.59 NL　　 3.0 < I ≦ 5.0
1.61 + 0.31Teq / T0 　　 0.0 < I ≦ 0.5


 1 . 60 + 0 . 19T / T 0.5 < I ≦ 1.5
　　
b=
eq 0
(4.3.52)
1.33 + 0.08Teq / T0 　　 1.5 < I ≦ 3.0
0.94 + 0.02Teq / T0 　　 3.0 < I ≦ 5.0
where Teq: period based on the maximum deformation of the seismic isolation level.
Kobayashi et al.10) proposed a new amplification factor βi instead of ai in equation (4.3.44).

They used equation (4.3.46) of this design recommendation as the basis. In this equation, heq
is a value expressed as a percentage.
 β −1  N − β
β i =  i +
 (4.3.53)
 N −1 N −1
s
β= +t (in the range where β >u, β =u) (4.3.54)
I2
s = 0.26heq + 0.29 (in the range where s >5.0, s = 5.0)
t = 0.60 (4.3.55)
@Seismicisolation
@Seismicisolation
u = 0.09heq + 1.28 (in the range where u >3.0, u = 3.0)
Equation (4.3.54) is set so that it gives average values of the shear force coefficient
distribution according to the time history response analysis. If the restoring force
characteristics of the seismic isolation level differ from the bilinear model, the bilinear model
can be applied by replacing with an equivalent bilinear model that uses an equivalence rule.
Furthermore, the variation in the shear force coefficient distribution with respect to the input
seismic motions becomes significant in the range where isolation factor I is small. The extent
of this variation is determined by the coefficient of variation CV of β in accordance with
isolation factor I, and can be evaluated by multiplying the correction factor ν in the following
equation by β of equation (4.3.53).
 0.25 (0 ≦ I < 3)

CV = − 0.06 I + 0.43　　
(3 ≦ I < 6) (4.3.56)
 0.07 (6 ≦ I )

 1 + CV 　　(average value + σ )
ν = (4.3.57)
1 + 2CV 　
　　( average value + 2σ )
Furthermore, it has been confirmed that amplification in the vertical direction of the shear
force coefficient due to fluid type dampers can be evaluated by equations (4.3.53) to (4.3.55).
Therefore, if hysteretic dampers are used in combination with fluid type dampers, then
calculating their respective amplification factors sβi and hβi and multiplying by the shear
force for each damper, as indicated by equation (4.3.58), means it is possible to obtain the
design shear force coefficient distribution that takes into consideration the response
amplification of the superstructure10).
Qiso Ai ( s β i ⋅s Q y + h β i ⋅h Qm )+ f Qm
αi = ⋅ (4.3.58)
Mg s Q y + h Qm + f Qm
where sβi: amplification factor due to hysteretic dampers, hβi: amplification factor due to fluid
type dampers.
@Seismicisolation
@Seismicisolation
5 5
4 4
Uniform
一定震度
3 Inverted triangle
逆三角形 3
Story
Story
Design recommendation
指針
2 Notification
告示 2
Kikuchi et al.
菊地ら
1 1
Nishimura et al.
西村ら
Kobayashi et al.
小林ら
0 0
0 0.2 0.4 0.6 0 3000 6000 9000
Shear force coefficient
Shear force (kN)
Fig. 4.3.16 Shear force coefficient Fig. 4.3.17 Shear force

(5-story models) (5-story models)
15 15
12 12
9 9
Story
Story
6 6
3 3
0 0
0.0 0.2 0.4 0.6 0 6000 12000 18000
Shear force coefficient
Shear force (kN)
Fig. 4.3.18 Shear force coefficient Fig. 4.3.19 Shear force

(15-story models) (15-story models)
Figures 4.3.16 to 4.3.19 compare the calculation equations for the design seismic load of
5-story and 15-story seismically isolated building models (design base shear coefficient of
0.165 and 0.130, respectively)12). In the calculations, the 4th term representing the effect of
the near-surface soils in calculation equation (4.3.48) of Kikuchi et al. has been ignored, and
equation (4.3.52) is used in the calculation equation of Nishimura et al. Also, in the
calculation equation of Kobayashi et al.11), the values (average values) for ν = 1 are shown. It
can be seen that the shear force coefficient distribution and the shear force distribution differ
greatly depending on the calculation equations. It is necessary to select the design equation
that is suitable for the structural properties of the building being designed. Also, it is
necessary to set the design seismic load of the superstructure appropriately while carrying out
an interactive comparison with a time history response analysis.
@Seismicisolation
@Seismicisolation
References
1) International Conference of Building Officials, “Earthquake regulations for seismic Isolated structures”, Uniform
Building Code, 1991
2) ASCE Standards ASCE/SEI 7-10, “Minimum Design Loads for Buildings and Other Structures”, 2010
3) Takim Andriono and Athol J. Carr, “REDUCTION AND DISTRIBUTION OF LATERAL SEISMIC INERTIA
FORCES ON BASE ISOLATED MULTISTOREY STRUCTURES”, Bulletin of The New Zealand National Society
for Earthquake Engineering, Vol.24, No.3, pp. 225-237,1991
4) D. G. Lee, J. M. Hong and J. Kim, “Vertical distribution of equivalent static loads for base isolated bui1ding structures”,
Engineering Structures, Vol. 23, pp.1293-1306, 2001
5) Kelby York and Keri L. Ryan, “Distribution of Lateral Forces in Base-Isolated Buildings Considering Isolation System
Nonlinearity”, Journal of Earthquake Engineering, Vol.12, pp.1185-1204, 2008
6) Architectural Institute of Japan：Recommendation for the Design of Seismically Isolated buildings 3rd Ed., pp.80-93，
2001
7) Mineo Takayama, Keiko Morita：A Study on the Regulations for Seismic Design of Base-Isolated Buildings,
8) Masaru Kikuchi, Tatsuhiko Ohsawa, Yoshihisa Kitamura, Takeaki Koshikawa, Masaiki Ueda ： Seismic Load
Distribution of Isolated Buildings with Hysteretically-Damped Isolation Bearing, AIJ Hokkaido Chapter
Architectural Research Meeting, No.78, pp.83-88, 2005.7
9) Takuya Nishimura, Kazuo Tamura, Masaaki Saruta, Kazuhiko Morikawa, Masanori Iiba：Evaluation of Shear Force
Distributions of Isolated Buildings Part.1 Part.2, Summaries of Technical Papers of Annual Meeting, Architectural
Institute of Japan, B-2, pp.231-234, 2010.9
10) Masahito Kobayashi, Gou Tanizaki, Shingo Matsuda：Vertical Distribution of Seismic Design Load for Seismically
Isolated Buildings Corresponding to Diversity of Seismic Isolation Devises, Journal of Structural and Construction
Engineering, AIJ, No.676, pp.859-868, 2012.6
11) R. I. Skinner, W. H. Robinson and G. H. McVerry, “An Introduction to Seismic Isolation”, Wiley, pp.33-48, pp.171-199,
1993
12) Architectural Institute of Japan：Structural Response and Performance for Long Period Seismic Ground Motions,
pp.233-235, 2007
@Seismicisolation
@Seismicisolation
Section 4.4 Response Prediction by Time History Response Analysis
4.4.1 Time History Response Analysis Method
The time history response analysis enables analysis models that are more detailed and
complex to be set in stages according to the shape and properties of the building. The
analysis also enables the actual behavior of the building to be directly determined. Therefore,
it is an effective method for confirming the safety of the design in more detail.
Apart from natural rubber bearings, in most cases the restoring force of the isolators has
nonlinear properties depending on the amount of deformation. The damping force of
hysteretic dampers is nonlinear, which is similar to that of the isolators, and some viscous
dampers have nonlinear properties that depend on the relative velocity. In the time history
response analysis of these types of seismically isolated buildings, the nonlinear properties
must be taken into consideration. Methods of time history response analysis include
numerical integration methods such as the acceleration method, the Runge-Kutta method,
and methods based on rigorous solutions (Nigam-Jennings method). To deal with the
nonlinear property of the seismic isolation devices and the superstructure, the acceleration
method is often used. The acceleration method is a method in which the change in
acceleration within a small time step is assumed, and the response values in each small
interval are obtained. Its advantages are that the time step can be made larger compared with
other numerical integration methods, and strong nonlinearities can be handled. Normally, the
assumption for the change in acceleration is an equation referred to as the Newmark-β
method1), and often the linear acceleration method (β = 1/6) or the average acceleration
method (β = 1/4) is used. As the time history response analysis progresses, the numerical
calculation may become unstable, such as divergence in the integral or unbalanced forces,
due to the relation of the size of the time step to the natural frequency of the vibration system
and the extent of nonlinearity of the structure. It is necessary to appropriately select the time
step and the numerical integration method by taking into consideration the properties of the
seismic isolation devices and the restoring force characteristics, modeling of the building and
its scale, and calculation time.
References
1) N. M. Newmark : A Method of Computation Dynamics, Journal of the Engineering Mechanics Division, Proceedings
of the ASCE, Vol. 85, No. EM3, pp.67-94, July, 1959.
@Seismicisolation
@Seismicisolation
4.4.2 Analysis Models
The analysis model must be capable of representing the properties of the seismically isolated
structure, but it is also desirable that the structure be appropriately modeled in accordance
with the objective of the analysis. For example, can the model take into consideration
relevant phenomena such as torsional vibrations or the response to simultaneous inputs in
two horizontal directions? Or, what kinds of response values are to be obtained as results? Or,
are response values for each individual member to be directly obtained, and not just the
overall structure response such as the story shear force? In addition to the maximum response
values for design, are response results such as the response waveforms and cumulative
energy to be obtained? As a result of the dramatic progress in the calculation processing
power of computers, the restrictions on the scale of analysis models when carrying out time
history response analyses are being eliminated. However, it is necessary to be aware that the
quantity of information that can be obtained from a detailed model is very large, within
which the actual necessary information may be buried. The following is an explanation of the
modeling corresponding to the superstructure, seismic isolation devices, and lower structure.
(1) Modeling the superstructure
For modeling not only seismically isolated structures but any building structure, various
methods can be considered from the simplest single-degree-of-freedom system model to the
most detailed three-dimensional frame model. When selecting the analysis model, the
model should have sufficient accuracy to achieve the objective of the analysis. Detailed
modeling is not necessarily appropriate modeling. In almost all cases of dynamic analysis
modeling for the design of seismically isolated structures, it is sufficient to model the
superstructure as a multi-degree-of-freedom model. The forces in the members of the
superstructure, such as in the columns and beams, can be obtained from a static analysis
using as external forces the seismic forces obtained from a dynamic response analysis with
the multi-degree-of-freedom model, or design seismic forces determined as the envelope of
these response seismic forces. Also, if analyzing multidirectional input of ground motions
or torsional vibrations, a detailed model other than a multi-degree-of-freedom system is
necessary.
The following is an explanation of the various types of modeling of the superstructure in

accordance with the procedures in the main text of this design guideline.
1) Single-degree-of-freedom model (Fig. 4.4.1)
The SDOF model considers the dynamic behavior of the seismic isolation level and the
seismic isolation devices. For elastically designed seismically isolated structures in
which the superstructure has sufficient stiffness, the behavior of the superstructure is
virtually that of a rigid body, so the dynamic behavior of the seismic isolation level and
the seismic isolation devices can be determined by using a single-degree-of-freedom
model.
@Seismicisolation
@Seismicisolation
2) Multi-degree-of-freedom model (Fig. 4.4.2)
In the multi-degree of freedom model (MDOF), the mass at each story is concentrated at
a single point and the story stiffness is represented by a spring connecting each story.
This model is very useful for modeling seismically isolated structures. The following
two methods of modeling can be considered, depending on the treatment of the stiffness
of each story:
- method using bending and shear elements for the stiffness of each floor of the
superstructure
- method using an equivalent shear element for the stiffness of each floor of the
superstructure.
If elasto-plasticity of the superstructure is taken into consideration, often an equivalent

shear model is used, or the elasto-plastic properties of the shear stiffness only of the
bending and shear system are given.
Fig. 4.4.1 Fig. 4.4.2 Fig. 4.4.3 MDOF model for

Single-degree-of- Multi-degree-of-freedom torsional vibration
freedom (SDOF) model (MDOF) model
3) Torsional vibration model (Fig. 4.4.3)
The torsional vibration model assumes each floor to be rigid. The components of motion
u, v, θz of each floor are considered, and input motions are applied in two horizontal
directions. This model can also be represented as an equivalent shear system or a
bending and shear system, as in the previous model. In seismically isolated buildings
with a large in-plane dimension, a difference in the deformation of the seismic isolation
level can be produced between the center of the building and the edges due to a phase
difference in the input ground motions and eccentricity in stiffness or mass of the
seismic isolation level. The torsional vibration model is often used as a method that is
comparatively simple for investigating this type of behavior.
4) Three-dimensional models (Figs. 4.4.4, 4.4.5)
@Seismicisolation
@Seismicisolation
A structure can be modeled as a complete three-dimensional structure. For practical

purposes, a method in which the floor in each story is assumed to be rigid in-plane (the
so-called rigid floor assumption) should take into consideration the components of
motion u, v, θz, and the vertical components of motion at each column position can also
be considered. By using this type of model, it is possible to analyze the response with
consideration of the irregular layout of seismic shear walls and the uplift phenomena at
isolators (Fig. 4.4.4). Also, a three-dimensional model (parallel MDOF model) can be
considered. In this model, the superstructure is modeled with a set of
multi-degree-of-freedom models that take into consideration the in-plane stiffness of
each frame of the superstructure, where each frame is connected by using the rigid floor
assumption (Fig. 4.4.5). In three-dimensional models, the superstructure can be elastic
or plastic. The most rigorous model is considered to be a superstructure modeled as a
three-dimensional frame, with the nonlinearity of the columns and beams tracked over
time.
Fig. 4.4.4 Three-dimensional Fig. 4.4.5 Parallel MDOF model

frame model
5) Vertical motion and uplift models (Figs. 4.4.6, 4.4.7)
Seismic response analysis is sometimes carried out to evaluate the variation in axial
forces acting on isolators and the forces produced in large-span beams or slabs during
the input of vertical motions. Models used for vertical motion response analysis include
models in which the vertical stiffness of each story is evaluated from the column axial
stiffness, and in which the degrees of freedom in the vertical direction are added to the
MDOF model in Fig. 4.4.2. Or, they are models in which intermediate nodes are
provided on the plane frame or three-dimensional frame models, as shown in Fig. 4.4.6,
and can represent the vibration modes of beams or slabs in the vertical direction.
The generation of tensile forces in isolators at the edges of buildings is caused not only
by vertical motions but also by the effect of building overturning moments due to
horizontal motions. Therefore, to investigate the generation of tensile forces on
isolators, it is necessary to input ground motions with a combination of vertical motions
and horizontal motions. In this case, stiffness in the vertical direction is added to the
bending and shear system multi-degree-of-freedom model as in Fig. 4.4.7, and
sway-rocking models that include the position of the isolators in plan can be considered.
Also, input motions of a combination of vertical motions and horizontal motions can be
applied to plane or three-dimensional frame models. For simultaneous input of
horizontal motions and vertical motions, the phase properties of the input ground
motions have a great effect on whether tensile forces are generated. The combination of
@Seismicisolation
@Seismicisolation
vertical motions and horizontal motions is explained in Section 4.4.6, "Vertical Motion
Response."
When the seismic isolation level is fixed and only horizontal motions are considered,
damping that is proportional to about 1% to 3% of the stiffness (in nonlinear models,
damping that is proportional to the instantaneous stiffness) for the primary natural
frequency in the horizontal direction is often applied as the internal viscous damping of
the superstructure. In the case of vertical motions only, similar damping is applied for
the primary natural frequency in the vertical direction. When both horizontal motions
and vertical motions are input simultaneously and when damping proportional to the
stiffness is defined for the primary natural frequency in the horizontal direction, the
primary natural frequency in the vertical direction is higher compared with the
horizontal direction, and the damping will be over-evaluated. Therefore, damping ratios
can be independently defined in the horizontal and vertical directions by using modal
damping, or appropriate damping can be ensured in both directions by using Rayleigh
damping. For the seismic isolation devices, which are described later, internal viscous
damping is normally not applied.
Fig. 4.4.6 Frame model for Fig. 4.4.7 MDOF model for
vertical vibration sway-rocking vibration
(2) Modeling the seismic isolation devices
1) Isolators
Normally, it is assumed that the relation between the horizontal force and horizontal
deformation of natural rubber bearing is linear elastic. However, to represent the
hardening properties for large deformations, multi-line type nonlinear elastic models are
used to vary the stiffness in several stages.
The relation between horizontal force and horizontal deformation for multifunctional
isolators that integrate the bearing function and the damping function is nonlinear, such
as in lead rubber bearings or high-damping rubber bearings, and so it is necessary to
appropriately model the restoring force properties. For modeling multifunctional
isolators, bilinear models or restoring force models can be used to vary the yield force or
the stiffness in accordance with the amplitude of the maximum displacement (modified
bilinear model) (Fig. 4.4.8). The yield stress, the initial stiffness, and the stiffness after
yielding, which define the bilinear model, are determined by design properties that
@Seismicisolation
@Seismicisolation
reflect the test results. However, compared with natural rubber bearings, the dependence
on axial force and the dependence on velocity of these multifunctional isolators are
large, so the error in replacing the shape of the hysteresis loop with a bilinear type is
large. Furthermore, in a bilinear model, it is not possible to express the hardening at
large deformation. Therefore, the various factors that affect rubber bearings are not
considered. Alternatively, when the objective of the analysis is to consider these various
factors or to track the large deformation behavior, curved system restoring force models,
such as the modified HD model1) or the Kikuchi-Aiken model2), have been proposed.
For elastic sliding bearings, a perfectly elasto-plastic restoring force model can be used.
In this model, the stiffness of the rubber bearing is the elastic stiffness, and the
maximum frictional force (= friction coefficient × supported load) is the yield load.
Rigid sliding bearings have rigid plastic properties, but in a time history response
analysis, a perfectly elasto-plastic restoring force model is often used with an extremely
large value for the elastic stiffness. Also, models have been proposed that consider the
change in friction coefficient from static friction to dynamic friction and the velocity
dependence of the friction coefficient3).
For problems with rocking vibrations or vibrations in the vertical direction, it is

necessary to define the relation between the force in the vertical direction and the
deformation. The relation for all isolators without tension force or uplift is assumed to
be linear elastic. The elastic stiffness in the vertical direction of isolators is extremely
high, and the value is the same as the stiffness of a reinforced concrete column on one
floor supporting the same vertical force. Therefore, for vibration problems in the vertical
direction, it is not sufficient to just consider the deformation of the isolators, but it is
also necessary to consider the axial deformation of the columns of the superstructure
and the deformation of the soil and the piles of the foundation structure. On the other
hand, if tension forces or uplift is generated on the isolators, the vertical stiffness
becomes extremely small, and so it is necessary to take this nonlinearity into
consideration. Nonlinear elastic models with bilinear or trilinear properties on the tensile
side, as shown in Fig. 4.4.9, are often used for modeling the restoring force properties of
isolators in the axial direction.
Force
Skeleton curve
Displacement
Fig. 4.4.8 Modified bilinear Fig. 4.4.9 Axial hysteresis model

hysteresis model for isolator shear for isolator vertical behavior
behavior
For the analysis with simultaneous input in two horizontal directions or torsional
vibrations, natural rubber bearings can be represented by two orthogonal linear shear
springs. For multifunctional type isolators, such as lead rubber bearings or
@Seismicisolation
@Seismicisolation
high-damping rubber bearings, the restoring force properties in a certain direction are
clearly affected by the deformation in the normal direction, so the horizontal biaxial
coupling effect must be considered and two orthogonal nonlinear shear springs must be
provided.
A practical method to analyze the horizontal biaxial coupling effect is the multiple shear
spring model (MSS), shown in Fig. 4.4.104). This model depicts the isotropy and the
biaxial coupling affect by arranging multiple springs having the same nonlinear
properties within the horizontal plane at equal angles. The advantage of the MSS model
is that the model can be easily extended, not just for bilinear models with unidirectional
properties, but also for hysteretic models such as trilinear models and modified bilinear
models, to consider the horizontal biaxial coupling effect5), 6). In the case of a model
with bilinear unidirectional properties, each shear spring stiffness ks and yield load qy in
the MSS model that are equivalent to the unidirectional stiffness Ks and yield load Qy
can be approximated by the following equations.
n
i
k s = K s / ∑ sin 2 π (4.4.1)
i =1 n
n −1
i
q y = Qy / ∑ sin π (4.4.2)
i =0 n
where n is the number of divisions in the circumferential direction.
Fig. 4.4.10 Multiple shear spring model
Understanding the horizontal biaxial coupling effect for multifunctional isolators is

being researched. Horizontal bidirectional loading tests have been carried out under
various conditions, and various mechanical models have been proposed to understand
the horizontal biaxial coupling effect7),8).
2) Dampers
Dampers include those that absorb the elasto-plastic hysteretic energy of lead or steel
rods, hysteretic dampers that use friction, and viscous dampers such as oil dampers or
viscous material dampers.
To represent the load deformation relation of hysteretic dampers, bilinear type or

trilinear type restoring force models are often used. However, curved surface type
restoring force models are also used to more faithfully reproduce the test results9).
@Seismicisolation
@Seismicisolation
Viscous dampers are often modeled as linear or nonlinear dashpots, or as Maxwell or

Voigt models.
When receiving bidirectional input motions or accommodating torsional vibrations, first

it is necessary to be assess whether the damper has directionality and, if so, to model
this directionality. Piston type oil dampers are clearly effective in only one direction, so
they have directionality. Elasto-plastic type dampers using steel rods or lead exhibit a
virtually uniform resistance force (energy absorption capability) with respect to
deformation in all directions within the horizontal plane. Measures can be taken to
eliminate directionality. However, the restoring force in each direction indicates that
bidirectional deformation cannot be treated independently, and it is necessary to take
into consideration the biaxial coupling effect that reflects this mutual interaction.
Regarding this biaxial coupling effect, modeling can be carried out in the same way as
for a multifunctional isolator, such as a lead rubber bearing or a high-damping rubber
bearing, which does not have directionality. In other words, when the properties can be
represented as perfectly elasto-plastic when force is applied in one direction, the
plasticity theory, in which the yield surface is circular, can be used. In the case of the
bilinear type when the secondary stiffness is not zero, the Ziegler kinematical hardening
model10) can be used. Alternatively, the restoring force model can be approximated with
an MSS model consisting of bilinear type restoring force property springs.
(3) Modeling the lower structure
In the case of normal base isolation, the lower structure below the seismic isolation level is
often not modeled and the lower ends of the seismic isolation devices are fixed. This is
because the period of the predominant vibration mode in the response of the lower structure
is quite short compared with the seismic isolation period, so the lower structure has very
little effect on the response of the superstructure and the seismic isolation level.
In cases where the lower structure contains structural stories, as in intermediate-level

seismic isolation, the modeling is similar to that of the superstructure. For example, if the
superstructure is a multi-degree-of-freedom model, the lower structure is also a
multi-degree-of-freedom model.
When the building response properties are to be evaluated by including the soil-structure
interaction, such as on sites with soft soil or a danger of liquefaction, a soil-structure
coupled model is used. In this model, everything is modeled, including the soil and the
piles. Also, a detailed soil-structure coupled modeling method is to model the soil as finite
elements, but the size of the model becomes extremely large and it is difficult to model the
nonlinearity of the building sufficiently. However, this coupled finite element model is an
effective means for studying the structure below the seismic isolation level, such as
retaining walls, underground basements, foundations, piles, etc11). On the other hand, some
methods evaluate the ground response or the interaction effect, carry out a separate
idealization of soil springs, apply sway and rocking soil springs that represent the
interaction effect on the seismic isolation model, or evaluate the nonlinear stiffness and
damping between the piles and the soil and applying soil springs to the pile side surfaces.
References
@Seismicisolation
@Seismicisolation
1) Yasuo Takenaka, Kazuhiko Yamada, Kazuhide Yoshikawa：Modified H-D Model：A New Smooth-Curve Hysteresis
Model of Laminated Rubber Bearings for Base Isolation, Journal of Technology and Design, AIJ, No.14, pp.87-92,
2001.12
2) Masaru Kikuchi and Ian. D. Aiken：An analytical hysteresis model for elastomeric seismic isolation bearings,
Earthquake Engineering and Structural Dynamics, Vol.26, pp.215-231, 1997.
3) Hiroshi Hibino, Masayoshi Takaki, Shouji Katsuta: Modeling of Friction Characteristics of Elastic Sliding Bearings
Based on Loading Test Using Full Scale Specimen，Journal of Structural and Construction Engineering, AIJ, No.574，
pp.45-52, 2003.12
4) Akira Wada, Keiichi Hirose: Elasto-Plastic Dynamic Behaviors of the Building Frames Subjected to Bi-Directional
Earthquake Motion, Journal of Structural and Construction Engineering, AIJ, No.399, pp.37-47, 1989.5
5) Hironao Seto, Yasuo Takenaka：MSS-Modified-Bilinear Model for Bi-Axial Nonlinear Hysteresiss of High Damping
Rubber Bearing and Lead Rubber Bearing, Summaries of Technical Papers of Annual Meeting, Architectural Institute
of Japan, pp.805-806, 1996.9
6) Masaru Kikuchi, Ken Ishii, Masashi Yamamoto, Harumi Yoneda, Shigeo Minewaki, Hideaki Kato：Bi-axial
Horizontal Response for the Seismically- Isolated Building with High-damping Rubber Isolation Bearings
(Part.1~Part.3)，Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan, pp.249-254,
2010.9
7) Hideaki Kato, Takahiro Mori, Nobuo Murota, Ken Ishii, Masaru Kikuchi：A Hysteresis Model of High-damping
Rubber Bearings using an Integral Deformation History Constitutive Law, Journal of Structural and Construction
8) Masashi Yamamoto, Shigeo Minewaki, Harumi Yoneda and Masahiko Higashino ： Nonlinear Behavior of
High-Damping Rubber Bearings Under Horizontal Bi-Directional Loading: Full-scale Tests and Analytical Modeling,
Earthquake Engineering and Structural Dynamics, Vol. 41, pp.1845-1860, 2012.
9) Masaru Kikuchi, Yoshihisa Kitamura, Keiichi Inoue, Masaiki Ueda：An Analytical Model for Seismic Isolation
Devices with Hysteretic Damping, Journal of Structural and Construction Engineering, AIJ, No.565, pp.63-71, 2003.3
10) Hans Ziegler: A modification of Prager’s hardening rule, Quarterly of Applied Mathematics, Vol. XVII, No.1,
pp.55-65, 1959.
11）Kiyoshi Fukutake, Kazuo Tamura：Analytical Evaluation of Isolation System Responses on Soft Ground, Journal of
Technology and Design, AIJ, No. 13, pp.43-48, 2001.7
@Seismicisolation
@Seismicisolation
4.4.3 Seismic Ground Motion Selection and Input Method
The time history response analysis is easily affected by the individual frequency properties
and the phase properties of the input ground motions. Therefore, it is desirable to set several
input ground motions. The ground motions that should be used include recorded waves from
the past, and spectrum fitting waves or site waves produced by the various methods described
in Section 4.2.
Recorded waves from the past can include recorded events obtained near a construction site
or significant recorded events such as El Centro, Taft, and Hachinohe. Recorded seismic
events near a construction site can be obtained from strong-motion seismograph networks
(K-net1), KiK-net2)) installed throughout Japan, from the Japan Meteorological Agency or
from local governments. Recorded seismic events obtained near a construction site have the
advantage that they reflect the properties of the site, but it is necessary to judge whether the
ground motions are from epicenters that should be considered and whether they have an
input level that should be considered. In the case of using the input level of ground motions,
the amplitude of the recorded event can be amplified. Among the significant recorded events,
it is necessary to be aware that in some old records, the duration time is short and they do not
have sufficient period components corresponding to the natural periods of seismically
isolated buildings. However, they are significant in that they have been widely used in design
in the past, so they can be used as benchmarks for comparison of the response to other
The response values from response spectrum fitting waves can vary greatly depending on the
method of applying the phase properties, so several phases should be assumed and several
ground motions should be prepared. For example, by using the phase information of actual
recorded events, it is possible to produce ground motions in which the spectral properties are
the same but the epicenter is close to or far from the building site. A recorded seismic event
obtained near the building site can be used. Also, by using the phase information of the NS
and EW directions of the seismic record, it is possible to produce ground motions that can be
used for investigating bidirectional horizontal input.
For site waves, it is desirable that at least two types of epicenters are postulated for inland
near-field earthquakes and inter-plate earthquakes, and that several ground motions are
produced.
If the input ground motions selected and produced by the methods described above are
unidirectional horizontal ground motions, it is desirable that the response values are studied
not only in the X and Y directions but also for input in the 45° direction and the 135°
directions. In particular, the response values obtained for input in the 45° direction and the
@Seismicisolation
@Seismicisolation
135° direction are important for studying the seismic isolation devices located in the corners
of the building. For simultaneous input in two horizontal directions, the NS and EW
components of the envisaged input ground motions can be simultaneously obtained by using
the empirical Green's function method as the method of producing the ground motions, and
the ground motions can be used with directional correction only, without amplifying the input
level. If horizontal bidirectional ground motions produced by other methods or recorded
events are directly used, it is necessary to set the input level and to standardize level by using
the maximum amplitude value of the main axial direction. The multiplication factor when
standardizing by using the maximum amplitude value of one component is applied to the
other direction, and as shown in Fig. 4.4.13(a), which is described later, a method of
multiplying by cosθ for a direction that is set at an angle θ and the other direction by sinθ can
be considered.
If the selected input ground motions are produced at the depth of engineering bedrock, it is
necessary to consider the amplification properties of the subsurface layer. To produce the
input ground motions to the building by seismic response analysis of the soil, often an E+F
wave (upward transmitting wave + downward transmitting wave) is applied at the level of
the bottom of the foundation, as shown in Fig. 4.4.11.
Fig. 4.4.11 Development of structure input motions from site response model
References
1) National Research Institute for Earth Science anf Disaster Preveution：http://www.k-net.bosai.go.jp/k-net/
2) National Research Institute for Earth Science anf Disaster Preveution：http://www.kik.bosai.go.jp/kik/
@Seismicisolation
@Seismicisolation
4.4.4 Torsional Response
In conventional structures, torsional vibrations due to the eccentricity in stiffness or mass of

the superstructure lead to stiffness degradation in the corner area that has the largest
inter-story drift. The eccentricity causes the concentration of damage and further degradation
in the corner area. In seismically isolated structures, even when eccentricity is present in the
superstructure, it is possible to reduce the effect of torsional vibrations by a design that
minimizes the eccentricity in the seismic isolation level by the arrangement of seismic
isolation devices. However, when seismic isolation devices with elasto-plastic properties are
used, it is necessary to ensure that the eccentricity of the seismic isolation level by using the
equivalent stiffness does not change greatly at each deformation level of seismic isolation
devices.
When evaluating the torsional response, the time history response analysis is effective. When
carrying out the time history response analysis that takes torsional vibrations into
consideration, it is necessary to construct the model in accordance with the properties of the
superstructure, the arrangement in plan of the seismic isolation devices, and objectives of the
analysis. When the superstructure has significant eccentricity, even when using a uniaxial
stick model with 3 DOF in the two translational directions and the torsional direction, as
shown in Fig. 4.4.3, it is necessary to consider the position of the center of gravity at the
position of each mass, and the position of the center of stiffness for the stiffness at each story.
Sometimes it is necessary to analyze the superstructure as a three-dimensional model. For
modeling the seismic isolation level, several horizontal biaxial coupled models such as MSS
models, are arranged in accordance with the plan arrangement of the seismic isolation
devices, and it is necessary to evaluate the response in the translational direction and the
direction normal to that direction.
The basic properties are examined for the case of superstructure eccentricity as an example of
the seismic response analysis for torsional vibrations due to eccentricity in the seismic
isolation level. According to the examination of the time history response analysis, the
horizontal deformation increase ratio due to eccentricity varies virtually proportionally to the
eccentricity distance. For the same eccentricity distance ratio, the horizontal deformation
increase ratio due to eccentricity tends to increase as the plan shape of the seismic isolation
level becomes long and narrow. Also, if the eccentricity ratio of the seismic isolation level is
about 3%, the obtained result shows that the response displacement increase ratio is about 1.1
compared with the case of no eccentricity.
@Seismicisolation
@Seismicisolation
4.4.5 Response in Two Horizontal Directions
(1) Restoring force properties of rubber bearings when subject to horizontal forces in two
directions
Normally, in dynamic analysis models of seismically isolated buildings, the load

deformation relation of the seismic isolation level (initial stiffness, secondary stiffness,
yield displacement) is set based on the results of horizontal unidirectional compression
shear tests under constant axial stress in full-scale rubber bearing test specimens. To
investigate the response in two horizontal directions, the horizontal biaxial coupling effect
of the restoring force properties is taken into consideration by using MSS models.
However, the ultimate properties in two horizontal directions do not differ much from the
properties in one horizontal direction, so the effect of two horizontal directions is not
considered.
The load deformation relation of high-damping rubber bearings when deformed in two
horizontal directions exhibits the characteristic hysteresis shape where the intercept load
increases greatly on both the long axis and the short axis. Figure 4.4.12 shows the load
deformation relation in the long axis and the short axis directions when high-damping
small-scale test specimens of a rubber bearing are deformed in two horizontal directions
(long axis: short axis = 2:1 elliptical orbit), and the results of reproducing this behavior by
two methods by using the MSS model (model 1) and the horizontal biaxial coupling model
(model 2) proposed by Yamamoto et al.1). According to Fig. 4.4.12, the MSS model does
not sufficiently express the increase of the intercept load that is seen in the test results. To
produce this property, it is necessary to use the horizontal biaxial coupling model1). Also,
the effect of horizontal two-dimensional coupling is significant in the load deformation
relation in the short axis direction.
From the above, when taking into consideration the horizontal bidirectional properties in
seismic isolation design, attention must be paid not only to the reduction in the ultimate
properties, but also to changes in the restoring force properties. However, mechanical
properties that are capable of expressing the horizontal bidirectional properties are not yet
generalized, so at present it is necessary to determine the trend of the response values
obtained from the current mechanical models, and to take into consideration a margin for
deformation in two horizontal directions during the design.
(2) Response of seismically isolated buildings during horizontal bidirectional ground

motion
Seismic response analysis was carried out for horizontal input in two directions for a case
using the MSS model (model 1) in accordance with the conventional method, and a case
using the horizontal biaxial coupling model (model 2)1).
The input ground motions used in two directions were taken from the notification spectrum
fitting waves of the Ministry of Land, Infrastructure and Transport (MLIT) produced for the
NS component and the EW component of recorded seismic events using various phase
information. As shown in Fig. 4.4.13(a), the seismic input energy for unidirectional input
and bidirectional input had to be virtually constant, so a certain angle θ was set and the NS
component of the notification spectrum fitting wave was multiplied by cosθ and the input in
@Seismicisolation
@Seismicisolation
the X direction, and the EW component was multiplied by sinθ in the Y direction and both
inputs were applied simultaneously. Angle θ is a parameter that represents the ratio of the
NS component and the EW component, and in the following it is referred to as the
distribution parameter.
Experiment
Experimental Analysis
40 40
20 20
Shear force (kN)
Shear force (kN)

0 0
Major axis Minor axis

-20 -20
-40 -40
-100 -50 0 50 100 -100 -50 0 50 100
Displacement (mm) Displacement (mm)
(a) MSS model (Model 1)
40 40
20 20
Shear force (kN)
Shear force (kN)
0 0
Major axis Minor axis

-20 -20
-40 -40
-100 -50 0 50 100 -100 -50 0 50 100
Displacement (mm) Displacement (mm)
(b) Yamamoto model1) (Model 2)

Fig. 4.4.12 Simulation analyses of bi-directional loading tests of high-damping rubber
bearings2)
5 rows
55 columns
columns
2525elastomeric
elastomericisolators
isolatord
inin
5 x5 5x grid
5 grid
(a) Input ground motions (b) Structure model

Fig. 4.4.13 Horizontal bi-directional seismic response analysis
@Seismicisolation
@Seismicisolation
A total of 25 units of laminated rubber bearings in the above two models were arranged in
five rows by five rows in the X and Y directions, and the multi-degree-of-freedom dynamic
analysis model shown in Fig. 4.4.13(b) was produced. The distribution parameter θ for this
model was varied from 0° to 90°, and the seismic response analysis was carried out for
simultaneous horizontal input in two directions, and the variation in response values due to
the differences in the components in the two directions was investigated3).
From the response values for input in two horizontal directions, Fig. 4.4.14 shows the
average shear stress obtained by dividing the shear force in the laminated rubber bearings
by the cross-sectional area, the torsional stress obtained by dividing the torsional moment
produced in the laminated rubber bearings by the polar modulus of the section, and the
maximum shear stress when the two are added together as a time history waveform.
When the partition parameter θ is either 0° or 90°, unidirectional input comes from either
the NS component or the EW component and no torsional moment is produced within the
laminated rubber bearings. On the other hand, when partition parameter θ is between 0° and
90°, the laminated rubber bearings deform in two directions and a torsional moment is
produced, and a shear stress is produced in the side surface of the laminated rubber
bearings. The unidirectional input (0° or 90°) and smaller at intermediate θ produces a
V-shaped distribution shape, in which the average shear stress is large but the torsional
stress has an chevron-shaped distribution shape with the torsional stress greatest when the
partition parameter θ is between 30° and 60°. In the model 2 results (denoted by ● in Fig.
4.4.14) for the maximum shear stress, which is the sum of the two, the maximum is not
necessarily during unidirectional input. In this figure, the maximum value is at θ = 70°, a
value that is about 10% greater than that for unidirectional input. On the other hand, in
model 1 (denoted by ○ in Fig. 4.4.14) of the conventional model (MSS model), this
tendency is not found. This fact suggests that in the analysis of input motion in only one
direction or in the analysis of horizontal simultaneous input in two directions by using
conventional modeling, the maximum shear stress produced in the laminated rubber
bearings may be evaluated on the unsafe side. From the above, when taking into
consideration the horizontal response of seismically isolated buildings in two directions by
using conventional modeling, it is necessary to incorporate a certain margin for the
response values obtained from the unidirectional response of the laminated rubber bearings.
2.0 2.0 2.0

(MPa)
(MPa)
Model 1
(MPa)
1.5 1.5 1.5

(MPa)
(MPa)
(MPa)
Model 2
stressstess
stressstess
stess
stress
1.0 1.0 1.0

Shearshear
Shearshear
shear
Shear
Torsional
Average
Maximum
0.5 0.5 0.5
0.0 0.0 0.0

0 30 60 90 0 30 60 90 0 30 60 90
Parameter θ (deg) Parameter θ (deg) Parameter θ (deg)
(a) Averaged shear stress (b) Additional stress due to (c) Maximum shear stress
torsional moment
Fig. 4.4.14 Response shear stress in elastomeric isolator
@Seismicisolation
@Seismicisolation
References
1) Masashi Yamamoto, Shigeo Minewaki, Harumi Yoneda, Masahiko Higashino, Akira Wada：Full-scale and Analytical
Modeling of High-damping Rubber Bearings Under Two Horizontal Directional Loading, Journal of Structural and
2) Masaru Kikuchi, Ken Ishii, Masashi Yamamoto, Harumi Yoneda, Shigeo Minewaki, Hideaki Kato：Bi-axial
Horizontal Response for the Seismically-isolated Building with High-damping Rubber Isolation Bearings
Part.1~Part.3，Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan, B-2, pp.249-254,
2010. 9
3) Takahito Nakamura, Masaru Kikushi, Ken Ishii, Yoriko Suzuki：Mechanical Properties of Lead Rubber Bearings
under Biaxial Loading (Part.1: Biaxial Loading Tests), Summaries of Technical Papers of Annual Meeting,
Architectural Institute of Japan, B-2, pp.505-506, 2011.8
4.4.6 Vertical Motion Response
When it is necessary to evaluate in detail the fluctuation in the axial forces of isolators, as in
super high-rise seismically isolated buildings, or to evaluate the maximum forces produced in
large-span beams or slabs, a vertical time history response analysis is carried out. For
evaluation of the vertical forces of the isolators and the superstructure members, the analysis
is carried out with simultaneous horizontal and vertical inputs by using the analysis models
shown in Fig. 4.4.6 or Fig. 4.4.7. Or, separate horizontal and vertical analysis models can be
produced and a separate time history response analysis carried out, and then the maximum
responses for the horizontal motions and vertical motions are combined.
Methods of combining the separately obtained response values for horizontal motions and
vertical motions can include (1) the method of adding time histories, (2) the method of
adding absolute values, (3) the SRSS method, (4) the combination coefficient methods, etc1).
The method of adding time histories is to simply add the rocking response value and the
vertical response value at the same analysis time to obtain the maximum values. Various
ground motions range from those in the time period in which the main motions of the
horizontal motions and the vertical motions virtually coincide, to those in which a time
difference of up to several seconds is produced. So, large differences can be produced in the
results due to the earthquake ground motions when the method of adding time histories is
adopted2). In the method of adding absolute values, the maximum values for each horizontal
motion response and vertical motion response are simply added. When the time in which the
main motions of the horizontal motion and the vertical motion are virtually coincidental, the
results obtained are close to those of the method of adding time histories; of the various
methods of combining, this method evaluates on the safest side. The SRSS method obtains
the square root of the sum of squares of the maximum response values for the horizontal
@Seismicisolation
@Seismicisolation
motions and the vertical motions. In the combination coefficient method, the smaller of the
horizontal motion maximum response value and the vertical motion maximum response
value is multiplied by a combination coefficient less than 1. The SRSS method and the
combination coefficient method are convenient methods in practice. The method of adding
the absolute values is generally an over-evaluation and does not have a theoretical basis.
To date, a clear opinion has not been obtained regarding the analysis model properties in the
vertical direction for seismically isolated buildings. The vertical transfer functions of the
seismic isolation level obtained from recorded earthquake ground motions show large
differences depending on the building. In some cases, the primary mode in the vertical
direction is clearly seen, and in some cases several closely spaced modes are seen, and large
variations appear in the damping ratios in the range 2% to 20%3)-6). This presumed reason is
that the out-of-plane stiffness of floor slabs and beams is low, so each member vibrates in
their own natural vibration mode in the vertical direction7), which suggests a limitation to the
stick model with a fixed foundation. To accurately evaluate the vertical motion response, it is
necessary to use frame models or three-dimensional models8), or coupling models taking into
consideration the interaction effect with the soil9).
References
1) Takashi Nishiyama, Tatsuya Sugiyama, Takuya Kashiwazaki：Study on Combination of Horizontal and Vertical
Earthquake Loads using Earthquake Records, Summaries of Technical Papers of Annual Meeting, Architectural
Institute of Japan, B-2, 1994.7
2) Yuichi Ikeda, Sumio Nagahashi：Effects of Simuataneous Action of Horizontal and Vertical Ground Motion on Uplift of
Isolator of Base Isolated High-rise Apartment Houses, Journal of Structural and Construction Engineering, AIJ, No.515,
1999.1
3) Arihide Nobata, Akira Teramura, Yuzuru Yasui：Vertical Response Property of Base-isolated Building Based on
Observed Earthquake Rrecords, Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan,
1997.9
4) Kiyotaka Matsumoto, Tsuneo Takazawa：A Study on Vertical Characteristics of Base Isolation Building: Earthquake
Records and Microtremours, Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan,
pp.513-514, 2003.9
5) Tadayuki Morishita, Yoshihito Saito, Takaharu Yoshida, Hiroaki Ryujin：Properties of Vertical Vibration Through
Earthquake Observation on a Base Isolated Building, Summaries of Technical Papers of Annual Meeting, Architectural
Institute of Japan, pp.293-294, 2004.8
6) Mitsuru Uryu, Toshihiko Yamazaki, Toshinari Kondo, Ryuji Nakanishi：A Study on the Response Characteristics of
Base-Isolated Structures for Vertical Earthquake Motion (Part.10 A Study Based on Excitation Test and Seismic
Response Analysis), Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan, pp.381-382,
2004.8
7) Haruyuki Kitamura, Yang ZhiYong, Hideyuki Tada：An Observation of Vertical Vibration Response on Base Isolated
Buildings, Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan, pp.817-818, 1996.9
@Seismicisolation
@Seismicisolation
8) Rei Kurose, Haruyuki Kitamura, Daiki Satou, Masaki Tabei：Modeling Method of Vertical Response Analysis of
Base-Isolated Building Based on Multi-Points Earthquake Observation Record：Part.2 A Proposition of Analysis
Method, Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan, pp.321-322, 2008.9
9) Masayuki Nagano, Masanori Iiba, Toshihide Kashima：Vertical Response Characteristics of a Seismically Isolated
9-story Building During Small and Large Earthquakes and its Dynamic Soil-structure-interaction Effects, Journal of
Structural and Construction Angineering, AIJ, No.647, 2010.1
4.4.7 Effect of Variation in the Seismic Isolation Devices on the Response
Seismic isolation devices have the advantage that loading tests are carried out on all devices
before they are installed in the building. If the mechanical properties are outside the specified
values, the devices are not used, so it is possible to clearly determine the variation of the
mechanical properties at the time of design of the building. On the other hand, seismically
isolated buildings have a low degree of redundancy and the majority of the energy input to
the building is absorbed in the seismic isolation level. As a result, the effect of the
characteristics of seismic isolation devices on the overall seismic response of the building is
extremely large. Therefore, during the design of seismically isolated buildings, it is important
to appropriately evaluate the stiffness, yield load, and equivalent damping ratio of isolators
and dampers and their variation, and to reflect this in the analysis models.
For seismic isolation devices manufactured in Japan, the manufacturing variation of each
individual product with respect to standard values and the amount of variation of the main
properties due to the temperature environment in which they are used are published for the
main properties such as stiffness, and yield load. In addition, for laminated rubber bearings
and elastic sliding bearings, the properties of the rubber will vary with age due to oxidation
degradation, so the amount of variation in the stiffness and the equivalent damping ratio after
60 years is predicted based on chemical reaction kinetics. Also, the variation in properties of
rubber bearings 20 years after installation has been reported for seismically isolated buildings
that were constructed in the comparatively early stages1), 2). This published data should be
referred to for evaluation of the variation in seismic isolation devices.
The factors that govern the seismic response of seismically isolated buildings due to the
variation in the various properties of the seismic isolation devices depend on the types of
seismic isolation devices used and their combination, the range of variation in installation,
and the properties of the input ground motions. In an example of a sensitivity analysis using
the Monte Carlo method for single-degree-of-freedom models of natural rubber bearings and
steel dampers, it was found that the variation in the yield capacity of the steel dampers was
most dominant3). On the other hand, in practice, the variation in the main properties of
seismic isolation devices are often considered to be completely independent, and several
@Seismicisolation
@Seismicisolation
analysis cases have been analyzed by using a time history response analysis with the upper
limit values and the lower limit values of each property, and each combination case has been
checked to determine that the design criteria are satisfied. A convenient method for
evaluating the variation in the seismic response due to variations in the properties of the
seismic isolation level has been proposed by using the response spectrum (SA-SD curve) of
the input ground motions4). When using this method, it is possible to first focus on the
variation factors with high sensitivity, and then reduce the number of analysis cases of the
time history response analysis to carry out an efficient investigation of the variation.
References
1) Hiroki Hamaguchi, Satoru Aizawa, Yuusuke Samejima, Takashi Kikuchi, Shigenobu Suzuki, Toshikazu Yoshikawa：A
Study of Aging Effect on a Rubber Bearing After About Twenty Years in Use, Journal of Technology and Design, AIJ,
No.30, pp.393-398, 2009.6
2) Tomotaka Wake, Mitsuru Miyazaki：Survey of Aging for LRB Completed 20 Years Ago, Summeries of Technical
Paper of Annual Meeting, Architectural Institute of Japan, pp.391-392, 2008.9
3) Tsutomu Hirotani, Masaaki Ohba, Yasuyuki Murazumi, Junji Suhara：Study on Structural Design for Seismic Isolated
FBR plant.：Part.3 Influence on Seismic Response by Variation of Material Properties. Evaluation of Variation of
Seismic Response, Summeries of Technical Paper of Annual Meeting, Architectural Institute of Japan, pp.1037-1038,
1995.8
4) Kohju Ikago, Norio Hori, Atsushi Nakaigawa, Norio Inoue：A Study on Effects to Seismic Response of Base-isolated
Structures Caused by Differnces of Damping Mechanism and Deviations of Device Characteristics, AIJ Tohoku
Chapter Architectural Research Meeteing, 2009.6
4.4.8 The Effect on Response of Changes in Properties Due to Repeated Deformation
In multi-functional isolators that have a damping function, such as lead rubber bearings,
high-damping rubber bearings, and sliding bearings, the seismic input energy is absorbed as
hysteretic energy by the isolators. The hysteretic energy is then converted into heat energy.
When subjected to long period ground motions, particularly earthquake ground motions with
large amplitude and long duration, changes occur in the properties, such as stiffness and
resistance force, due to the rise in temperature of the seismic isolation devices during
repeated deformation. These changes can affect the response of seismically isolated buildings.
Figure 4.4.15 shows the results of a sine wave test with repeated loading carried out on a lead
rubber bearing (LRB), and Fig. 4.4.16 shows the results of a seismic response wave shaking
test1).
According to the test results, the yield load is reduced with repeated deformation, and the
hysteretic energy absorbed per loop is reduced. In these tests, the results showed a
temperature rise up to 150°C in the lead plug portion, and after the test was terminated and
@Seismicisolation
@Seismicisolation
the temperature returned to ambient temperature, it was confirmed that no change occurred in
the basic properties before and after the repeated loading test. This property is seen in
isolators where the plug portion is heated and the plug itself is affected, such as lead-plug
rubber bearings, and tin-plug rubber bearings. The properties during repeated loading are
greatly affected by heat, and the results are also affected by the size of the seismic isolation
devices as a scale effect, so it is necessary to carry out tests on full-scale seismic isolation
devices as much as possible, and to carry out dynamic loading tests at the actual velocities
and actual displacements. In high-damping rubber bearings, the energy is absorbed by the
rubber bearings as a whole, and the effect of the temperature rise due to heating has been
confirmed to be comparatively small. However, it is necessary to be aware of the change in
properties during loading due to the effect of load hysteresis and vibration intervals.
Q(kN) Ratio with 3rd cycle

1200 1.2
Shear Force (kN)
Ratio to 3rd cycle
0.8
0
0.4
Qd
Kd
-1200 0.0
-500 0 500 0 5 10 15 20
Displacement (mm)
D(mm) Cycle Cycle
(a) Load-Displacement Rlation (b) Changes in Qd, Kd

Fig 4.4.15 LRB φ1000 results under sinusoidal input (period=3 sec，γ =100%)
Q(kN)
1200
Shear Force (kN)
-1200
-500 0 500
Displacement (mm)
D(mm)
Fig 4.4.16 LRB φ1000 results under earthquake response input
(Nagoya Sannomaru (EW) response wave)
In consideration of the above property changes in connection with the response of

seismically isolated buildings, the matter of most concern is mainly an increase in horizontal
deformation of the seismic isolation level during an earthquake. Furthermore, the input
energy could also increase depending on the period properties of the input ground motions
and the seismic isolation level. The results of a seismic response analysis study taking into
consideration the change in properties due to repeated deformation of lead rubber bearings
are presented. An evaluation was performed by a time history seismic response analysis
considering the rise in temperature of the lead rubber bearings. The test was conducted under
the assumption that all the hysteretic energy was consumed as heat in the lead plug and the
yield load of the lead plug was corrected at each time interval based on the set relation
between the lead temperature and the yield stress after calculating the temperature in the lead
@Seismicisolation
@Seismicisolation
plug by performing an analysis of heat transfer in the rubber bearings2). According to the
analysis results, the maximum response displacement of the seismic isolation level did not
necessarily increase significantly for the set building model and input ground motions, but it
is considered that in some cases the amplitude would be greatly increased due to the input
amplification factor and the period properties of the input ground motions. In particular, the
effect can be large for input ground motions having a predominant period2).
Various methods have been proposed for evaluating the response when considering the
change in properties due to the temperature rise of lead rubber bearings, such as evaluation
by a time history response analysis that accounts for the reduction in yield load from the
absorbed hysteretic energy3), or evaluating the response results that account for the rise in
temperature while using the energy balance equations from the seismic response analysis
results that did not consider the temperature rise4).
From the above, it is necessary to consider the effect of the temperature rise of the seismic
isolation devices in the actual design, or to take the property changes due to repeated
deformation into consideration, or to carry out an evaluation that sets the likely reduction in
properties of the seismic isolation devices in advance in order to achieve safer seismically
isolated buildings.
References
1) Yasuo Takenaka, Akihiro Kondou, Eiji Takaoka, Makiko Hikita, Haruyuki Kitamura, Takahito Nakamura：
Experimental Study on Heat-mechanics Interaction Behavior of Laminated Rubber Bearings, Journal of Structural
and Construction Engineering, AIJ, No.646, pp.2245-2253, 2009.12
2） Akihiro Kondo, Yasuo Takenaka, Eiji Takaoka, Makiko Hikita, Haruyuki Kitamura, Mitsuru Miyazaki：
Heat-Mechanics Interaction Behavior of Laminated Rubber Bearings under Large and Cyclic Lateral Deformation，
3） Tomonori Honma, Rei Kurose, Haruyuki Kitamura, Takahito Nakamura, Kazuhide Yoshikawa, Yasuo Takenaka：
Heat-Mechanics Interaction Behavior of Laminated Rubber Bearings under Large and Cyclic Lateral Deformation，
4） T. Honma, A. Kondo, Y. Takenaka, H. Kitamura：Seismic Response Prediction Method Considering Heat-Mechanics
Interaction Behavior of Lead Rubber Bearing, 5th World Conference on Structural Control And Monitoring, 2010.7
4.4.9 Plasticity of the Superstructure
Normally, if a seismically isolated structure is adopted, the response value of the

superstructure is reduced compared with that of the fixed-base case, so the structural design
is carried out by using design seismic forces that are smaller than that in the conventional
design. If the load resistance is reduced too much, plasticity of the superstructure will occur
@Seismicisolation
@Seismicisolation
even if a seismically isolated structure has been adopted. Also, if the earthquake ground
motion is greater than predicted, the seismically isolated building will impact the retaining
wall due to excessive response deformation of the seismic isolation level, or the seismic force
on the superstructure will increase due to hardening of the rubber bearings.
The nonlinear responses of two types of building models are compared in Fig. 4.4.17. For the
seismic isolation model and the fixed-base model of a 5-story and a 15-story building1), the
relation between the maximum input ground motion acceleration and the first-floor
inter-story deformation was compared for gradually increasing input ground motions. The
input ground motions were prepared by matching the response spectrum properties to the
notification spectrum by following the JMA Kobe NS component phase information. The
building model was a multi-degree-of-freedom model, with the restoring force properties of
the Takeda model assuming an RC superstructure. Also, the load-carrying capacity of the
superstructure of the seismic isolation model was set at three levels, 1.0, 0.75, and 0.5 times
that of the fixed-base model.
The maximum acceleration of the original earthquake ground motion was 381 Gal. At that
input level, the response values of each seismic isolation model was 1/2 or less than the
fixed-base case, so a sufficient seismic isolation effect can be expected. When the input
ground motion level was increased, the response values for both the seismic isolation and the
fixed-base cases increased, but the trend differed greatly between the two. The major
difference between the two was that the response of the fixed-base model changed at
virtually a constant slope, but the seismic isolation model showed a ground motion
acceleration level at which the response values suddenly increased. This ground motion
acceleration level was the level at which the superstructure reached the yielding capacity, and
the lowered capacity of the superstructure appeared at that level. Also, as the input level
increased beyond yielding, the response values of the seismic isolation model increased
dramatically, and in some cases, exceeded the response values of the fixed-base model. This
type of trend, which has been confirmed for other earthquake ground motions and
seismically isolated building models, is one of the universal vibration properties of
seismically isolated structures. So, it is necessary to take care in setting the load-carrying
capacity of the superstructure2).
The nonlinear behavior of the seismically isolated structure described here can be explained
by specific factors in the nonlinear behavior from the point of view of nonlinear steady state
vibration of two mass vibration systems3). The superstructure of a seismically isolated
structure can be interpreted as vibrating due to the long period seismic forces through the
filter of the seismic isolation level. The seismic isolation effect of seismically isolated
buildings as a whole is exhibited by vibrating in the period region with a low amplification
rate by removing the predominant periods of the earthquake ground motions. On the other
@Seismicisolation
@Seismicisolation
hand, long period external forces that are close to static forces act on the superstructure.
Normally, when a fixed-base building becomes plastic, hysteretic damping acts effectively
and the response displacements are reduced. The current seismic design code incorporates
this effect as a structural property factor Ds, which represent the structural characteristics
considering damping and ductility of each story, so it is possible to reduce the necessary
load-carrying capacity. However, under the situation where long period external forces that
are close to static forces act, hysteretic damping due to multi-repeated deformation is not
expected. The seismically isolated building is under this situation; when the stiffness is
reduced due to plasticity, it is not possible to suppress the deformation, and so the stiffness
reduction and plasticity further progress. From the point of view of redundancy when a
building is damaged, this can be interpreted as the fixed-base structure has higher redundancy
than the seismically isolated structure.
Fixed, 1.0V
Isolated, 1.0V
Isolated, 0.75V
Isolated, 0.50V
6.0 6.0
5.0 5.0
Interstory drift (cm)
Interstory drift (cm)
4.0 4.0
3.0 3.0
2.0 2.0
1.0 1.0
0.0 0.0
0 200 400 600 800 1000 0 200 400 600 800 1000
Input ground acceleration (Gal) Input ground acceleration (Gal)
(a) 5-story building model (b) 15-story building model

Fig. 4.4.17 First-story inter-story drift versus peak ground acceleration
Notes:
1) Input ground motion is JMA Kobe NS component matching the Building Standard Law (MLIT) design spectrum.
2) V is the lateral shear capacity of the fixed-base structure.
The above discussion has no meaning as long as seismically isolated buildings keep the
elastic property. It is necessary to be aware that seismically isolated buildings have the
property that when the superstructure becomes plastic, the inter-story deformation
dramatically increases, so excessive reduction in the load-carrying capacity must be avoided,
and it is necessary to design so that elasticity is maintained as much as possible.
References
1) Architectural Institute of Japan：Structural Response and Performance for Long Period Seismic Ground Motions
2) Masaru Kikuchi, Kazuo Tamura, Akira Wada：Safety Evaluation of Base-Isolated Structures, Journal of Structural and
@Seismicisolation
@Seismicisolation

3) Masaru Kikuchi, Cameron J. Black, Ian D. Aiken：On The Response of Yielding Seismically Isolated Structures,
Earthquake Engineering and Structural Dynamics, Vol. 37, pp. 659-679, 2008.
@Seismicisolation
@Seismicisolation
Section 4.5 Wind Design
4.5.1 Scope of Application and Basic Policy
In this section, the wind design of seismically isolated buildings is described. In seismically
isolated buildings, the wind response characteristics are different from those of the seismic
response, so the wind characteristics must be evaluated and reflected in the design. The wind
response includes the velocity dependence and amplitude dependence of the seismic isolation
devices, and the horizontal creep deformation produced by external static forces. The wind
load characteristics include a mean component that can be treated as a static external load
and a fluctuating component with a peak lower than that of the seismic load, and power
ranging from low frequencies to high frequencies, including the natural frequency of the
building. Also, the duration can exceed several hours, in contrast to the case of an earthquake,
where the duration ranges from several tens of seconds to at most several minutes. Horizontal
creep deformation is produced in some seismic isolation devices due to static external forces,
and some devices have amplitude dependence and velocity dependence. Therefore, it is
necessary to study the horizontal creep deformation, velocity dependence and amplitude
dependence of seismic isolation devices with consideration of the wind load characteristics in
environments not subjected to earthquake ground motions.
As a result of installing a seismic isolation level in a seismically isolated building, the

stiffness of the building is reduced compared with a normal building of the same scale. It is
possible that a wind load generating perceptible vibrations occurs comparatively frequently,
in other words, several times per year. Also, it is possible that the reduction in the wind
velocity generates unstable aerodynamic vibrations. It is necessary to examine these wind
characteristics in detail.
Based on the above, the basic policy of wind design in this recommendation is as follows.
(i) For extremely rare strong winds, the design is intended to ensure stability against the
wind. Here, stability refers not just to the maximum response, but also to the fatigue
damage of the isolation devices due to multiple repetitions. By taking into consideration
that the duration of the wind load is long and its frequency of occurrence is high,
information on the wind load can be obtained for evaluating the fatigue damage for one
extremely rare strong wind. However, because it is necessary to evaluate fatigue damage
for rare strong winds in conjunction with the earthquake load, such an evaluation is
outside the scope of application of this recommendation.
(ii) Wind load is calculated with reference to "Recommendations for Loads on Buildings"1),
published by AIJ, and the enforcement ordinances of the Building Standards Law. In
particular, in some cases of high-rise seismically isolated buildings sensitive to wind
load, it is necessary to consider the wind load in the across-wind direction and the wind
load in the torsional direction.
@Seismicisolation
@Seismicisolation
(iii) Three levels2) are set for the response level of the seismic isolation level (rank A, rank B,
rank C), and their definition and matters for investigation have been presented. In any
case, there is no problem if all members including the seismic isolation devices are
designed in the conventional way within the elastic range. However, if it is necessary to
set the yield load of the seismic isolation level low so that a seismic isolation effect is
produced during an earthquake, the seismic isolation devices will not necessarily be
restricted to be within the elastic range under the wind load, and plasticity could be
produced. In this case, it is necessary to evaluate the fatigue damage due to the
maximum response values of the seismic isolation devices and repetitions over a long
duration by time history response analysis. If the level of plasticity of the seismic
isolation devices due to wind load is slight, then it will not be a problem to design
generally within the elastic range, and the response can be evaluated by the static
equilibrium of forces by the conventional equivalent static wind load. Also, the data on
the behavior properties under repetition and horizontal creep deformation are being
continually enhanced. Therefore, as indicated below, the behavior of the seismic
isolation level is classified into three ranks: within the elastic range (rank A), virtually
elastic and only slightly exceeding the elastic limit (rank B), and greatly exceeding the
elastic limit (rank C). (For details, see Item 4.5.3)
(iv) Habitability, which can be a problem in high-rise buildings in particular due to wind
oscillations, should additionally be examined. Also, unstable aerodynamic vibrations
should be examined, because the potential for their occurrence increases as the natural
frequency reduces. In addition, an example problem of the trade-off with seismic loads
in terms of design has been provided and its importance discussed. The seismic isolation
level is provided in seismically isolated buildings, so it is unavoidable that the stiffness
is reduced compared with normal buildings of the same scale. Perceptible vibrations are
generated in buildings with low rigidity by wind velocities that occur comparatively
frequently, namely, several times per year and this is particularly so for high-rise
buildings. This, of course, reduces the habitability of the building, so it is necessary to
sufficiently examine the seismically isolated building more than a normal building of the
same size. Also, the possibility of generation of unstable aerodynamic vibrations
increases as the natural frequency reduces due to the reduction in rigidity. The
occurrence of unstable aerodynamic vibrations is virtually determined by the value of
the non-dimensional velocity Vr (= V/f0B, V: wind velocity, f0: building natural
frequency, B: building aspect width), which is determined by the building shape. As Vr
has a constant value, if the natural frequency f0 is low it is easy to understand that the
wind velocity V at which unstable aerodynamic vibrations occur is low.
References
1) Architectural Institute of Japan：Recommendation for Loads and Buildings, 2004
2) The Japan Society of Seismic Isolation：Guidelines for Wind-resistant Design of Base-Isolated Buildings, 2012
@Seismicisolation
@Seismicisolation
4.5.2 Calculation of Wind Load
Wind loading on seismically isolated buildings is calculated with reference to AIJ’s

"Recommendations for Loads on Buildings"1). Wind tunnel tests can be used for calculating
the wind force coefficients.
The standard for the design of wind is obtained from the enforcement ordinances for the
Building Standards Law. The standard wind velocity for a rare strong wind is at 10 m above
the ground. The ground surface roughness category is taken into consideration. The standard
wind velocities defined here correspond to V0 in the enforcement ordinances for the Building
Standards Law, and are used as the design wind velocity in a rare strong wind. The standard
wind velocity for an extremely rare strong wind is 1.25 times a rare strong wind. On the other
hand, the wind load in an extremely rare strong wind is defined as 1.6 times the wind load
during the rare strong wind, but the relation to wind velocity is not stated. The calculation of
the wind load here is required not only for evaluation of the maximum response values, but
also for evaluation of the safety against fatigue damage. In other words, it is convenient to
define the standard wind velocity with respect to an extremely rare strong wind. Therefore,
when a return period of 50 years is defined as rare and a return period of 500 years is defined
as extremely rare, 1.25 times the standard velocity in a rare strong wind is the standard
velocity in an extremely rare strong wind.
The duration of the wind load is defined to include the wind properties at the site and various
design conditions (for example, the specification for maintenance inspection). The duration
here refers to a single strong wind event, but depending on the maintenance and inspection
conditions, it may be necessary to consider the cumulative damage. The performance of
dampers in the seismic isolation devices could possibly be reduced due to fatigue damage
caused by repeated wind loading over a long period of time. However, to avoid this situation,
sufficient performance must be exhibited even after being subjected to repeated loading.
Specifically, it is necessary to confirm that damage is not caused by an extremely rare strong
wind that may or may not be experienced once during the service life. For this purpose, it is
important to appropriately consider the duration of the wind loading.
The calculation of wind loading is based on AIJ's "Recommendations for Loads on

Buildings”, but it may also be evaluated based on other methods such as a time history
response analysis or a frequency response analysis.
References
1) Architectural Institute of Japan：Recommendation for Loads and Buildings, 2004
@Seismicisolation
@Seismicisolation
4.5.3 Design of the Seismic Isolation Level
The wind load acting on a building is normally smaller than the seismic load. So, in
structural engineering practice, the building’s load-carrying capacity is determined based on
the seismic load, and checking against the wind load is frequently omitted. However, in
seismically isolated structures, the horizontal stiffness of the seismic isolation level is
extremely small compared with each of the stories of the superstructure, so it is necessary to
verify the safety of the seismic isolation level and the seismic isolation devices with respect
to wind load.
In the "Guidelines for Wind-resistant Design on Base-isolated Buildings"1), the response

level of the seismic isolation level to the wind load is classified into three ranks (rank A, rank
B, rank C), and the seismic isolation level and the seismic isolation devices are verified for
safety in accordance with their respective rank. Figure 4.5.1 shows the schematic diagram for
the response of the seismic isolation level to the wind load and the seismic isolation level
ranks1).
The classification into specific response levels and the methods of verification are as follows.
Rank A: Cases where the behavior of the seismic isolation level with respect to wind load
is within the elastic range
Rank B: Cases where the elastic limit of the seismic isolation level is exceeded when
subjected to wind load, but the behavior with respect to the fluctuating
component of the wind load is elastic
Rank C: Cases where the elastic limit of the seismic isolation level is exceeded when
subjected to wind load, and the behavior is elasto-plastic with respect to the
fluctuating component of the wind load
In the case of elastic behavior of the seismic isolation level when subjected to a wind load
(rank A), the safety of the seismic isolation level can be sufficiently checked in accordance
with conventional wind design. Also, in cases where the elastic limit is slightly exceeded due
to a wind load (rank B), the response of the seismic isolation level does not differ greatly
from elastic behavior at the macroscopic level2), and it is known that the response of the
seismic isolation level can be evaluated by the static equilibrium of forces in accordance with
the conventional equivalent static wind load2). On the other hand, in cases where the elastic
limit of the seismic isolation level is exceeded and the behavior is elasto-plastic (rank C), the
wind load is changed by the plasticity of the seismic isolation devices. In addition, the
@Seismicisolation
@Seismicisolation
properties of the seismic isolation level change as a result of long-term loading of the seismic
isolation devices. So, it is necessary to evaluate the safety by considering these changes. As a
rule, in the case of rank C, a time history response analysis is carried out to verify the seismic
isolation devices against the maximum deformation, fatigue damage, and energy absorption
performance.
Shear force of seisimic balance point curve of seismic isolation

isolation level (Q) level with mean wind load
skeleton curve of ( without creep deformation )
seismic isolation level
(for seismic design)
Rank C
fluctuating
Rank B wind load
fluctuating
Rank A wind load
fluctuating
wind load
creep Rank B’ mean
deformation
mean
mean
Horizontal deformation of
seismic isolation level (δ)
Fig. 4.5.1 Schematic diagram of wind-load induced response of seismic isolation level and
ranks of seismic isolation level
When selecting the seismic isolation devices for the seismic isolation level, it is necessary to
consider the characteristic properties with respect to the wind load, such as fatigue damage,
reduction in stiffness, reduction in load-carrying capacity due to long-term repeated loading
during wind, the effect of increased temperature associated with the absorption of energy,
horizontal creep properties, and the low amplitude properties. Furthermore, it is necessary to
verify that the performance is sufficient with respect to cumulative deformation and
cumulative absorbed energy.
The safety of each individual seismic isolation device must be verified by considering the
characteristic properties of the wind load (sum of static and dynamic components, long
duration), which are different from the seismic load.
In "Guidelines for Wind-resistant Design on Base-isolated Buildings"1), the following five
@Seismicisolation
@Seismicisolation
items are given for evaluation of the seismic isolation devices with respect to wind load.
(i) Confirm that performance of the seismic isolation devices is maintained (during
strong winds and afterwards).
(ii) Appropriately consider fatigue damage, reduction in stiffness, reduction in

load-carrying capacity, the effect of temperature rise, creep deformation in the
horizontal direction, and small amplitude properties, against repeated loading over
long period of time.
(iii) The change in properties of the seismic isolation devices caused by repeated
deformation or temperature rise, due to the wind response must not be excessive. If
the response values are affected by the change in properties, this should be
considered.
(iv) The cumulative damage to the seismic isolation devices due to fatigue, as a result of
the strong winds that occur comparatively frequently during the service life and
earthquakes must be considered.
(v) The residual displacement after strong winds must be evaluated.
References
2) Yasuo Takenaka, Masao Iizuka, Masayasu Suzuki, Kazuhide Yoshikawa, Kazuhiko Yoshikawa：Convenient Method
For Estimating Base-Isolated Tall Buildings Responses to Strong Winds Incorporating Creep Characteristics of
Lead-Plugged Rubber Bearing, Journal of Structural and Construction Engineering, AIJ, No.561, pp.89-94, 2002.11
4.5.4 Other Matters to Be Investigated
(1) Investigation of habitability
In high-rise buildings with a large aspect ratio, it is necessary not only to investigate the
structural safety with respect to large earthquakes and strong winds, but also the habitability
with respect to normal seasonal winds. Particularly, the natural period of seismically
isolated structures is long compared with normal buildings of the same size, and the mass
of the building tends to be lighter. So, it is necessary to treat habitability more carefully.
Habitability is investigated by a performance evaluation curve (see Fig. 4.5.2) from

"Guidelines for the Evaluation of Habitability to Building Vibration"1) (hereafter
"Guidelines for the Evaluation of Habitability") that uses the maximum response
acceleration of the horizontal vibrations due to a wind load having a return period of 1 year.
The specific investigation procedure is shown in the following (1) to (4)2).
@Seismicisolation
@Seismicisolation
50
Responce Acceleration (0-p) (cm/s2)
10
5 Frequency
H-90 (Hz) 0.1≦f ＜1.5 1.5≦f ＜2.5 2.5≦f ≦10
Revel
H-70 -0.5 0.8
H-10 1.17 f 0.96 0.462 f
H-50
-0.5 0.8
H-30 H-30 1.67 f 1.37 0.658 f
-0.5 0.8
H-10 H-50 2.15 f 1.76 0.846 f
1
-0.5 0.8
H-70 2.76 f 2.26 1.087 f
-0.5 0.8
H-90 3.94 f 3.22 1.548 f
0.5
0.1 0.5 1 5
Frequency (Hz)
Fig. 4.5.2 Performance evaluation curve of horizontal vibration of a building induced by

wind explained in the AIJ Guidelines 1)
(1) The structural engineer sets design criteria for habitability as targets, based on a
comprehensive judgment of the building owner’s needs, and cost. The criteria are
quantified as a probability of the perception of vibration.
(2) An appropriate wind direction is set in accordance with the plan shape of the building,
and the maximum response accelerations due to translational vibrations in the horizontal
direction are estimated for both the along-wind and the across-wind direction.
(3) Sets of the primary natural frequency in each direction and the maximum response
accelerations obtained in (2) are located on the performance evaluation curve, and the
probability of perception of vibrations is obtained.
(4) It is checked whether the obtained probability of perception of vibrations satisfies the
criteria. If the criteria are not satisfied, some measure to reduce the response
acceleration is taken.
(a) Setting Criteria
When evaluating the habitability by using the "Guidelines for the Evaluation of
Habitability”, the criteria are prescribed as probabilities of the perception of vibration.
To set the criteria, therefore, it is necessary to connect the probability of the perception
of vibration and the values for evaluation as habitability performance (good, poor,
normal). In this section, the probability of the perception of vibration is connected with
the values for evaluation to determine the criteria, and the results of past research are
introduced.
Figure 4.5.3 shows an example of setting performance grades based on perception

surveys2). In reference 2, the habitability performance with respect to horizontal
vibrations of a building is classified into four ranks, ranked 1 (tolerable) to rank 4 (very
good) based on the perception survey of occupants. Rank 3 (good) is the recommended
level for occupants.
@Seismicisolation
@Seismicisolation
According to reference 2, the probability of perception of vibration in the range 30% to

60% is a guide to setting the criteria. However, the target performance for each
individual building is determined by the structural engineer in accordance with the
wishes of the building owner, and reference 2 should not affect this freedom to decide.
Actually, many buildings are constructed with a probability of perception of vibration in
excess of 90% (H-90), so it is desirable to set the criteria with a balance on cost and
performance.
50
Habitability grade of horizontal vibration of building induced by wind
(in case of a residence and an office)
Maxmum Acceleration (cm/s2)
in case of a residence and an office

10 Frequency Perception
0.1≦f ＜1.5 1.5≦f ＜2.5 2.5≦f ≦30
(Hz) Probability
Rank4 (preferable level)
H-90
5 Border of
30％ 1.67 f
-0.5
1.37 0.658 f
0.8
Rank4 and 3
Rank1 H-70
(undesirable) Rank3 (desirable level)
H-50
Border of -0.5 0.8
Rank2 (standard) H-30
Rank3 and 2
60％ 2.43 f 1.98 0.952 f
Rank3 (desirable) H-10 Rank2 (standard level)
Border of -0.5 0.8
Rank4 (preferable) 80％ 3.22 f 2.63 1.263 f
1 Rank2 and 1
Rank1 (undesirable level)
Limit of -0.5 0.8
about 97％ 5.29 f 4.32 2.077 f
0.5 Rank1
0.1 0.5 1 5
Frequency (Hz)
Fig. 4.5.3 Example of habitability grade of horizontal vibration of building based on residents’
consciousness 2)
(b) Probability of Perception of Vibration
The probability of perception of vibration is a numerical index representing the

percentage of people that would perceive a vibration in one location. If the primary
natural frequency of the building is f1 [Hz], and the maximum response acceleration is
amax [cm/s2], the probability of perception of vibration Hpp [%] shown in “Guidelines for
the Evaluation of Habitability” as the index for the evaluation of habitability
performance is given by equation (4.5.1), which uses an error function erf.
−0.5
 2.15 × f1 ： 0.1 ≦ f1 < 1.5
  a   
H PP = 1 + erf  ln max  0.668  × 50 , C =  1.76 ： 1.5 ≦ f1 ≦ 2.5
   C    0.8
 0.846 × f1 ： 2.5 < f1 < 10 (4.5.1)
The habitability evaluation curve of Fig. 4.5.2 shown in “Guidelines for the Evaluation
of Habitability” is a graph corresponding to Hpp = 10, 30, 50, 70, 90 [%] as a guide to
the probability of the perception of vibration over the range of the maximum response
acceleration of f1 = 0.1 Hz to 5 Hz. In practice, the results of the evaluation of the
habitability (probability of perception of vibrations) are frequently confirmed by
plotting the maximum response acceleration of the building on a graph such as Fig.
4.5.2. However, if the criterion is set to a level other than that shown on the performance
evaluation curve, for example, Hpp = 60 [%] or 80 [%], or if it is preferable to indicate a
specific probability of the perception of vibration, it is necessary to also calculate the
results by using equation (4.5.1).
@Seismicisolation
@Seismicisolation
(c) Estimating the Response Acceleration
If the building has a rectangular shape in plan, the maximum response acceleration in
the along-wind and across-wind direction can be estimated in accordance with
"Recommendations for Loads on Buildings"3) (hereafter the "Recommendation for
Loads") with the natural frequency of the building, the natural vibration modes, and the
damping ratio as parameters. Here, the natural frequency and the natural vibration
modes of a seismically isolated structure are determined based on the results of an
eigenvalue analysis for the elastic behavior of the seismic isolation level. Also, the
damping ratio for small amplitude levels can be considered to be the primary damping
ratio of the superstructure, as determined by referring to "Damping in Buildings"4).
If the building has a plan shape other than a rectangular shape, response prediction using
an approximate calculation from "Recommendation for Loads” is not accurate, so it is
desirable to estimate the response acceleration by using a wind tunnel test or
computational fluid dynamics (CFD).
(d) When the Criteria Are Not Satisfied
When a result does not satisfy the criteria obtained in the investigation of habitability
with respect to wind, it is necessary to i) improve the estimation accuracy of the
response acceleration, or ii) take a measure to reduce the response acceleration.
For i), if the response has been predicted by using the simplified methods of
"Recommendation for Loads,” a re-analysis is performed by using values of the wind
load coefficient from wind tunnel tests or computational fluid dynamics (CFD). Also,
constructing a dynamic analysis model of the building, and carrying out a time history
response analysis by using the wind loads obtained from wind tunnel tests, can be
considered.
For ii), it is considered that the most effective method is to apply damping to the
building. The response acceleration due to wind load is inversely proportional to the
square root of the damping ratio. For example, if 2% damping is applied by installing
dampers in a building with a 1% damping ratio, the magnitude will be
1% ( 1% + 2% ) = 0.58 times the response acceleration before installation of the
dampers. Also, if possible, the wind load coefficient can be reduced by modifying the
shape of the building in plan, and modifying the vibration characteristics by adjusting
the stiffness or mass is also considered to be effective.
If the criteria are severe, one option is to review the criteria by considering the balance
between cost and performance.
(2) Investigation of unstable aerodynamic vibrations
When strong winds act on a super high-rise building with a large aspect ratio that is light
and comparatively flexible and has low damping, additional aerodynamic forces are
produced in accordance with the vibration of the building. These aerodynamic forces act as
@Seismicisolation
@Seismicisolation
negative damping, and unstable aerodynamic vibrations that amplify the vibrations could
possibly be generated4). If the possibility of generation of unstable aerodynamic vibrations
is to be checked at the design stage, it is necessary to use appropriate wind tunnel tests. For
normal buildings that have a fixed base, according to the research that has been carried out
to date, the conditions for generating unstable aerodynamic vibrations in the building in
accordance with the "Recommendation for Loads"3) can be conveniently checked. On the
other hand, little research has been carried out on unstable aerodynamic vibrations of
seismically isolated buildings, so it is not clear whether it is valid to investigate unstable
aerodynamic vibrations in seismically isolated structures by using the "Recommendation
for Loads"3) in the same way as for normal buildings. Therefore, when designing super
high-rise seismically isolated buildings, it is necessary to perform a careful investigation
incorporating the latest research results5).
References
1) Architectural institute of Japan：Guidelines for the Evaluation of Habitability to Building Vibration, 2004
2) Chizuko Noda, Takashige Ishikawa：Habitability Grade of Horizontal Vibration Based on Residents’ Consciousness,
Journal of Environmental Engineering, AIJ, No.648, pp.131～137，2010.2
3) Architectural institute of Japan：Recommendations for Loads and Buildings, 2004.9
4) Architectural institute of Japan：Damping in Buildings, 2000
5) Junji Katagiri, Takeshi Ohkuma, Hisa Marukawa, Hachinori Yasui：Unstable Aerodynamic Response of Base- isolated
High-rise Buildings, Journal of Structural and Construction Engineering, AIJ, No.681, pp.1637～1644，2012.11
@Seismicisolation
@Seismicisolation
Section 4.6 Points to Note Regarding Special Seismically Isolated Structures
4.6.1 Seismically Isolated Detached Houses
The first seismically isolated buildings in Japan were reinforced concrete (RC) detached
houses constructed in 1983. Thereafter, for about 10 years, multistory buildings became the
mainstream for seismic isolation, and applications for detached houses were limited to a few
constructed for research purposes. However, with the impetus from the 1995 Hyogo-ken
Nambu Earthquake, house builders carried out technical developments on special seismic
isolation devices and seismically isolated structures for the detached housing1) constructed
after the year 2000. Furthermore, because their performances were evaluated as beneficial in
several major earthquakes2), 3), this type of building spread rapidly. At the end of 2010, the
number of seismically isolated detached houses was estimated to be about 4500.
When the structural stability of seismically isolated buildings is confirmed by using structural
calculations, as a rule the response of the seismic isolation level is within limiting values, and
the superstructure is evaluated integrally by the behavior of a single mass. However, in the
case of small-scale timber or steel structures (4th category buildings in Japanese building
standard law), structural calculations are not required in accordance with the notification. A
two-story timber building with a floor area 500 m² or less, or a one-story building
constructed from a material other than timber with a floor area of 200 m² or less, is classified
as a 4th category building.
For the structural design of detached seismic isolation houses, the notification method is
easier for the scale of the building, but depending on the judgment of the structural engineer,
calculation methods based on time history response analysis or limit state design can also be
selected. Regardless of the design method, it is desirable that the structural design be
performed with a good balance between isolation systems, superstructures and foundation
structures so that the characteristics of the seismically isolated structures can be well used.
(1) Superstructure
For seismically isolated detached housing, various structural systems can be used for the
superstructure, such as reinforced concrete, structural steel, and timber. Timber structures
include framed structures, stud wall and framework assembly construction, panel
construction, and log construction. Furthermore, for each of these construction methods,
besides the conventional method of in-situ construction, prefabrications methods for
industrialized detached houses are available. For designing seismically isolated detached
houses, therefore, it is important to understand the mechanical properties of the
superstructure. Normally, the superstructure is small-scale and light, so the cross-sections of
the structural members are small, and the stiffness is small in the in-plane direction,
out-of-plane direction, and vertical directions. When this superstructure is separated from
the foundation structure as an independent structure, various types of forces are transmitted
@Seismicisolation
@Seismicisolation
to the seismic isolation devices, so a stiff floor plate for seismic isolation (hereafter
platform) is normally provided as a base for the superstructure; so, the superstructure is not
directly installed on these seismic isolation devices.
(2) Platform (floor plate for seismic isolation)
In the case of seismically isolated detached housing, the platform is normally composed of
a steel frame structure, or a laminated timber structure with metal reinforcement at the
joints or a thick reinforced concrete floor plate. The deformation of the platform plus the
overall building inter-story drift angle consists of deformation of the superstructure. So,
confirming the necessary bending stiffness and strength in the out-of-plane direction needs
to be carried out by structural calculation. The frame structures can be replaced with lattice
beam models, and the long-term loads and seismic loads from the superstructure and the
lower structure should be applied as external forces, and the deformation of the frame
structure should be checked. It is desirable to place the isolators installed under the platform
(seismic isolation level) directly below the columns of the superstructure; but the isolators
on the outer periphery are frequently placed to the inside of the center of the columns
because of external appearance and waterproof details, and the external corners become
eccentric in two directions. The eccentricity of the center of the columns and the center of
the bearing members produces bending moments, shear forces, and deformation on the
platform structure, so it is necessary to provide appropriate strength and stiffness.
(3) Seismic isolation devices
Because the building mass of seismically isolated detached housing is small, the concepts
for using isolators frequently differ from those of multistory buildings. The combinations
vary depending on the structural engineer, but many examples of combinations use sliding
type isolators, followed by roller type and rubber bearings, in that order. Frequently, rubber
bearings are used as the restoring force member (not allowed to support a vertical load),
and systems with a gravity self-centering-force are also used. In the case of damping
members, many examples use sliding bearings (friction), but rubber bearings (high
damping) or dampers (oil, rotary damping tube) are also used either on their own or in
combination4).
Structural safety against strong winds is an important item for seismically isolated detached
houses, as it is for normal houses, and so it is necessary to provide horizontal restraint
device (displacement control mechanism) in the seismic isolation level5).
(4) Wind design
In small-scale seismically isolated buildings, such as seismically isolated detached houses,

the shear force in the seismic isolation level due to wind pressure forces in extremely rarely
occurring strong winds (1.6 times that of the standard wind velocity) frequently exceeds the
seismic shear forces in the seismic isolation level. If a device for restraining the seismic
isolation level against wind is provided as a countermeasure against the wind, it is
necessary to check the safety in terms of structural load resistance. Rope systems or
counterfort wall systems can be employed as devices for restraints against the wind. In the
design of devices for restraint against the wind, in addition to carrying out structural
calculations using the specified design strength for specified building materials, materials
meeting the standard strength stated in an official specification such as JIS should be used.
@Seismicisolation
@Seismicisolation
(5) Design of clearances
Many detached houses are built on small narrow sites that have many external facilities,
such as carports for parking cars and sheds for service equipment. Also, gates and fences
may be provided on the site at the boundary line with an adjacent site. Furthermore, not
only large trees but also growing trees may be present. It is essential to design so that the
impact of the house on these amenities during a large earthquake is prevented, housing
eaves do not pass over the boundary, and safety space is ensured so that people do not
become trapped.
4.6.2 High-Rise Seismic Isolation
Every year several tens of high-rise seismically isolated buildings exceeding heights of 100
m are being constructed. Several buildings among them have an aspect ratio (height direction
divided by short side direction) exceeding 4.0; in many cases, the technical problems of these
structures are solved by increasing the natural period of the seismic isolation level.
Furthermore, to reduce the difficulty of designing this type of high-rise seismically isolated
building, mechanisms for controlling and reducing the uplift forces on the seismic isolation
devices are being actively developed.
Specifically, the following issues require attention in terms of the design of high-rise
1) If the design shear force of the superstructure approaches the maximum response shear
force of the seismic isolation level, the shear force can be easily transmitted to the
superstructure due to the increase in the response shear force of the seismic isolation
level, which can easily cause plasticity of the superstructure.
2) If the response shear force of the superstructure increases, excessive tensile forces can
easily be generated that exceed the allowable values for the seismic isolation devices
due to the increase in the overturning moment and the vertical seismic motions.
3) If the yield shear force of the seismic isolation level is set smaller to suppress the
response shear force of the superstructure, the seismic isolation effect can be easily
obtained for the superstructure; however, cumulative damage or an increase in
deformation due to long period long duration seismic motions or external wind forces
can cause problems for these structures.
@Seismicisolation
@Seismicisolation
4.6.3 Intermediate-Level Seismic Isolation
Buildings in which the seismic isolation level is provided at an intermediate level of the
building are normally referred to as intermediate-level seismically isolated structures. Here,
structures that disregard the seismic response amplification of the lower structure from the
foundation to the seismic isolation level are referred to as base isolated structures. Also,
structures that must consider the seismic response amplification of the lower structure are
referred to as intermediate-level seismically isolated structures (Fig. 4.6.1). In
intermediate-level seismically isolated structures, connections with the ground surface, as in
base isolated structures, have no architectural restrictions, so they have the merit of
increasing the degrees of freedom of architectural design. Thus, it is possible to connect
different structural forms via the seismic isolation level6). In addition, because the seismic
performance of the building as a whole improves, including the structure below the seismic
isolation level, it is widely adopted for seismic retrofit7) and urban redevelopment projects. In
other words, intermediate-level seismic isolation is one of the most favorable methods for
improving the current extremely high-complexity and high-density urban environment.
Because of the seismic isolation level, intermediate-level seismic isolation is a highly

stratified structural system, so this system has unique vibration properties and seismic
response properties that vary depending on the location of the seismic isolation level8), 9). In
particular, typical characteristics are the mass damper effect6), which reduces the response of
the lower structure by the mass damper effect of the structure above the seismic isolation
level, and the mode coupling effect10), which produces amplification of the superstructure
response due to the coupling effect between the vibration modes of the superstructure and the
lower structure. These phenomena have been confirmed for cases in which the seismic
isolation level has been provided in an intermediate story or a high story of high-rise
buildings, and are strongly affected by the mass ratio and period ratio of the superstructure
and the lower structure that sandwich the seismic isolation level11). Furthermore, compared
with base isolated buildings of the same scale, the response reduction effect of the dampers in
intermediate-level seismic isolation is further reduced. Figure 4.6.2 shows the relation
between the maximum shear force reduction ratio a/ae (a: maximum shear force coefficient
of the seismic isolation level, ae: maximum shear force coefficient of the seismic isolation
level without dampers) of the seismic isolation level due to the damper effect coefficient ae
(corresponding to the amount of damping) and the maximum displacement reduction ratio
δm/δe (δm: maximum displacement of the seismic isolation level, δe: maximum displacement
of the seismic isolation level without dampers) in accordance with the mass ratio µ (=
superstructure mass / lower structure mass), which represents the position of the seismic
isolation level12). The higher the level of the seismic isolation level, the lower the mass ratio
is, so the response reduction effect due to the dampers is reduced. This is because the
dampers provided in the seismic isolation level must absorb the seismic energy not only of
@Seismicisolation
@Seismicisolation
the structure above the seismic isolation level, but also that of the lower structure.
Upper
structure
Upper
structure
Upper Seismic
structure Lower
isolation structure
layer Lower
structure
Lower
structure
(a)Base isolation (b)Intermediate-level seismic isolation

(The seismic response amplification (The seismic response amplification
due to lower structure is ignorable) due to lower structure is not ignorable)
Fig. 4.6.1 Base isolation and intermediate-level isolation
ae = 8(a s / a e )
(Intermediate
-level isolation)
(Base isolation)
Fig. 4.6.2 Response reduction effect of the seismic isolation level
To determine these vibration and seismic response characteristics, a simple response
@Seismicisolation
@Seismicisolation
prediction method that does not require time history response analysis is effective12), 13). It is
considered that a more rational design method is to simply evaluate the response reduction
effect of the dampers of the seismic isolation level and the response amplification due to the
mode coupling action, set parameters that are effective for response control, predict the
response, and based on this, perform a detailed analysis using the time history response
analysis.
The seismic performance of intermediate-level seismically isolated structures varies greatly

depending on the position of the seismic isolation level, but most structures are
multi-purpose buildings (mixture of housing, commercial facilities, and offices). So, in many
cases, the position of the seismic isolation level is determined by the building use purpose.
The advantages of seismically isolated structures in intermediate-level seismic isolation
include excellent structural safety and functional integrity after an earthquake. In order to
sufficiently exhibit these advantages, multiple measures must be taken at the preliminary
design stage (including fireproof coating of the seismic isolation devices, fire prevention
compartmentalization of the seismic isolation level, EV shafts, and stairs that pass through
the seismic isolation level).
4.6.4 Column-Head Seismic Isolation, Pile-Head Seismic Isolation
Column-head seismic isolation is a structural method in which seismic isolation devices are
placed on the first-floor column heads. Figure 4.6.3 shows an example of column-head
seismic isolation. Column-head seismic isolation has the advantage that it is possible to
effectively utilize seismic isolation level as multi-purpose spaces, such as car parks or pilotis.
The seismic isolation effect of real buildings by applying column-head seismic isolation was
confirmed in the 2011 Great East Japan Earthquake14). Pile-head seismic isolation is a
structural system where the seismic isolation devices are placed on pile heads, and the
foundation beams connecting the pile heads are either simplified or omitted15). Figure 4.6.4
shows an example of pile-head seismic isolation. Compared with conventional seismically
isolated structures in which the seismic isolation devices are placed below the foundation of
the superstructure, both column-head seismic isolation and pile-head seismic isolation
require little underground construction, and so the quantity of excavation is reduced and the
construction time is shortened. However, because the story that includes the seismic isolation
devices is space in use, it is necessary to take measures, such as fireproof coating of the
seismic isolation devices, to ensure the necessary fire resistance performance of the room
space. Also, since foundation slabs restrain the pile heads, it is desirable to take measures so
that the horizontal displacement of the pile head of each pile is uniformly restrained. As is
similar to conventional seismically isolated structures, it is necessary to sufficiently consider
the ability to follow the displacements at the layers where seismic isolation displacement is
@Seismicisolation
@Seismicisolation
allowed.
2nd floor
bottom floor
fire-resistive covering
seismic isolator
seismic isolator
slab
pillar
pile
1st floor
Fig 4.6.3 Column-head isolation system Fig 4.6.4 Pile-head isolation system
In column-head seismic isolation and pile-head seismic isolation, the double foundation
conventionally adopted for seismically isolated structures is omitted, so it is not possible in
some cases to assume that the seismic isolation device connections are rigidly fixed. In these
cases, rotation is produced at the seismic isolation device connections. If rotation is produced
at the ends of the seismic isolation devices, it is necessary to appropriately evaluate the effect
of the rotation on the horizontal properties of the seismic isolation devices. The effect of
rotation at the ends of seismic isolation devices on the horizontal properties has been reported
in recent years. Examples of applying a constant rotational angle to the ends of rubber
bearings and experimentally measuring the horizontal properties, as well as examples of the
horizontal loading tests coupling the rotational angle at the ends with the horizontal
displacement have been reported16) - 23). When a rotation is produced at the end of a rubber
bearing, an additional shear force is produced by the component of force of the vertical force
due to the tilting of the axial center, so the apparent horizontal stiffness (shear stiffness) is
reduced compared with the case where no rotation is produced at the ends. The relation
between the relative rotation angle between flanges (constant rotation angle) and the rubber
bearing horizontal stiffness20) is shown in Fig. 4.6.5.
Rate of change of horizontal stiffness
Rate of change of horizontal stiffness
1.2 1.2
1.0 1.0
0.8 0.8
from 0 (rad)
from 0 (rad)
Natural Rubber Bearing Natural Rubber Bearing

0.6 φ500 S2=4 0.6 φ500 S2=4
compression compression
0.4 1N/mm2 γ=50% 0.4 10N/mm2 γ=50%
γ=100% γ=100%
0.2 γ=200% 0.2 γ=200%
γ=230% γ=230%
0.0 0.0
0 1/200 1/100 3/200 1/50 0 1/200 1/100 3/200 1/50
Relative rotation(rad) Relative rotation(rad)
Fig 4.6.5 Relation between horizontal stiffness and relative rotation of natural rubber
bearing
In this figure, the vertical axis shows the ratio of the horizontal stiffness relative to a
rotational angle of 0° for various values of shear strain γ. It can be seen that when the
rotational angle increases, or the compressive stress increases, the stiffness changes greatly.
@Seismicisolation
@Seismicisolation
However, at the shear strain of 100% in Fig. 4.6.5, the stiffness ratio of the horizontal
stiffness at the compressive stress of σ = 10 MPa and θ = 1/60 rad is about 0.86. Tests have
also been carried out on test specimens with different secondary shape factors17), 23). The tests
showed that the horizontal stiffness due to the rotation of the ends reduces as the compressive
stress increases, even with different secondary shape factors23). The reduction in the
horizontal stiffness KH due to the effect of rotation (insufficient bending stiffness of the
structure) has been investigated analytically, and the following equation has been proposed24),
25)
.
2
KH 1 + 4γ  σ 
=1− 
σ


（4.6.1）
KH0 1+γ  cr 
where KH0: horizontal stiffness when the compressive stress σ = 0
2
1 H  TR  κ
σ cr = G πS1S 2 H  2
= ζGS1S 2
2 TR   2(1 + 2κS1 G / Eb )
κ
ζ =π
8(1 + 2κS12G / Eb )
S1, S2: primary and secondary shape factors, G: elastic shear modulus of the rubber, TR:
total rubber thickness
H: layer thickness of rubber layers and intermediate steel plates, κ: correction
coefficient, Eb: rubber elastic bulk modulus
E I E r Eb  2 
γ =  rb  K r , Erb = , Er = E0 1 + κS12 
 TR  E r + Eb  3 
E0: rubber Young's modulus

I: second moment of area of the rubber bearing cross-section, Kr: bending stiffness of
the structure
The tests applying a tilt or rotation to rubber bearings20), 23) were carried out up to a maximum
tilt of 0.02 rad. Although the horizontal stiffness is reduced when a rotational angle is applied,
the results obtained do not show large reductions in the damping performance. If the
rotational angle increases in proportion to the shear deformation of a rubber bearing, the
effect becomes larger (such as the reduction in horizontal stiffness) compared with a constant
rotational angle. It is considered that for rotational angles up to 0.01 rad in the experimental
research carried out to date, the angles have no major effect on the hysteretic properties in the
horizontal direction. However, a task for investigation in the future is whether mechanical
@Seismicisolation
@Seismicisolation
properties due to rotation or tilting change with time. Also, as experimentally confirmed, the
bending stiffness of rubber bearings tends to be reduced as the compressive stress or the
horizontal deformation increases.
Horizontal loading tests that apply a rotational angle to sliding bearings have also been
performed. The thicker the elastomer for making the compressive stress uniform, the smaller
the effect of the rotational angle during horizontal load hysteresis26). For pile-head seismic
isolation, a seismic isolation device has been proposed for which the rotational constraint of
the pile head due to the device is extremely small, and by combining a mechanism that
allows a relative rotation angle to be produced at the pile head in series with a sliding bearing,
no P-∆ moment is produced due to the axial load of the superstructure on the pile head15).
4.6.5 Vertical Seismic Isolation
Vertical seismic isolation has been researched in the past as a countermeasure to prevent
vertical vibrations of automobiles and rail vehicles while traveling. In addition,
high-precision equipment, such as semiconductor manufacturing equipment, suspended
foundations for shaking tables, display cases for works of art, and fittings for housing
valuable articles, have been installed on floors of buildings with vertical seismic isolation. In
particular, the number of cases of data centers using seismically isolated floors with the
objective of stabilizing data output has been increasing in recent years. (It is said that the
error rate of data storage increases when subjected to vibrations of 200 to 250 Gal.)
These methods of vertical seismic isolation are referred to as elastic support methods,
because the objects to be protected from vibration are supported by flexible springs, such as
coil springs or leaf springs, anti-vibration rubber, or air springs. Normally, in the elastic
support method, the natural frequency used is 1 to 10 Hz for coil springs or leaf springs, 4 to
15 Hz for anti-vibration rubber, and 0.7 to 3 Hz for air springs. Immediately after the
enactment of the Vibration Regulation Act in 1976, the elastic support method was actively
introduced into factories using production equipment, such as press machines or forging
machines, that generate large impact vibrations. However, the problems caused were still
large, and because the structures were flexibly supported, those subjected to vibration could
easily oscillate, and it was difficult to stably support them. Also, the operability of equipment
suffered because the equipment itself moved, which was not desirable from the point of view
of the machine life.
Around the year 2000, much research and development on buildings was conducted to apply
3-dimensional seismic isolation systems to a fast-breeder reactor (FBR) nuclear power plant
to improve the seismic safety of the internal equipment, and the research results were
@Seismicisolation
@Seismicisolation
summarized27). In the case of normal buildings, in contrast, seismically isolated structures

became widespread for data centers and art galleries, although the seismic isolation was
mainly in the horizontal direction. Currently, effective vertical seismic isolation devices have
not been developed because the vertical accelerations are only about 1/2 of the acceleration
of the horizontal motions, and the effect of horizontal motions on damage to buildings is
greater than that of the vertical motions. Also, a system to suppress rocking motions in a
building was a technical task for seismic isolation in the vertical direction. In particular, to
seismically isolate a whole building in the vertical direction, it is necessary to accurately
determine the mass of the building to be supported, and there was the problem of how to
maintain balance at the building height level. When coil springs or leaf springs are used, it is
difficult to restore and adjust the building height level when the building mass differs greatly
from that estimated. On the other hand, when air springs are used, the air pressure is adjusted
by using a compressor in accordance with the applied load, so it is possible to adjust the
building height level. To reduce the perceived feeling of discomfort when people move
during daily use of the building, it is necessary to use frictional damping in combination with
a trigger device for vibrations excited by foot traffic.
In March 2011, a building was constructed28)- 30) with a 3-dimensional seismic isolation
system that suppresses rocking motions while allowing vertical motions. The system uses a
combination of rubber bearings and air springs and an oil damper, in which two oil tanks are
connected by a cross pipe. This building was the first example of a building seismically
isolated in both the horizontal and vertical directions.
4.6.6 Retrofit
The seismic isolation technique is one method of seismic retrofit for existing aged buildings
or brick buildings, if their seismic performance is insufficient. By adopting seismic isolation
retrofit, the external forces acting on the superstructure during an earthquake are greatly
reduced, and in some cases the existing parts can obtain the required seismic resistance even
without strengthening. Therefore, seismic isolation retrofit is being applied to buildings, such
as disaster prevention centers, that must maintain their function during and after an
earthquake. Furthermore, for seismic retrofit measures, the effect on the external appearance
of a building of seismic isolation is extremely small, so it is being applied to the conservation
of historical buildings. A major feature of seismic isolation retrofit is that during the retrofit
construction, the building can continue to be used.
Normally, the seismic isolation level is based below the foundations of an existing building;
but, if part of the existing building is suitable for the seismic isolation level, that part can also
be effectively used. Figure 4.6.6 shows the normal construction sequence of installing the
@Seismicisolation
@Seismicisolation
seismic isolation devices below existing foundations. First, the existing concrete of the floor
on the lowermost level is removed, and excavation is carried out around the foundations.
Next, existing foundations are reinforced, and the building is temporarily supported by steel
tubular piles that have been pressed in controlling the load resistance and settlement by using
the dead load of the building as the reaction force. Then, a foundation slab is constructed, and
the seismic isolation devices are installed and integrated with the existing foundation. Finally,
while the whole mass of the building is supported temporarily by the steel tubular piles, the
mass of the building that had been transferred onto the steel tubular piles is transferred onto
the newly installed seismic isolation devices. In this way, it is important to plan the
construction sequence and construction safety at each stage of the work, and to sufficiently
consider countermeasures against earthquakes during construction. It is also necessary to
comprehensively plan by paying attention to the external forces set during the original design,
and the strength and vibration properties of the existing structure, in order to obtain the
seismic isolation effect that will satisfy the target seismic performance.
1. Underground excavation and Received provisional 2. Foundation slab concrete placement
3. Seismic isolation system installation and 4. Complete

Up and down foundation concrete placement
Fig. 4.6.6 General construction procedure to install isolation devices under an existing
foundation
4.6.7 Seismic Isolation on Weak Soils
Most seismically isolated buildings have been constructed on ground conditions of

category-1 or category- 2. However, as seismically isolated structures spread, some
construction may take place on weak soils, such as category-3 ground or ground having the
potential for liquefaction. The properties of the input seismic motions are related to the
@Seismicisolation
@Seismicisolation
earthquake source properties, properties of the transmission path, and the soil properties near
the building, and at the same time they are affected by the interaction between the building
and the foundations. The natural periods of seismically isolated structures are relatively long,
so it is known that the radiation damping effect cannot be expected. In addition, in the case of
weak ground, the periods of the seismic motions are lengthened and become closer to the
seismic isolation period, so the effect of the building-ground interaction on the input ground
motion properties increases and cannot be neglected. By considering the response of
seismically isolated structures on weak ground based on conventional knowledge, the
following general trends are indicated.
1) On weak ground, the natural periods of the soil–foundation system are increased, so the
periods of the seismic motion input to the superstructure are also increased. Therefore,
resonance with the period of the seismically isolated building becomes easier, and the
response of the seismic isolation level is larger than the response on hard ground.
Normally, the response acceleration ratio of a seismically isolated structure relative to a
normal seismic resistant structure on weak ground tends to increase, so it becomes
difficult to exhibit the seismic isolation effect.
2) The ground deformation on weak ground increases, so the pile stresses due to forced
deformation from the ground increase. On the other hand, in the case of seismically
isolated structures, inertia forces from the superstructure acting on the piles are reduced
in comparison with normal structures. Therefore, the pile stresses due to ground
deformation become dominant on weak ground, and so the pile stress reduction effect
due to the relative reduction in inertia forces is reduced.
3) If ground liquefaction occurs in a major earthquake, the ground deformation is further

increased, so as a rule it is desirable that improvement of the ground condition is carried
out in strata where liquefaction can occur in order to reduce the ground deformation.
4) To maintain the seismic isolation performance on weak ground, it is effective to reduce

the stiffness of the seismic isolation level so that the seismic isolation period is longer
than the dominant period of the ground during an earthquake. However, if the period is
increased excessively, it is necessary to be aware that the deformation of the seismic
isolation level could be excessively increased. On alluvial plains, long period seismic
motions are affected by the deep ground structure, so this must also be investigated.
5) The horizontal stiffness of the seismic isolation level of seismically isolated structures is
low, and so it is possible that horizontal displacements could be produced if the building
is inclined by differential settlement. Therefore, it is necessary to ensure that large
differential settlement does not occur, even if the ground or the pile springs change due
to the effect of liquefaction.
6) If liquefaction of the ground occurs, it is known that the acceleration at the ground
surface is reduced, so the assumption that liquefaction will occur is not necessarily on
the safer side when setting the input seismic motions. Therefore, it is essential to
consider the seismic motions where it is assumed that liquefaction does not occur.
@Seismicisolation
@Seismicisolation
References
1) Shigeru Hirano, Tatsuya Miyake et al.：A Study on Estimation Method of Seismic Response Behavior of Base-Isolated
Wooden House Based on Full Scale Vibration Tests, Journal of Structural and Construction Engineering, AIJ, No.529,
pp.65-72, 2000.3
2) Takanori Oyokawa, Shigeru Hirano et al.：A Research on Seismic Response of Base-isolated Wood Houses in Site
Earthquakes, Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan, pp.849-852, 2009.8
3) Kazuyuki Matsumoto, Shigeru Hirano：A Research on Seismic Response of Base-Isolated Wood Housees in Actual
Earthquake (Aug. 11,2009) Paet.3: Comparison of The Measured and Analyzed Responses，Summaries of Technical
Papers of Annual Meeting, Architectural Institute of Japan，2010.9
4) Shinichi Kiriyama, Shinji Nakata, Tsutomu Hanai, Nobuo Fukuwa ： Comparison of Seismic Performance of
Base-Isolated House with Various Devices, Journal of Structural Engineering, AIJ, Vol.50B, pp.561-574，2004.3
5) Takehiro Takahashi, Kazunori Anahara, Yoshihide Fukahori：Experimentation and Simulation of Seismic Impact
Generated in a Base-Isolated House by a Horizontal Displacement Contro, Journal of Structural and Construction
6) Katsuhide Murakami, Haruyuki Kitamura, Hitoshi Ozaki, Tomohiko Yamanishi：Design of a Building with Seismic
Isolation System at the Mid-story, Journal of Technology and Design, AIJ, No.7, pp.51-56, 1999.2
7) Mitsuo Kuroda, Ishirou Narita, Shouji Tanaka, Souichi Kawamura, Keiji Ogura, Akihiko Yajima：Seismic Isolation
Retrofit At Mid-story and Base: Planning and Outline of Works, Journal of Technology and Design, AIJ, No.5, pp.18-22,
1997.12
8) Keiji Ogura, Masaharu Takayama, Osamu Tsujita, Yuichi Kimura, Akira Wada：Seismic Response of Mid-story Isolated
Buildings, Journal of Structural and Construction Engineering, AIJ, No.516, pp.99-104, 1999.2
9) Tadamichi Yamashita, Youchi Mukai, Yutaka Inoue：Evaluation and Synthesis for A Seismic Responses of Structural
System with Inter-story Isolation Devices, Journal of Structural and Construction Engineering, AIJ, No.591, pp.35-42,
2005.5
10) Masahito Kobayashi, Yasukazu Izawa, Tadaki Koh：The Prediction Method of Earthquake Responses on Mid-story
Isolated System Considering Modal Coupling Effect, Journal of Structural and Construction Engineering, AIJ, No.572,
pp.73-80, 2003.10
11) Masahito Kobayashi, Daisuke Sasaki：Making a Seismic Design Database of Mid-story Isolated Buildings and Structural
Property Evaluation Based on Response Prediction Method, Journal of Technology and Design, AIJ, No.29, pp.65-70,
2009.2
12) Masahito Kobayashi, Tadaki Koh：Earthquake Response Prediction and Rationalization of Dynamic Design of Mid-story
Isolated Buildings, Journal of Structural and Construction Engineering, AIJ, No.592，pp.51-57，2005.6
13) Katsuhide Murakami, Haruyuki Kitamura, Yutaka Matsushima：The Prediction for Seismic Responses of the Two-mass
Model with the Mid-story Isolation System, Journal of Structural and Construction Engineering, AIJ, No.549，pp.51-58，
2001.11
14) Masaaki Saruta, Keiichi Okada, Yutaka Nakamura, Masanobu Hasebe：Observed Response of Seismic Isolated
Buildings during the 2011 off the Pacific Coast of Tohoku Earthquake, Summaries of Technical Papers of Annual
15) Yukihiro Nishimura, Mitsuo Miyazaki, Takayuki Sugawara：Development of Pile-top Isolation System, Summaries of
Technical Papers of Annual Meeting, Architectural Institute of Japan, pp.383-388, 2004.8
@Seismicisolation
@Seismicisolation
16) Takafumi Miyama, Yasumasa Kato, Shinichi Iizuka, Hisayoshi Ishibashi, Seiichiro Yasuda, Kazuo Yagishita, Eiichi Inai,
Satsuya Soda：Static Load Tests of Rubber Bearing for the column Top Isolation System, Summaries of Technical
Papers of Annual Meeting, Architectural Institute of Japan, pp.567-570, 1998.9
17) Mistuo Asano, Shigeo Minewaki：Estimate of Horizontal Stiffness of Laminated Rubber Bearing Concerned with
Flexibility of Installation Element, Journal of Technology and Design, AIJ, pp.57-62, 1999.6
18) Masayuki Yanagi, Yozo Shinozaki, Toru Nakajima, Yoshitaka Muramatsu, Ichiro Nishikawa：Shearing Test of Natural
Rubber Bearing Concerned with Column Top Isolation System, Summaries of Technical Papers of Annual Meeting,
19) Takafumi Miyama：Study on the Force-Deformation Relations of Base-Isolation Rubber Bearings with Forced Rotation
Angles At Their Top and Bottom Ends, Journal of Structural and Construction Engineering, AIJ, No.556, pp.43-50,
2002.6
20) Yoshihisa Kitamura, Masaaki Saruta, Takuya Nishimura, Ken Ishii, Masaru Kikuchi, Takahito Nakamura：Mechanical
Characteristic of Rubber Bearing with Angled Flange Plates, Summaries of Technical Papers of Annual Meeting,
21) Ken Ishii, Masaru Kikuchi, Yoshihisa Kitamura, Takahito Nakamura：A New Analytical Model for Seismic Isolation
Bearings Subjects to Eccentric Bending Moments and Shear Deformation, Journal of Structural and Construction
22) Yoshiro Kai, Takuya Nishimura, Osamu Takahashi, Hiromasa Aida, Yoritaka Sasaki, Takafumi Fujita：Development of
Three Dimensional Seismic Isolation Building Part.6: Modeling Technique of Laminated Rubber Bearing under
Rotational Deformation, Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan,
pp.959-960, 2009.8
23) Eiji Takaoka, Makiko Hikita, Yasuo Takenaka：Mechanical Characteristics on Laminated Rubber Bearings Subjected to
End Rotation, Journal of Structural and Construction Engineering, AIJ, No.651, pp.905-912, 2010.5
24) Takahito Nakamura, Mitsuru Miyazaki, Shuhei Kaneko：Mechanical Characteristic of Rubber Bearing with Angled
Flange Plates (Part.3) Experiment for Shear Property, Summaries of Technical Papers of Annual Meeting, Architectural
Institute of Japan, pp.375-376, 2008.9
25) M. Imbimbo, J. M. Kelly : Stability Aspects of Elastomeric Isolators, Earthquake Spectra, Vol.13, No.3, 1997.7
26) Eisaku Nakamura, Eiji Takaoka, Yasuo Takenaka, Toshinao Kawai, Sakae Ueda：Mechanical Characteristics of Base
Isolation Devices Subjected to End Rotations: Part.2: Loading Tests of Sliding Bearings, Summaries of Technical
Papers of Annual Meeting, Architectural Institute of Japan, pp.369-370, 2008.9
27) Junji Suhara, Ryoichiro Matsumoto et al.：Study on Three-dimentional Seismic Isolation System Applied to Advanced
Nuclear Power Plant-Development of Seismic Base Isolation System, Journal of Structural Engineering, AIJ, Vol.52B，
2006.3
28) Tetsuya Tomizawa, Osamu Takahashi, Hiromasa Aida, Junji Suhara, Itaru Kurosawa, Yasuo Tsuyuki, Takafumi Fujita：
Construction of Civil Building Using Three-dimentional Seismic Isolation System, Structural Engineers World
Congress 2011, 2011.4
29) Tetsuya Tomizawa, Osamu Takahashi, Hiromasa Aida, Junji Suhara, Keiichi Okada, Yasuo Tsuyuki, Takafumi Fujita：
Earthquake Observation Record in a Building named “Chisuikan” using Three-dimentional Seismic Base Isolation
System, The 14th Asia Pacific Vibration Conference, Proceedings VOL.3, pp.1365-1371, 2011.12
@Seismicisolation
@Seismicisolation
30) T. Tomizawa, O. Takahashi, H. Aida, J. Suhara, M. Saruta, K.Okada, Y. Tsuyuki, T. Suzuki, T. Fujita：Vibration Test in a
Building named “Chisuikan” using Three-dimentional Seismic Isolation System, 15th World Conference of Earthquake
Engineering, 2012.9
@Seismicisolation
@Seismicisolation

AIJ., Design Recommendations For Seismically Isolated Buildings, 2016 PDF

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

AIJ., Design Recommendations For Seismically Isolated Buildings, 2016 PDF

Încărcat de

Drepturi de autor:

Formate disponibile

20180130salehinalireza

Architectural Institute of Japan

Copyright © 2016 Architectural Institute of Japan. All Rights Reserved.

Preface to the English Version

Architectural Institute of Japan

Architectural Institute of Japan

Seismically isolated structures exhibit sufficient effectiveness of seismic isolation with

Members of Committees Concerned

Research Committee on Structures

Chairman Mitsumasa Midorikawa

Managing Committee on Earthquake and Structural Dynamics

Chairman Yuji Miyamoto

Sub Committee for Seismically Isolated Structures

Chairman Masaru Kikuchi

Working Group for English Edition of the Design Recommendations

Chairman Mineo Takayama

Part I Design Recommendations for Seismically Isolated Buildings

Chapter 1 Outline of Seismically Isolated Buildings 1

Chapter 2 Design – General 5

Chapter 3 Structural Design 10

Chapter 1 Summary of Seismically Isolated Buildings 17

Chapter 2 Seismic Isolation Devices 30

Chapter 3 Design – General 90

Chapter 4 Structural Design 103

Part I Design Recommendations

Part I Design Recommendations for Seismically Isolated Buildings

Chapter 1 Outline of Seismically Isolated Buildings

1.1 Characteristics of Seismically Isolated Structures

(1) Principle of seismically isolated structures

A seismically isolated structure is a structure in which a seismic isolation level is

(2) Performance and effect of seismic isolation in structures

As a result of many years of research, it is now possible to increase the deformation

(3) Design of seismically isolated structures

Design, including architectural design, is a creative activity. It is not merely selecting a

1.2 Outline of Seismic Isolation Devices

(1) Characteristics of devices and performance evaluation

When a seismically isolated structure is subjected to earthquake ground motions, the

1.3 Applicability of Seismically Isolated Structures

Marginal space (clearance) should be provided around a building with a seismically

Chapter 2 Design – General

2.1 Seismic Isolation Performance Considerations

(1) Seismic safety considerations

a) Expansion joint damage

b) Damage to piping in the seismic isolation level

d) Exceeding the deformation limit of the seismic isolation device (isolator)

f) Plastic deformation of the superstructure

The overall intended performance of seismically isolated structures is to ensure safety

acceleration, the effectiveness of a reduction in the horizontal force acting on the

2.2 Basic Scheme

(1) Conditions for the superstructure

In a conventional seismic-resistant structure, plastic deformation is allowed in the frame

(2) Ground conditions

When constructing ordinary buildings, an important point is to select a construction site

(3) Location of the seismic isolation level

(4) Site clearance

Unlike ordinary buildings during an earthquake, a displacement of several tens of

2.3 Arrangement and Selection of Seismic Isolation Devices

(1) Arrangement of seismic isolation devices

The properties of seismic isolation devices include bearing capacity, deformation

(2) Selection of seismic isolation devices

(3) Ensuring performance of seismic isolation devices

2.4 Building Equipment Design

It is presumed that the seismic performance of seismically isolated structures is higher