Seismic Isolation For Architects-Routledge (2017)
Seismic Isolation For Architects-Routledge (2017)
Seismic isolation offers the highest degree of earthquake protection to buildings and their
inhabitants. Modern applications of the technology are less than 50 years old and uptake in
seismically active regions continues to soar.
Seismic Isolation for Architects is a comprehensive introduction to the theory and
practice in this field. Based on the latest research findings and the authors’ extensive experience,
coverage includes the application, effectiveness, benefits, and limitations of seismic isolation,
as well as the architectural form, design aspects, retrofitting, economics, construction, and
maintenance related to this method.
The book is written for an international audience: the authors review codes and practices
from a number of countries and draw on examples from 11 territories including the USA,
Chile, Argentina, Italy, Japan, and New Zealand. Aimed at readers without prior knowledge
of structural engineering, the book provides an accessible, non-technical approach without
using equations or calculations, instead using over 200 drawings, diagrams, and images to
support the text. This book is key reading for students on architecture and civil engineering
courses looking for a clear introduction to seismic isolation, as well as architects and engineers
working in seismically active regions.
Between them, the authors have been actively involved in the design and construction of
three seismically isolated buildings and have visited and studied over 60 more in 11 different
countries.
“I believe that this very well written and documented book will be very useful to architects
worldwide. In fact, it explains the features and advantages of seismic isolation in a very useful
way to architects, by stressing the fact that, thanks to this technique, it is possible not only
to make buildings much safer at limited additional construction costs (if any), but also to
allow for adopting architectural solutions that could never be applicable to conventionally
founded buildings.” – Alessandro Martelli, President of the Italian Association GLIS, Founding
President and present Vice-President of the Anti-Seismic Systems International Society (ASSISi),
and former Professor of "Constructions in Seismic Areas" at the Faculty of Architecture of the
University of Ferrara, Italy
Seismic Isolation
for Architects
Acknowledgements ix
1 Introduction 1
Purpose and timing of the book 1
Current approaches to the seismic design of buildings 2
Introducing seismic isolation 4
Tale of two hospitals 11
Reality check 13
Computer modelling 55
Physical testing 56
Contents
7 Retrofitting 114
Introduction 114
Location of the isolation plane 116
Case studies of retrofitted unreinforced masonry buildings 120
International Library of Children’s Literature, Tokyo 120
Iasi City Hall, Romania 124
Case-studies of retrofitting reinforced concrete frame buildings 126
Rockwell International Building 80 126
Rankine Brown Building 127
National Museum of Western Art, Tokyo 130
China Basin Landing, San Francisco 135
Limitations 137
11 Conclusions 190
Confidence in seismic isolation 190
Benefits of seismic isolation 190
Design freedom and limitations 190
Uncertainties 191
Savings and additional costs of seismic isolation 191
Looking ahead 191
Index 193
Contents
vii
This page intentionally left blank
– ACKNOWLEDGEMENTS –
W E thank all those who have helped us along the journey towards the publishing of this
book. In particular, the photographic and computer technicians at Victoria University
of Wellington Schools of Architecture and Design are thanked for their support. Also thanks
to Duncan McNaughton for drawing all the diagrams, and to the individuals and companies
who have provided images. Unacknowledged images are provided by the authors.
Acknowledgements
ix
This page intentionally left blank
– 1 –
Introduction
M ANY books and articles on the seismic isolation of buildings have been written for
structural engineers. However, these publications are highly technical in nature and
therefore are unsuitable for the vast majority of those who design, construct, own, insure and
inhabit buildings in seismically active regions. The purpose of this book, then, is to introduce
a relatively new game-changing technology to a wide audience concerned about how buildings
and their contents are affected by earthquakes.
A research and writing project such as this, which seeks to address all issues relevant
to seismic isolation, could have been undertaken at any time over the last 35 years. Yet it
would have inevitably left unanswered many important questions relevant to the seismic
isolation of buildings. In particular, how confident can we be in the ability of seismic isolation
to reduce damage during earthquakes, and second, how does this relatively new approach to
protect buildings from earthquakes compare to more conventional ones? Is seismic isolation
really worth adopting?
Within a period of 18 days ending on 11 March 2011, the answers to these questions
suddenly became much clearer. On 22 February the city of Christchurch experienced a
devastating earthquake located only 10 km from its centre, and at a shallow depth of 5 km.
While only one base-isolated building in Christchurch, the Christchurch Women’s hospital,
was tested by the earthquake, so was the entire building stock of Christchurch. Hundreds of
buildings, many designed in accordance with one of the world’s most advanced seismic codes,
survived without collapse. But tragically, for different reasons, most have subsequently been
demolished. This situation raises considerable uncertainty regarding the appropriateness of
modern philosophies of seismic design.
Then, on the 11 March, Japan was struck by the massive Magnitude 9 Tōhoku
Introduction
Earthquake, centred off the east coast. Most of the destruction was caused by the tsunami
that destroyed coastal areas, but earthquake shaking damaged many buildings, and for the
first time, tested hundreds of seismically isolated buildings on an unprecedented scale.
Although subjected to a lesser intensity of shaking than they had been designed for, these
1
2
buildings performed very well. The two earthquakes occurred within days but were located
Introduction
thousands of kilometres apart along the Pacific tectonic plate. Together they both demon-
strated both the effectiveness of seismic isolation as well as deficiencies in current design
approaches to earthquake attack, accentuating the benefits of seismic isolation.
B EFORE examining seismic isolation in detail, a primary aim of this book, we need to first
explain the vulnerability of most buildings around us to earthquakes, and then explore
the strengths and weaknesses of conventional as well as new structural concepts to protect
buildings against earthquakes.
The buildings we inhabit tend to instil a false sense of safety, particularly with respect
to earthquakes. A number of factors converge to soothe our fears about the seismic safety of
a building. We note how buildings stand without any visible signs of distress. They support
their self-weight plus the loads of all the equipment, stock and other contents their occupants
impose. It is rare for buildings to collapse when gravity is the only force acting. We also
correctly understand that the chances of an earthquake occurring during a given period of
time, perhaps today or even during our lifetime, range from infinitesimal to minor. We might
also believe that because the buildings around us have experienced one or more earthquakes
during their lives, they can therefore be assumed safe. People who are more knowledgeable
about building construction might believe that because buildings are built using modern
materials of reinforced concrete or structural steel, they are safe.
Most of these perceptions are only partially correct. The reality is that the risk of a
building collapsing during an earthquake is far higher than one might imagine. The unfor-
tunate fact is that most buildings in the world, unless they have been seismically retrofitted
or built since the early 1980s, are unsafe in earthquakes. In some countries even more modern
buildings are at-risk. The vulnerability of unreinforced masonry buildings is well known. The
undesirable combination of heavy and brittle material and a lack of strong tying between
walls, and floors to walls means these buildings are usually the first to suffer earthquake
damage. But what is less well known is the hazard more modern buildings pose.
It was only after the mid-1970s that codes of practice internationally rectified a critical
flaw in the way buildings were designed against earthquakes. It was then that the importance
of ductility was recognized – that ability of the structure of a building to survive earthquake
over-load without collapse yet suffering damage. Although buildings designed prior to this
period did possess adequate strength, for what would by today’s standards be understood as
small earthquakes, in many cases they will partially or fully collapse when the shaking of a
large earthquake exceeds their strength (Figures 1.1 and 1.2). Rather than their structural
elements such as beams, columns or structural walls bending, stretching or shearing in
a ductile manner during a large earthquake, they break or snap suddenly, in a brittle way.
1.1 Example of structural damage in a building with inadequate shear walls at the garage
level. Loma Prieta earthquake, San Francisco USA, 1989
Source: US Geological Survey
Introduction
1.2 The seriously damaged Olive View Hospital moment frame building. San Fernando
Valley earthquake, California USA, 1971
Source: National Oceanic & Atmospheric Administration (NOAA)
3
4
If such buildings survive a damaging earthquake without at least partial collapse, it will not
Introduction
be because of how the primary structural members were designed and detailed. Surviving
buildings will have likely benefited from, for example, additional strength provided by walls
required for fire protection along site boundaries or to confined stairwells, as well as
structurally desirable characteristics such as symmetry and regularity.
Today, modern codes require that buildings avoid collapse when their primary seismic-
resisting structural systems experience overload during an earthquake. And overload is likely
given how low design load levels are where compared to loads occurring during design
earthquakes. Typical modern earthquake-resistant building design begins with the selection
of one of three common structural systems (Figures 1.3 and 1.4). One system will resist
horizontal earthquake forces in one direction in the plan of the building, and the same or
another structural system will act in the direction at 90 degrees to it. The numbers and sizes
of structural members depend on many factors including the size, height and weight of the
building, and the level of seismicity of the site. With strength in both orthogonal directions,
the structure can cope with earthquake attack from any direction. Where parallel structural
systems, such as shear walls, are separated in plan, so they don’t act along the same line, they
prevent the building from excessive twisting.1 Then the structural engineer calculates the sizes
and details of the structural members so that they are either so strong as not to be overloaded
in an earthquake or, more commonly, designs them to be ductile. That is, even if a member
is damaged it maintains most of its strength and doesn’t break. But, as graphically illustrated
by the Christchurch earthquake, ductility means damage which can lead to demolition.
Even before the limitations of current earthquake-resistant design approaches were so
starkly revealed, engineers had been developing new devices such as dampers, rather like car
shock absorbers, and modifying existing structural systems to survive medium to large
earthquakes with little or no structural damage. The use of dampers and ‘damage-free’ systems
are growing in popularity, but are still uncommon (Figures 1.5 and 1.6). These new systems
certainly represent a big improvement over conventional systems in which structural damage
is inevitable during a large earthquake, but they are unable to offer enhanced protection
for architectural elements such as claddings, interior fit outs and building contents. Seismic
isolation alone protects both structure and architectural elements.
S EISMIC isolation is a term describing any sort of isolation, including by far its most
common form, base isolation. Most buildings are isolated at their bases, but occasionally
the isolation plane is located in a middle storey. Seismic isolation involves the partial
separation of a building from the ground underneath it to reduce the intensity of the
earthquake shaking it experiences. Seismic isolation is like placing the superstructure of a
building on the surface of foundations that have been oiled or greased to make them slippery.
(a) Shear walls
Introduction
1.3 The three most common structural systems for resisting horizontal forces
5
6
Introduction
When the most damaging earthquake waves shake the ground horizontally back and forth,
the superstructure is protected by the sliding interface. If only it were that simple in practice!
Rather than grease, elastomeric and or sliding bearings form the slippery isolation plane. The
dynamic impact of a building superstructure against perimeter retaining walls is avoided by
constructing physical separation gaps (Figure 1.7).
Seismically isolating a building is rather like moving a ship from a dry dock, where it
is resting on the ground, and launching it into water. Once released from its foundations and
floating on elastomeric or other bearings, an isolated structure is far less vulnerable to the
effects of horizontal ground shaking. Usually, most of the energy of an earthquake is manifest
in high frequency vibration. But since the bearings of a seismically isolated building are very
flexible for horizontal movement, the building doesn’t resonate like a conventional fixed-base
building. Rather than experiencing violent high frequency shaking at its base, which is then
amplified up its height, an isolated building enjoys a far gentler ride and an absence of
increased accelerations and movement on its upper floors (Figure 1.8). Its low frequency
response means less earthquake energy is transferred into the superstructure. This reduces
or even eliminates damage to structural members, architectural elements and building
contents as exemplified in the following two well-documented case studies.
1.4b Example of a type of structural system: single-bay moment frame
Introduction
7
8
Introduction
Introduction
Isolated Fixed-base
building building
Introduction
Elastometric
bearing
Ground stationary
Original
position
1.7 The essence of seismic isolation. An isolated building is protected from earthquake
shaking by moving at its base. Movement in a conventional building is absorbed up its
height
Isolated building Conventional building
Acceleration
Time
W E begin by recounting the response of the Ishinomaki Red Cross Hospital to the 2011
Tohoku earthquake. The 6-storey 420 bed hospital was built in 2006 (Figure 1.9). It
is located 4.5 km from the east coast of Japan, just far enough inland to escape the tsunami.
It was the closest hospital to the earthquake epicentre, some 100 km away. Fortunately, the
effect of the earthquake on the hospital was filmed. The video begins with scenes in an office
during the building shaking and then shows the immediate preparation for the arrival of the
injured.2 In spite of its seismic isolation, the intensity of shaking within the building was
considerable. Some workers remained at their desks while others steadied vibrating computer
monitors and mobile equipment. Apart from papers and files falling from desks and shelves,
no damage was observed to the structure or architectural finishes. Thanks to this excellent
seismic performance and the reliable back-up electricity generation, hospital functioning was
unaffected. Nearby, buildings designed to the most recent 1981 Japanese code suffered slight
structural damage, but major damage to building contents, suspended ceilings and external
cladding.3 During the quake the hospital moved to-and-fro relative to the ground beneath it
by up to 260 mm, approximately half the design earthquake movement.4
Christchurch Women’s Hospital was opened in 2005 (Figure 1.10). Consisting of one
main basement level and seven storeys above ground, earthquake forces in the superstructure
are resisted by reinforced concrete moment frames in the longitudinal direction (Figure 1.9).
Introduction
Steel V-braced frames, in the lowest four levels, combine with full-height concrete moment
frames to resist transverse loads. The hospital rests mainly on lead-rubber bearings, designed
to undergo plus or minus 420 mm horizontal displacement during the design earthquake
11
Reality check
A LTHOUGH these two case studies illustrate outstanding seismic performance of seismically
isolated structures as compared to those with conventional fixed-bases, there is no sense
in which seismic isolation is a miracle solution. As noted above, particularly in smaller
earthquakes, the shaking inside a building can be vigorous. Although only a small proportion
of the earthquake energy at the base of the building is transferred into its superstructure,
structure such as walls, moment or cross-braced frames are still required to resist horizontal
inertia forces within the superstructure. Unfortunately for architects and building owners,
the structural members of these systems are of similar dimensions to those in conventional
buildings. This is because seismic design codes allow the design forces of conventional duc-
Introduction
tile buildings to be reduced by factors between four and six in acknowledgement of the
ductility they are designed for, to prevent collapse. Therefore, the structures of seismically
isolated buildings are designed to similar levels of force, but the really important difference is
13
that they remain undamaged during the design earthquake. Conventional structures, on the
14
other hand, will be seriously damaged and, as sadly observed in Christchurch, may be
demolished.
Introduction
Notes
1 Charleson, A.W., 2008, Seismic design for architects: outwitting the quake, Oxford: Elsevier.
2 Japanese Red Cross Society 2011, ‘Ishinomaki Red Cross Hospital – recording of the Great East Japan
Earthquake’, video recording. Available from: www.youtube.com/watch?v=Pc1ZO7YwcWc (accessed: 30
December 2015).
3 Motosaka, M. & Mitsuji, K., 2012, ‘Building damage during the 2011 off the Pacific Coast of Tohoku earthquake’,
Soils and foundations, vol. 52, no. 5, pp. 929–44.
4 Someya, T., 2013, ‘Seismically isolated hospital offers ray of hope in disaster – Ishinomaki Red Cross Hospital’,
Proceedings of the 13th World Conference on Seismic Isolation, Energy Dissipation and Active Vibration
Control of Structures, 24–27 September, Sendai, Japan, p. 3.
5 Gavin, H.P. & Wilkinson, G., 2010, ‘Preliminary observations of the effects of the 2010 Darfield earthquake
on the base-isolated Christchurch Women’s hospital’, Bulletin of the New Zealand Society for Earthquake
Engineering, vol. 43, no. 4, 360–7, p. 362.
6 Whittaker, D., 2013, ‘Recent developments in seismic isolation in New Zealand’, Proceedings of the 13th World
Conference on Seismic Isolation, Energy Dissipation and Active Vibration Control of Structures, 24–7
September, Sendai, Japan, p. 1.
7 Kam, W.Y., Pampanin, S. & Elwood, K., 2010, ‘Seismic performance of reinforced concrete buildings in the
22 February Christchurch (Lyttelton) earthquake’, Bulletin of the New Zealand Society for Earthquake
Engineering, vol. 44, no. 4, pp. 239–77.
8 McIntosh, J.K., Jacques, C., Mitrani-Reiser, J. et al., 2012, ‘The impact of the 22nd February 2011 earthquake
on Christchurch hospital’, Proceedings of the 2012 Technical Conference of the New Zealand Society for
Earthquake Engineering, p. 6.
9 Charleson, A.W. & Allaf, N.J., 2012, ‘Costs of base-isolation and earthquake insurance in New Zealand’,
Proceedings of the 2012 Technical Conference of the New Zealand Society for Earthquake Engineering,
Christchurch, p. 6.
– 2 –
Brief history
A LTHOUGH seismic isolation as practiced today has a short history of less than 50 years,
evidence of far earlier attempts to isolate buildings from ground shaking are being
discovered. Detailing for isolation is found within the vernacular architecture of some
earthquake-prone communities. Some far older traditional buildings have been observed to
outperform modern buildings. For example, some ancient Iranian buildings are founded on
smooth flat stones or on sand layers under load-bearing walls, presumably to facilitate sliding.
A most compelling example from that region consists of houses supported on short lengths
of logs which are layered and oriented in plan to allow rolling in two orthogonal directions.1
These buildings performed well during the 1990 Manjil earthquake. They moved up to 200
mm on their foundations without damage.
The first modern application of seismic isolation was in an elementary school in Skopje,
Yugoslavia, founded on neoprene (artificial rubber) bearings in 1979. However the William
Clayton Building, Wellington, completed in 1981, was the first building whose isolation
system included all the five necessary characteristics of seismic isolation, as explained later
in this chapter (Figure 2.1). Seismic isolation was achieved using lead-rubber bearings, newly
invented by Dr Bill Robinson.
After those initial applications, the uptake internationally of seismic isolation was slow.2
Then the 1995 Kobe, Japan, earthquake struck, causing hundreds of buildings to collapse
and damaging thousands more. Only 10 seismically isolated buildings per year had been
Principles of seismic isolation
constructed before that earthquake; after the earthquake, the rate surged to 150 buildings per
year between 1995 and 2002.3 Renewed interest in and accelerated up-takes of seismic isolation
have been observed in other countries as well more recently. This has been due to a combin-
ation of loss of life during an earthquake, a huge amount of building damage, and the excellent
observed seismic performance of one or more local seismically isolated buildings. This
accelerated application of seismic isolation was also evident in China after the 2008 Wenchuan
earthquake, in Italy after the 2009 Abruzzo earthquake, and in New Zealand after the 2011
15
Christchurch earthquake.
16
Principles of seismic isolation
2.1 The William Clayton Building, New Zealnd, Wellington, 1981. The first example of the
comprehensive application of seismic isolation
• ‘emergency response,
• continued business operations,
• protect contents,
• reduce damage repair cost,
• protect architecture, and
• occupant safety – “peace of mind”.’
2.3 Angels of San Giuliano School, San Giuliano di Puglia, Italy, 2008
17
18
Principles of seismic isolation
2.8 Part of one of 185 residential buildings constructed after the 2009 L’Aquila
earthquake, Italy, as part of the CASE project. Multi-storey housing was built upon
seismically isolated ground floor slabs above basements
Source: FIP Industriale S.p.A., Italy
2.9 The car parking basement of a CASE project building. Double concave curved surface
sliders isolate the ground floor slab from the steel cantilever columns
Source: FIP Industriale S.p.A., Italy
Principles of seismic isolation
2.10 Glendale City Hall, Glendale, USA. The interior courtyard showing one of the
isolators used to retrofit the original 1927 building
Source: USGS
21
22
Principles of seismic isolation
2.11 San Francisco City Hall, San Francisco, USA, retrofitted 1999
buildings of all. Then there are buildings housing manufacturing or other functions for whom
downtime following an earthquake would be disastrous. Data centres, TV studios and even
newspaper print facilities require continuous operation. In some buildings, the focus of
seismic design is protection of the building contents, possibly far more valuable than the
building itself. Laboratories, libraries and semiconductor plants are examples. Some owners
desire to minimize damage and damage repair costs. Perhaps their investment in seismic
isolation is an alternative to paying earthquake insurance premiums and allowing early, if
not immediate, re-occupancy of a building following an earthquake. Government, commercial
and even retail buildings such as shopping centres benefit from this approach. There are also
many buildings of historic significance that have been seismically isolated to protect their
fabric from damage and preserve their architectural qualities. These buildings use seismic
isolation as their primary structural strategy. Several case-studies are presented in Chapter 7.
The final reason for implementing seismic isolation is occupant safety. Although modern
conventionally designed buildings provide a high degree of seismic safety, occupants can be
even safer and certainly experience far less trauma during an earthquake in a seismically
isolated building. This provision of increased personal safety is evidenced in growing numbers
of seismically isolated apartment buildings or condominiums and houses, particularly in
Japan.
Seismic isolation has been applied to both new and existing buildings. Certainly, the
technology is introduced more easily into a new building, but there are now hundreds of
2.12 Oakland City Hall, Oakland, USA, retrofitted 1995
buildings around the world retrofitted by seismic isolation. They include a wide diversity of
both heritage and newer buildings. As well as preserving history in the form of iconic
historic buildings, both building occupants and business continuity are being protected in
modern buildings. In each case the owner is convinced that seismic isolation will achieve one
or more of the six objectives of seismic isolation listed above.
Principles of seismic isolation
This brief overview of the history and diversity of seismically isolated buildings also
needs to acknowledge the seismic isolation of just parts or contents of buildings, rather than
buildings themselves. Many examples may be found where sensitive or fragile equipment is
isolated. Large specialized industrial equipment to smaller computer racks and server cabinets
can all be mounted on their own isolation devices. Even individual sculptures are isolated
within conventionally constructed museums. It appears that there are few, if any, types of
high-value buildings, equipment or contents that cannot benefit from seismic isolation.
23
24
Principles of seismic isolation
Movement capability
The superstructure of a building needs to be isolated from the violent to-and-fro horizontal
shaking of the ground by providing for relative movement between superstructure and
ground. Ideally, the ground and the foundations of an isolated building vibrate back-and-
forth in any direction without transferring kinetic energy into the building. When the ground
moves, the superstructure should remain motionless. In practice it is more complex, but
movement capability is provided by elastomeric bearings from natural or artificial rubber, or
Principles of seismic isolation
sliding surfaces of Teflon and stainless steel, and by complete horizontal separation of the
superstructure from the ground.
Vertical support
In order to allow almost unrestrained horizontal movement, an isolation system must support
25
the weight of the building. Hardware, such as most types of bearings, not only provides
26
Principles of seismic isolation
2.17 Ñuñoa Capital Building, Santiago, Chile. An apartment and office building
Source: Ñuñoa Capital-Inmobiliaria Armas
27
28
Principles of seismic isolation
movement capability but transfers the entire weight of a building to its foundations. Bearings
are normally located under columns and structural walls. They raise the superstructure above
the ground and define discrete points where gravity loads are transferred into the foundations.
Principles of seismic isolation
Re-centring
During earthquake shaking, which is random in extent and direction, a superstructure tends
to drift in a particular direction. Depending on the intensity and duration of an earthquake,
a superstructure could move further than the movement capability of the bearings, damaging
them. A re-centring or restoring force is always provided. It keeps bringing the superstructure
back to its original position. It prevents bearings incrementally having their movement
29
Restraint
Principles of seismic isolation
In theory, to achieve the maximum degree of seismic isolation, a superstructure should bear
on a frictionless surface. Unfortunately, horizontal forces other than from earthquakes also
can move buildings sideways. Wind is the most obvious example. A seismic isolation system
must prevent motion along the plane of isolation during wind gusts. Not only could the
building move excessively over its bearings but building occupants could suffer from motion
sickness. Such a ‘windblown’ building would, no doubt, also attract unwanted media attention!
Designers must therefore provide isolation systems with minimum horizontal restraint to
withstand code wind forces.
Damping
Finally, damping is necessary. Just as shock absorbers prevent intolerable vertical vibrations
when a car rides over bumps, damping reduces the overall dynamic response of a building.
It lessens the relative horizontal movement between the superstructure and foundations.
Bearings and other details can then be designed for far less movement, reducing both cost
and the width of gaps that accommodate movement.
H AVING identified the five requirements of a seismic isolation system we now introduce
the two principles behind the effectiveness of seismic isolation – period-shift and
damping.
First, it is necessary to appreciate the dynamic characteristics of earthquakes. Usually,
most of their dynamic energy is contained in high frequency vibrations, say between one
and ten cycles per second. These vibrations are very effective in causing dynamic resonance
in buildings that naturally vibrate horizontally at the same frequencies. Every building, when
jolted suddenly, vibrates at its own natural frequency. Or, expressed more commonly,
particularly by structural engineers, every building has its own natural period of vibration.
The natural period of a building is the time taken for a complete cycle of vibration, the inverse
of the natural frequency. A building with a natural period of 0.1 seconds has a natural
frequency of ten cycles per second.
For every earthquake, we can plot a graph of the horizontal acceleration response of
buildings with increasing natural periods (Figure 2.21). Such a graph is termed a ‘response
spectrum’. Response spectra from many earthquakes are plotted, averaged and simplified
before being introduced into countries seismic design codes. Although the shapes of response
spectra vary for each earthquake, site soil conditions and seismic code, they all display the
common feature that for periods longer than about 0.5 seconds, the acceleration response of
the building diminishes. Seismic isolation exploits this fact – that a building with a longer
period of vibration (a more flexible building) experiences less shaking than a shorter period
building. Compared to an otherwise identical conventional building, an isolated building is
allowed to move freely but to re-centre through horizontal spring action. Seismic isolation
therefore creates a very flexible building and introduces a ‘period-shift’. A period shift, often
between two to three seconds, dramatically reduces acceleration response. Period-shift is
the main reason for the effectiveness of seismic isolation in reducing superstructure
accelerations that otherwise cause widespread damage.
Second, and as mentioned previously, damping also reduces the dynamic response of
a building (Figure 2.22). Damping reduces resonance. It causes the amplitude of vibrations
to decay by absorbing dynamic energy. Damping in buildings is expressed as a percentage of
critical damping. Rather than a building vibrating to-and-fro after being suddenly pushed
sideways, a critically damped system just returns to its original position. Every building has
some inherent damping depending on its structural materials. For example, reinforced
concrete buildings possess greater damping than those with steel framing. But seismic isolation
systems typically provide three to five times more damping at the isolation interface.
Acceleration
response of building Period of
conventional building
A
Period of
isolated building
Reduction in
response
B
Principles of seismic isolation
2.21 A response spectrum shows how the acceleration response of a building to a typical
earthquake varies according to the natural period of the building. The effectiveness of
31
building
Damping of
isolated building
Reduction in
response
due to
increased
damping
0 0.5 1.0 1.5 2.0 2.5
Natural period of
vibration (secs)
2.22 A response spectrum of a typical earthquake showing the reduction in the response
of an isolated building due to increased damping
E ACH isolation system must satisfy the five requirements of movement capability, vertical
support, re-centring, restraint and damping. Typical seismic isolation hardware that is
described in more detail in Chapter 3 meets two or more of these requirements (Figure 2.23).
Lead-rubber bearings, for example, meet all five. Their height enables them to move horizon-
tally. When that occurs, the elasticity of their rubber causes them to spring back or re-centre.
Their steel plate-rubber sandwich construction enables them to support the building weight
without excessive vertical settlement, and the lead plug both restrains wind movement and
dampens vibrations by absorbing energy.
Curved sliders also integrate all five requirements, but via totally different mechanisms.
Movement is provided by a Teflon-coated slider that bears on highly polished stainless steel
surfaces. Gravity loads pass through the slider and to the lower stainless steel plate. The static
friction along this interface is sufficient to provide wind restraint, and then when exceeded
in an earthquake, dynamic friction is the method by which energy is absorbed and the system
damped. The concave curved surface means that whenever the slider displaces horizontally
it also rises. The weight from the building constantly drives the slider down the curved surface
to cause the re-centring action.
Lead cylinder
Steel plate
Rubber
PTFE coated
Stainless steel sliding surface sliding puck Curved stainless
steel surfaces
PTFE (recessed)
Confined elastometric
material
2.23 Different types of seismic isolation hardware: (a) Rubber bearing, (b) Lead-rubber
bearing, (c) Slider, (d) Curved slider, (e) Lead damper, and (f) Steel damper
33
34
Horizontal sliding bearings provide movement capability and support the building
weight. Their internal friction provides restraint and damping. However they need to be
Principles of seismic isolation
Notes
1 Naderzadeh, A., 2009, ‘Application of seismic base isolation technology in Iran’, Menshin (Japan), no. 63, pp.
40–7, p. 42.
2 Christopoulos, C. & Filiatrault, A., 2006, Principles of passive supplemental damping and seismic isolation, Pavia:
IUSS Press.
3 Pan, P., Zamfirescu, D., Nakashima, M. et al., 2005. ‘Base-isolation design practice in Japan: introduction to
the post-Kobe approach’, Journal of Earthquake Engineering, vol. 9, no.1, pp. 147–71.
4 Martelli, A., Forni, M. & Clemente, P., 2012, ‘Recent worldwide application of seismic isolation and energy
dissipation and conditions for their correct use’, Proceedings of the 15th World Conference on Earthquake
Engineering, Lisbon.
5 Zhou, F.L., Tan, P., Heisa, W.L. et al., 2013, ‘Lu Shan earthquake M 7.0 on 2012.4.12 and recent development
on seismic isolation, energy dissipation & structural control in China’, Proceedings of the 13th World
Conference on Seismic Isolation, Energy Dissipation and Active Vibration Control of Structures, Sendai.
– 3 –
Introduction
F OR most of its brief history, seismic isolation can be considered a passive system. That is,
the hardware is installed in buildings and sits there passively, waiting to be activated
by an earthquake. Even during earthquake shaking, isolation devices passively endure
the dynamic movement while simultaneously performing their damage mitigation and
lifesaving functions. They absorb and dissipate earthquake energy, dampening down
movements. They provide the horizontal flexibility which is at the heart of seismic isolation,
all the while bearing the weight of the superstructure above. And finally, their re-centring
forces are always bringing the building back to its original position above the founda-
tions. Once installed, protected against fire if necessary, and subject to regular inspection
and maintenance, isolation devices are left alone to passively play the roles outlined above
that they are designed for.
However, recently, research and development has focussed upon active systems. Usually
relying on movement sensors and computer-controlled hydraulic rams, active systems
counteract the ground movement. The ultimate goal of active seismic isolation is to keep a
building stationary while everything around and under it shakes (Figure 3.1). As you would
expect, active systems are high-tech and very expensive. Although they control wind vibrations
in several tall buildings, apart from a demonstration building discussed near the end of this
chapter, they have yet to find acceptance. But as individuals and companies become both Seismic isolation systems and hardware
become more risk-adverse and enamoured by high-tech solutions, their uptake is sure to
increase in the future, particularly shortly after damaging quakes.
Since active systems can be considered still under development, the focus of this chapter
is upon passive systems. We discuss the three categories of systems – elastomeric and dampers,
sliding, and other systems in the following sections.
35
36
Seismic isolation systems and hardware
Actuator
Sensor
Computer
Data flow
Sensor
3.1 An active isolation system keeps the building stationary during shaking. Sensors feed
information to a computer-controlled hydraulic ram
T HE most basic component of these systems is the elastomeric bearing. Used in bridges
for over 100 years, these bearings allow decks to move over piers and abutments during
thermal expansion and contraction. The main difference between bridge bearings and those
providing seismic isolation in buildings (and bridges) is their height. Seismic movements,
which are typically ten times greater than thermal movements, require a greater thickness of
rubber.
Elastomeric bearings consist of alternating layers of usually natural rubber and thin
steel plates, steel shims. The shims prevent excessive lateral bulging and settlement under
compression (Figure 3.2). The bearings range in diameters from 500 mm to 1500 mm
depending on the compression force acting and the horizontal design movement. The
approximate thicknesses of the rubber and steel shims are 10 mm and 3 mm respectively.
The bearings are therefore very stiff vertically yet flexible for horizontal movement. Typically
a bearing might settle up to 1–3 mm initially due to the column load acting upon it and a
further 50 per cent over the following decades. Yet it can move horizontally 300–500 mm.
These bearings are extremely robust. The horizontal rubber layers and protective exterior
layer are vulcanised to the steel shims. Under long-term loads they are stable. As the rubber
ages it hardens slightly, but a bearing is not detrimentally affected by oxidation.1
Elastomeric bearings placed under the columns of a seismically isolated building
are well-suited for supporting the building weight, providing horizontal flexibility and the
necessary re-centring force. However they may lack significant damping and might be too
flexible under wind gusts. Three approaches are taken to provide additional damping.
Steel plate
Bulge
Compression Compression
Original
shape Minor
bulge
3.2 The effectiveness of steel shims at reducing the settlement of an elastomeric bearing
The first is to insert one or more cylindrical lead plugs of approximately 60–150 mm
diameter into the bearing. This is the widely used lead-rubber bearing. The lead provides
rigidity against wind and dissipates large amounts of energy as it yields and recrystallizes
during an earthquake (Figures 3.3 and 3.4). During their development, these bearings were
subject to punishing test regimes to ensure satisfactory short and long-term performance. In
most installations two prototypes must survive testing so severe as to disqualify them for
eventual incorporation in the building.
Additional or alternative damping can also be provided by adjusting the properties
of natural rubber. Proprietary compounds are added and the curing process adjusted, or
synthetic rubber is used. High-damping rubber bearings are commonly used as seismic
isolators.
Provision of damping in conjunction with elastomeric bearings is also achieved with
supplementary external dampers. Lead dampers are particularly popular in Japan. These pure Seismic isolation systems and hardware
lead curved cylinders are typically 200 mm in diameter and about 1 m high. They are
thoroughly tested and accept maximum displacements of up to 800 mm.2 A less common
type of lead damper is the lead-extrusion damper. It is like a scaled-up car shock absorber
but filled with lead rather than oil (Figure 3.5). The damper connects a superstructure to its
foundations, across the isolation plane. When the ground moves, exceeding the yield force
of the damper, the lead against the bulge of the shaft melts and then recrystallizes after the
shaft has moved through the lead.3 Since the damper only works in the direction of its length,
at least four dampers, two in each orthogonal direction and widely separated in plan are
37
Orifice Lead
Shaft
without breaking. They dissipate energy by their steel members yielding. At Union House,
Auckland, which is founded on long sleeved piles like the Wellington Central Police Station,
horizontal restraint and damping is provided by mild steel tapered dampers (Figures 3.10 and
3.11).
Hydraulic dampers filled with oil or a silicon fluid are another option. They are less
frequently used than other dampers, possibly due to their cost, although in theory they are
the most effective. Their advantage is in reducing the lateral loads for which the superstructure
need be designed. This is because the force they transfer to the superstructure is proportional
to the velocity of movement. When there is little force in a nearly centred elastomeric bearing,
the oil damper will be moving fast and exerting its maximum force. Then, when the Seismic isolation systems and hardware
elastomeric bearing is nearing the end of its displacement, subject to maximum force and the
velocity is low, the hydraulic damping force has died away.
Sliding devices
T WO types of sliding devices are common in seismic isolation systems. First, relatively
simple sliding bearings, and second, the far more sophisticated curved sliders that, like
41
Damper
Basement
Isolated piles
Non-isolated piles
Damper reaction block
3.8 A basement plan showing the location of lead-extrusion dampers, Wellington Central
Police Station
Sliding bearings transfer compression load and allow horizontal sliding. In most cases
the sliding surfaces in contact are stainless steel and polytetrafluoroethylene (PTFE),
commonly known as Teflon. These types of bearings have been used for many years to
accommodate thermal movements in bridges. They are known as pot bearings due to the way
a thin layer of rubber is confined in a pot-shaped ring (refer to Figure 2.23(c)). The rubber
allows the sliding surfaces a small degree of rotation, and is confined to allow the bearing to
carry high compression loads. These sliding bearings are often used in seismic isolation and
are available from many manufacturers.
3.9 A steel damper integrated with an elastomeric bearing
Source: Nippon Steel & Sumikin Engineers
The main requirement of a sliding bearing is that it slides freely with a minimum of
friction. Although stainless steel and Teflon tend to stick together before being released by
horizontal movement, very low coefficients of dynamic friction in the order of one per cent
are achievable for low velocities, rising to a more typical and constant value of approximately
10 per cent at the higher velocities experienced during earthquakes. Higher friction values
can provide a valued source of additional damping. In contrast to conventional construction
materials, low coefficient of friction values occur when the compression stresses are greatest. Seismic isolation systems and hardware
the late 1980s when flat stainless steel surfaces were curved into concave forms. Now the
44
Seismic isolation systems and hardware
3.10 Union House, Auckland, New Zealand, 1983. Superstructure wind and seismic loads
are resisted by the exterior bracing system
3.11 Tapered steel damper that connects the superstructure to the foundation provides
damping when it yields during an earthquake
sliding bearing possessed re-centring capability, and along with damping provided by friction
between its sliding surfaces, it met all the requirements for seismic isolation. As the sliders
that support the building move along the concave surfaces the building rises. Gravity loads
from the weight of the building above force sliding towards the centre of the bearing thereby
causing re-centring (Figure 3.12).
The first friction pendulum bearing was applied in a 1994 retrofit project, and since
then the original design has been developed from a single pendulum operation with two
sliding surfaces to the Triple PendulumTM with its four surfaces (Figure 3.13). In this latest
version of what are generically referred to as curved sliders, the radii of curvature of the sliding Seismic isolation systems and hardware
surfaces, which determine the natural period of the bearings, vary and the surfaces can possess
different friction values. This improves the dynamic performance over a wide range of
earthquake motions and allows smaller diameter bearings. Another advantage of curved
sliders is their thinness (Figure 3.14). This means more compact construction detailing and
a reduction of forces acting on structural members adjacent to the bearings. Like elastomeric
systems these bearings have been, and in every project are, subjected to strenuous full-scale
testing regimes.
45
46
Weight from
building
Gain in
Seismic isolation systems and hardware
height
Puck surfaces
coated with
PTFE
Horizontal
movement
Highly polished
stainless steel
Slider is centred Slider maximum movement
Sliding surfaces
3.13 Schematic view of a Triple PendulumTM bearing centred, and then at its maximum
displacement
W HILE the seismic isolation systems and devices discussed above constitute the vast
majority of both those used currently and in the past, it is worth describing some that
are less common as they may be applicable in certain situations. Although their configuration
is very different from what has already been presented, they all encapsulate the basic principles
of seismic isolation, namely increasing flexibility (the natural period of vibration), allowing
large displacements along the isolation plane and providing increased damping.
The Sendai Mediatheque, Japan, designed by Toyo Ito, exploits the flexibility of its sub-
surface structure (Figures 3.15 and 3.16). To achieve the necessary delicacy of its iconic
braced-tube clusters yet protect them from seismic damage, some method of isolation was
3.14 A cut-away view of a curved slider
Source: FIP Industriale S.p.A., Italy
required. In this case, the interface of superstructure and foundations consists of four
basement ductile moment frames.4 The three beams per frame are designed to yield in a large
earthquake (Figure 3.17). As sacrificial elements, they suffer structural damage while Seismic isolation systems and hardware
absorbing earthquake energy. In the process they reduce and limit the forces transferred into
the superstructure. This bespoke isolation system does not provide the large period shift of
typical isolation systems. Perhaps this was one of the reasons for suspended ceilings collapsing
during the 2011 Tōhoku earthquake.5, 6
Another form of isolation can be achieved by letting structure, such as a structural
wall, rock to-and-fro. Once a structure starts rocking during an earthquake its natural period
increases substantially, partially isolating it from the earthquake shaking.7 Damping occurs
as the ends of the wall impacts the soil, but in some cases additional damping devices are
47
provided. Although allowing the structure to rock may protect it and save foundation costs,
Seismic isolation systems and hardware 48
some of the benefits of a full seismic isolation system, such as reduction of damage to contents,
will not be realized.
The Hermès building, Tokyo, designed by Renzo Piano Building Workshop, is a high
profile example of seismically isolating a building by allowing rocking (Figure 3.18). An
article elaborates:
At 50 m tall and with a main structural span of only 3.8 m, the unusual slenderness
of the structure results in high overturning moments during an earthquake and Seismic isolation systems and hardware
high levels of tension in the columns. The engineer . . . found inspiration in the
tall, thin wooden Buddhist pagodas of Japan . . . the same principle was adopted,
with the columns on one side of the frame being held in base joints that allow uplift
and rotation simultaneously and seismic energy to be absorbed by viscoelastic
dampers.8
In all the examples of seismically isolated buildings considered to date, isolation planes
are located at the bases of the buildings. However, in some situations it is advantageous for
49
the isolation plane to be located in a middle storey. An example of this approach is found at
50
Tubes
supporting
Seismic isolation systems and hardware
superstructure
Steel I Beam
Steel tubular
column
Pin joint
Basement
level
3.17 An elevation of a basement ductile steel moment frame that protects the
superstructure
the Iidabashi First Building, Tokyo.9 Eight floors of offices that rise above the ground floor
retail area provide generous (16 m) column-free spaces. On top of this structure, and
seismically isolated from it by rubber bearings and lead dampers, sits a further five levels of
apartments. When the main structure moves during an earthquake, the isolated upper floors
vibrate out-of-phase. The levels of force within the main structure are thereby reduced. The
isolation level accommodates building services, and the two external apartment elevator
towers that cantilever above the main structure are seismically separated from the apartments
but linked to them by seismic movement joints.
There are also many less common proprietary systems, often developed for special
situations. The RoGliderTM is an example (Figure 3.19). Suitable for both light and heavy
vertical loads, a puck coated with PTFE slides between two stainless steel plates. Rubber
membranes that connect the puck to the plates provide the re-centring force. At the Wanganui
Hospital these bearings were designed for a maximum displacement of 450 mm and their
10 per cent coefficient of friction provides restraint against wind load as well as dynamic
damping. Manufacturers’ catalogues can be consulted for other examples.
Researchers around the world are continuing to develop new systems. Every year, one
or two new ideas are presented in professional journals or at conferences. Recently, two
Seismic isolation systems and hardware
3.18 Hermès building, Tokyo, Japan, 2001. The structure is designed to partially self-
isolate by rocking
51
52
Seismic isolation systems and hardware
3.19 A RoGlider™ installed between a foundation pad and substructure framing under the
Wanganui Hospital, New Zealand
Active systems
A CTIVE isolation systems are used in small numbers of very tall buildings, mainly to
control wind vibrations. They are similar to, but more sophisticated versions of, passive
tuned-mass dampers, designed to vibrate out-of-phase with the superstructure and hence
reduce overall dynamic response. In active systems, sensors detect movement of the building
and the ground, and signals are sent to computer-controlled hydraulic jacks. The jacks force
a heavy mass in the building, or the whole building, back-and-forth with similar frequencies
to the earthquake shaking to keep the building as stationary as possible.
Obayashi Corporation’s Technical Research Institute in Tokyo is the first building to
use this technology (Figures 3.20 and 3.21). The system, designed to reduce building shaking
3.20 Technical Research Institute, Obayashi Corporation, Tokyo, Japan. An active isolation
system protects the building
to 1/30th of actual ground shaking has required very sophisticated design. All known
eventualities have been taken into account. In the event of an electrical power failure the
Seismic isolation systems and hardware
hydraulic jacks are powered from tens of high pressure oil cylinders, and in the worst case
scenario, the building relies upon a back-up passive elastomeric isolation system. Needless
to say, this system is extremely expensive, but the manufacturer is trying to reduce costs for
implementation in hospitals and manufacturing premises, such as for precision measuring
equipment. There is also an option of applying this technology to just one floor of a building.
Notes
1 Kelly, T.E., Skinner, R.I. & Robinson, W.H., 2010, Seismic isolation for designers and structural engineers,
Kanpur: NICEE, p. 93.
2 Pan, P., Zamfirescu, D., Nakashima, M. et al., 2005, ‘Base-isolation design practice in Japan: introduction to
the post-Kobe approach’, Journal of Earthquake Engineering, vol. 9, no. 1, 147–71.
3 Kelly, T.E. et al., p. 57.
4 Sasaki, M., 2002, ‘Sendai Mediatheque, Japan’, Structural Engineering International, no. 3, pp. 146–8.
5 ‘March 11 Earthquake at the sendai mediatheque.’ Available from: www.youtube.com/watch?v=TKgURstRt_A
(accessed: 21 January 2016).
6 ‘Sendai mediatheque’s damages of the Great East Japan Earthquake.’ Available from: www.youtube.com/
watch?v=0D4V803Z5O8 (accessed: 21 January 2016).
7 Rezai, M., Patterson, A. & Hubick, G., 2012, ‘Nonlinear seismic analysis and retrofit of BC Place Stadium using
rocking foundation and viscous dampers’, Proceedings of the 15th World Conference on Earthquake
Engineering, Lisbon.
8 Heneghan, T., 2001, ‘Japanese lantern’, The Architectural Review, vol. 220, no. 1255, pp. 78–81, p. 81.
9 Tasaka, M., Mori, N., Yamamoto, H. et al., 2008, ‘Applying seismic isolation to buildings in Japan – retrofitting
and middle-story isolation’, Proceedings of the 18th Analysis and Computation Speciality Conference, ASCE,
Vancouver, p. 6.
10 Hosseini, M. & Farsangi, E.N., 2012, ‘Telescopic columns as a new base isolation system for vibration control
of high-rise buildings’, Earthquakes and Structures, vol. 3, no. 6, pp. 853–67.
– 4 –
Introduction
W HEN modern seismic isolation was first applied to buildings in the 1980s, its claim for
effectiveness was based entirely upon computer modelling and laboratory testing of
isolation hardware, such as lead-rubber bearings. Those early applications of the then new
technology required a certain degree of faith in the theory of seismic isolation as well as in
the computer analyses that constituted the basis of the structural engineering designs.
Now, in 2016, the effectiveness of seismic isolation not only continues to be validated
in virtual environments, but it has been demonstrated both in the real world of full-scale
testing and, more significantly, in post-earthquake urban centres of several countries.
Seismically isolated buildings have now been tested by actual earthquakes ranging from those
of low intensities to more intense motions similar to design earthquakes. These buildings
protected by seismic isolation are on sites characterized by different soil conditions, in
particular soft soils, and have been subject to damaging earthquakes from nearby faults, as
well as those several hundreds of kilometres away.
The following sections demonstrate the effectiveness of seismic isolation as determined
by three quite different approaches.
As well as the effectiveness of seismic isolation being proven during every building
design, our initial and on-going confidence in seismic isolation is also grounded in the large
body of research that began in the 1980s. Since then, tens of PhD theses have explored aspects
of the topic, hundreds of research papers presented at conferences around the world and
numerous scholarly books written. Each of these publications, to some extent, is based on
the results of computer modelling. For example, a single table within one book summarizes
findings from 100 computer analyses.1 Research is on-going. At earthquake engineering
conferences around the world, we can guarantee that new research findings on seismic
isolation are being presented and discussed.
Physical testing
I T is comforting that the effectiveness of seismic isolation is not only reliant upon theoretical
and virtual studies. Physical testing has been a vital component in the early development
of seismic isolation, and in its on-going quality assurance programme. Initially, just devices
like lead-rubber bearings, curved sliders and many types of dampers were tested on their
own. These relatively small test set-ups were and still are challenging enough. They have to
replicate the large gravity forces acting on bearings plus simulate the dynamic motions that
occur during large earthquakes (Figures 4.2 and 4.3).
With the recent availability of hydraulically activated shaking tables, physical testing can
be scaled up. Now, not only can several isolation devices be tested simultaneously, but entire
isolated buildings experience recorded strong motion earthquake records. For example, the
largest shaking table in the world, E-Defense, subjects concrete and steel buildings up to five
storeys high to 3-D shaking. In one test of a 500-tonne steel frame building, the effectiveness
4.2 The testing of elastomeric bearings. Two bearings for the Hospital of Light, Lisbon
being tested
Source: FIP Industriale S.p.A., Italy
4.3 A bearing for the China Basin Landing office building, San Francisco being tested
57
of two isolation systems were compared and confirmed when compared to a conventional
fixed-base building: ‘Both isolation systems [elastomeric and curved slider] successfully
Effectiveness of seismic isolation
mitigated damage in the superstructure from a range of strong ground motions’,2 with lesser
measured floor accelerations than in the fixed-base case. Due to their different isolation
periods, the elastomeric system more strongly attenuated floor accelerations at lesser ground
acceleration, whereas the curved slider was more effective at higher ground accelerations. The
three test earthquake ground motions simulated ‘frequent’, ‘long-duration subduction’ and
‘near-fault’ motions; the last two of which are well recognised as providing the most severe
test of a seismically isolated system.
In another test, also at E-Defense, the earthquake performance of an isolated and a
fixed-base four-storey concrete building were investigated.3 The aims of this test were to check
the effectiveness of seismic isolation for both near-fault and long period motions that are
expected on sites underlain by soft soils that vibrate slowly. The focus was on preventing
damage to furniture and mobile equipment. Seismic isolation proved to be very effective for
the near-fault earthquake motion. Superstructure accelerations were only 32 per cent of the
ground accelerations while the fixed-base superstructure amplified ground accelerations
(measured at roof level) by 3.6 times. Therefore, the maximum levels of acceleration in the
seismically isolated superstructure were 1/10th of those in the fixed-base building, thereby
providing far greater protection to building contents. However, when the same buildings were
subject to extreme long-period earthquake vibrations the effectiveness of seismic isolation
diminished greatly. The long-period nature of the shaking resonated the flexible isolation
system. Although the isolation system was unable to reduce the level of ground accelerations
transferred into the superstructure, the maximum acceleration of the isolated building was
only 66 per cent of that in the fixed-base building.
This test illustrated the need for castors under mobile equipment to be locked whenever
stationary. But more significantly, it emphasises how the effectiveness of seismic isolation
depends on the dynamic characteristics of the earthquake waves, they themselves dependent
upon the proximity of the epicentre and the softness and depth of underlying soils. This is
why geotechnical and seismological advice must always be sought before deciding to adopt
seismic isolation.
4.4 Two identical buildings, one base-isolated and the other a conventional fixed-base
building, at Tōhoku University, Sendai, Japan
59
Several comparisons have been made between similar seismically isolated and con-
ventional buildings located within the same area. Ron Mayes (2007) cites two examples.6 First,
Effectiveness of seismic isolation
the seismically isolated Japanese Postal Computer Center suffered no damage during the
devastating 1995 Mw 7.1 Kobe earthquake. Instrumentation showed the isolation system
reduced ground accelerations by a factor of 3.4. A similar instrumented six-storey reinforced
concrete building in the next block amplified the ground shaking by a factor of 3.0 at the
roof. So compared to the conventional building, isolation reduced the shaking at roof level
by a factor of 10. A similarly impressive case is cited following the 1989 Loma Prieta
earthquake, California: ‘The University of Southern California (USC) Hospital – the world’s
first base-isolated hospital (Figure 4.5) – had no damage at all, while the Los Angeles Country
General Hospital complex 1 km away from the isolated USC hospital suffered $389 million
in damage.’ The benefits of seismic isolation also became apparent in the Rikkyo University
Chapel building, Tokyo during the 2011 Tōhoku earthquake.7 This historic brick masonry
building had been seismically retrofitted by a seismic isolation scheme. During the earthquake
its bearings and dampers moved 40 mm, and the superstructure experienced only 38 per cent
of the peak ground acceleration. Several metres away an instrument on the second floor of
a conventional brick building indicated an amplification of the ground accelerations by a
factor of 3.0.
4.5 USC Hospital, USA, the world’s first seismically isolated hospital, survived the 1989
Loma Prieta earthquake without damage
Source: Dynamic Isolation Systems Inc.
Of all countries, Japan provides the most information about the measured and observed
performance of seismically isolated buildings. Its active Japan Society of Seismic Isolation
(JSSI) not only promotes seismic isolation but initiates research and publishes reports. In his
state-of-the-art summary in 2009, Nagihide Kani (2009) summarized the performance of 23
buildings during six Japanese earthquakes from 2003–08. The recorded movements along the
isolation plane ranged from 40–200 mm, with the superstructure accelerations reduced to
between 25–50 per cent of the ground accelerations in instrumented buildings.8
In 2013, Kani reported this time on the performance of seismically isolated buildings
during the 2011 Tōhoku earthquake.9 Completed questionnaires were obtained from 213
seismically isolated buildings mainly in the most affected Tōhoku and Kanto areas. Of these
buildings, 43 are instrumented and 101 experienced displacement along their isolation planes.
Typical measured displacements varied from 50–200 mm with a maximum of 415 mm
recorded at Miyagi. The maximum residual displacement was around 20 mm. No structural
or non-structural damage was reported, apart from two instances of damaged plumbing
pipes, but some dampers were damaged. Malfunctioning movement joints were noted in 90
cases. Causes included a lack of maintenance, poor construction and problems with the joints
themselves. Remarkably, no building contents or pieces of furniture fell or moved, apart from
three items on castors. Building occupants viewed seismic isolation very positively.
Another paper discusses the performance of 20 instrumented buildings in Tokyo
ranging in height from two to 21 storeys during the same event.10 For low values of ground
acceleration some buildings behaved, as expected, like conventional structures and amplified
the ground accelerations up to a factor of 2.0. But once the peak ground acceleration increased
over 0.3 g, the isolation systems reduced superstructure accelerations to approximately 50
per cent of those measured below the isolation plane (Figures 4.6 and 4.7). Another study of
the records of eight selected seismically isolated buildings confirms these findings.11 It notes
there was little if any amplification of horizontal accelerations up the height of a seismically
isolated building. It can be reasonably assumed then, that unlike a conventional building, the
floors of seismically isolated buildings experience similar and low values of acceleration.
This section has summarized the effectiveness of seismic isolation as evidenced from
measurements and observations after earthquakes located almost entirely in Japan. The
numbers of isolated buildings surveyed, the research findings, including the extremely
positive feedback from their occupants, all represent impressive evidence that points to the Effectiveness of seismic isolation
overall effectiveness of seismic isolation. But before concluding this chapter we return again
to the Christchurch Women’s Hospital discussed in Chapter 1. Unfortunately it was not
instrumented at the time of the 22 February 2011 Christchurch earthquake that caused severe
structural and non-structural damage to most buildings in the city’s central business district,
leading to hundreds being demolished. However the Christchurch Women’s Hospital, even
though founded on relatively soft soils, was essentially undamaged and functional immediately
after the quake. The excellent seismic performance of this building is yet further evidence of
the effectiveness of seismic isolation.
61
62
Effectiveness of seismic isolation
4.6 The Yozemi Tower, Tokyo, Japan, which is seismically isolated by elastomeric bearings
and oil dampers, displaced 90 mm along its isolation plane. Its ground floor and top floor
accelerations were 30 per cent and 45 per cent respectively of the ground accelerations
4.7 Main building, Shimizu Corporation Institute of Technology, Tokyo, Japan, is
seismically isolated by six lead-rubber bearings at the top of ground floor columns. The
bearings deformed 90 mm and the maximum ground acceleration was reduced by the
seismic isolation system to 52 per cent at the second floor and 55 per cent at the top floor
Notes
1 Kelly, T.E., Skinner, R.I. & Robinson, W.H., 2010, Seismic isolation for designers and structural engineers,
Kanpur: NICEE, p. 154. Effectiveness of seismic isolation
2 Sasaki, T., Sato, E., Ryan, K.L., Okazaki, T. et al., 2012, ‘NEES/E-Defense base-isolation tests: effectiveness of
friction pendulum and lead-rubber bearings systems’, Proceedings of the 15th World Conference on Earthquake
Engineering, Lisbon, p. 10.
3 Sato, E., Kajiwara, K., Furukawa, S., Ji, X. et al., 2010, ‘Full-scaled shaking table test of a hospital made of a
base-isolated 4-storey concrete structure’, Proceedings of the 9th US and 10th Canadian Conference on
Earthquake Engineering, Toronto, paper 303.
4 Moroni, M.O., Sarrazin, M. & Soto, P., 2012, ‘Behaviour of instrumented base-isolated structures during the
27 February 2010 Chile earthquake, Earthquake Spectra, vol. 28, no. S1, S407–24.
63
64
5 Nakamura, Y., Hanzawa, T., Hasebe, M., Okada, K. et al., 2011, ‘Seismic isolation and protection, systems:
report on the effects of seismic isolation methods from the 2011 Tohoku-Pacific earthquake’, The Journal of
Effectiveness of seismic isolation
the Anti-Seismic Systems International Society (ASSIS), vol. 2, no. 1, pp. 57–74.
6 Mayes, R.L., 2007, ‘Ronald L. Mayes’, The Structural Design of Tall and Special Buildings, vol. 16, no. 1, pp.
3–36, p. 18.
7 Seki, M., Yoshida, O. & Katsumata, H., 2012, ‘Behaviour of Rikkyo Univ. chapel building retrofitted by seismic
isolation in the 3.11.2011 earthquake, Japan’, Proceedings of the 15th World Conference on Earthquake
Engineering, Lisbon.
8 Kani, N., 2009, ‘Current state of seismic isolation design’, Journal of Disaster Research, vol. 4, no. 3, pp. 175–81.
9 Kani, N., Ogino, N., Kitamura, Y. & Fukazawa, Y., 2013, ‘Performance of response-controlled buildings during
the huge 2011 earthquake’, Proceedings of the 13th World Conference on Seismic Isolation, Energy Dissipation
and Active Control of Structures, Sendai.
10 Matsuda, K., Kasai, K., Yamagiwa, H. & Sato, D., 2012, ‘Responses of base-isolated buildings in Tokyo during
the 2011 great East Japan earthquake’, Proceedings of the 15th World Conference on Earthquake Engineering,
Lisbon.
11 Iiba, M., Kashima, T., Morita, K., Azuhata, T. et al., 2013, ‘Behaviour of seismically isolated buildings based
on observed motion records during the 2011 great east Japan earthquake’, Proceedings of the 13th World
Conference on Seismic Isolation, Energy Dissipation and Active Control of Structures, Sendai.
– 5 –
Introduction
I N the previous chapter we discussed the effectiveness of seismic isolation. Our confidence
in the system is based upon three types of evidence: namely computer modelling, physical
testing of isolation components as well as full-scale building testing and, finally, measure-
ments and observations made on seismically isolated buildings during damaging earthquakes.
Having established that seismic isolation is worthy of our confidence, this chapter explores
the extent of its benefits and limitations. Seismically isolated buildings are compared to
conventional fixed-base buildings. The following chapters expand upon some of the topics
raised here.
The benefits of seismic isolation that make it attractive to building owners are
considered in two sections: first, benefits realised before the occurrence of a damaging
earthquake, and second, the benefits experienced during and after such an event. As far as
limitations are concerned, we explain how for reasons of geology, site location, building height,
and maintenance and cost considerations, seismic isolation may not be a sensible strategy,
and in some cases, may be impractical to implement.
S URPRISINGLY, seismic isolation can deliver benefits even before an earthquake strikes.
Many of these benefits are realized during the design phase of a building when they can
lead to improved architectural features. Take, for example, the size of and configuration of
structural members. Although seismically isolated buildings are generally designed to be as
strong as equivalent conventional buildings, they do not need the same capacity for ductility.
Their energy absorption is provided by their damping mechanisms. In conventional buildings
ductility is achieved by applying the Capacity Design approach. First, by ensuring a hierarchy
65
of structural damage whereby the most vital structural elements, such as columns that support
66
the whole building weight, have greater protection than say beams. Beams can support their
loads even if damaged at their ends. Second, brittle failure modes such as shear failure are
Benefits and limitations of seismic isolation
supressed by increasing the shear strength of structural members such as columns, beams
and structural walls. Then, earthquake energy can be absorbed by the yielding of structural
steel, or steel reinforcement in reinforced concrete structures, at non-critical discrete locations
throughout the structure.
Such ductile conventional buildings, at least if they rely upon moment frames for their
horizontal resistance, require regular inter-storey heights and columns stronger than beams.
But since seismically isolated buildings are normally designed to remain elastic, there is not
the same need for vertical regularity. Some of the traditional rules for sound seismic
configuration can be eased.1 For example, normally very undesirable soft storey and short
column configurations that are normally considered critical structural weaknesses may be
possible given that structural members are protected from damage by seismic isolation
(Figures 5.1 and 5.2). This potential is exploited in a seismically isolated Japanese building
with very slender columns and long-span beams (Figure 3.20).
As well as preventing structural damage, seismic isolation reduces greatly the horizontal
inter-storey drifts that occur during an earthquake (Figure 5.3). Where a need to reduce
horizontal flexibility rather than meet strength requirements drives the sizes of structural
members, seismic isolation enables smaller structural member sizes. Other significant archi-
tectural advantages arise in a building subject to less inter-storey drift. These include reduced
seismic separation gaps between architectural elements, such as glazing and mullions, or
exterior cladding units such as precast concrete panels, and between architectural elements
and structural members. The outcome is simplified, less visually prominent construction
detailing (Figure 5.4).
Seismic isolation then offers the possibility of more architectural freedom. Structural
configurations unthinkable in conventional buildings can be considered. More slender
columns and structural walls are also feasible, as well as simpler and thinner separation gaps.
These potential benefits can improve architectural aesthetics. Two examples designed by
Toyo Ito are discussed in more detail in the following chapter.
But pre-earthquake benefits include more than those related to architectural quality.
Seismic isolation offers peace of mind to building owners and occupants. This is reflected in
increasing numbers of seismically isolated Japanese apartment buildings. Seismic isolation,
marketed as one of the desirable features of these buildings, attracts both tenants and increased
rentals. There also may be other financial benefits for owners such as being able to negotiate
reduced earthquake insurance premiums or even self-insure, as discussed in Chapter 9.
Stiff and strong upper
floors due to masonry infills
RC frame with
infill or RC wall with
Benefits and limitations of seismic isolation
penetrations
Short column
Elevation
Total drift
Inter-storey
drift
Drift or
displacement at
the isolation
plane
5.3 Inter-storey drifts in conventional and seismically isolated buildings. Most of the total
drift of an isolated building occurs at the isolation plane, whereas in a conventional
building the total drift is the sum of all inter-storey drifts
Jamb mullion
fixed to structure Seismic frame Seismic movement
Width Width
Section Section
(a) (b)
5.4 A horizontal section through a conventional jamb mullion accommodating the small
inter-storey drifts of a seismically isolated building (a) compared to a conventional jamb
seismic mullion (b)
W HILE the pre-earthquake benefits of seismic isolation are tangible, how much more so
are those incurred during and after a large earthquake. They have been referred to in
previous chapters especially where comparing the performance of seismically isolated and
conventional buildings, but can be explicitly listed as:
Reduced accelerations and deflections are the norm in isolated buildings. For example,
Istanbul’s Sabiha Gökçen Airport, one of the largest seismically isolated buildings in the world,
Benefits and limitations of seismic isolation
5.5 Istanbul’s Sabiha Gökçen International Airport Terminal, Turkey, one of the largest
seismically isolated buildings in the world
Source: H. Darama, Arup
5.7 A bearing with its protective covering supports the steel framework
Source: H. Darama, Arup
Source: G. Mosqueda
72
Benefits and limitations of seismic isolation
5.9 Damage to ceiling mounted mechanical services at the Santiago International Airport
terminal building after the 27 February 2010 Chile earthquake
Source: G. Mosqueda
A S we have noted earlier, although seismic isolation is the best strategy to date for reducing
earthquake damage in buildings, it does have its limitations. This section begins by
explaining how the location and forms of seismically isolated buildings can reduce, or even
nullify, the effectiveness of seismic isolation. Then we consider other aspects of seismic
isolation that might be of concern to those interested in adopting the technology.
Geological conditions
Like conventional buildings, seismically isolated buildings must be constructed upon adequate
ground. Although the vertical forces induced in an isolated building caused by the tendency
to overturn under horizontal loading will generally be less severe than for a conventional
building, normal rules of geotechnical engineering apply. Not only should sites underlain by
fault lines or liquefiable soils be avoided, seismic isolation may not be appropriate for sites
with deep soft soils. The reason for this is that such soft-soil sites have their own long natural
periods of vibration that can cause seismically isolated buildings to resonate.
Examples of soft soils causing severe earthquake damage to high and flexible buildings
(with relatively long natural periods of vibration) are well documented. One of the best known
cases occurred in a localized area of Mexico City during the 1985 Mexico earthquake. A small
area of the city is built over a former lake bed underlain by soft clay. During this earthquake,
whose epicentre was over 350 km away, the soft soil responded to the bedrock shaking like
a bowlful of jelly sloshing when shaken. As well as the soft ground itself resonating with a
period of approximately 2.0 seconds, flexible buildings between 6 and 16 storeys with a similar
natural period amplified that ground shaking even further. Many of these multi-storey
buildings were severely damaged or collapsed. Adjacent low-rise masonry buildings that did
not resonate remained undamaged. Had a seismically isolated building with a natural period
around 2.0 seconds been subject to this shaking, it too would have been severely tested. One
of the primary aims of seismic isolation is to lengthen the natural period of the building to
avoid those natural periods (or frequencies) of ground shaking that contain significant
earthquake energy. Due to the change of shape of a soft soil site response spectrum, shifting
the period of the building is counter-productive (Figure 5.10).
Recent research following the 2010/2011 Christchurch earthquakes suggests that seismic
isolation with its far higher levels of damping (25 per cent compared to 5 per cent in
with perimeter retaining walls. Buildings less than 2 km from a fault are most affected. Further
74
Soil conditions
Acceleration
Acceleration
Time Time
Accelerogram recordings
Response
Isolated building Response Isolated building
period range period range
5.10 The effects of soft soil on earthquake shaking. To the left a rock site and to the right,
a soft soil site. Compared to rock, soft soil resonates. Shaking intensifies and the natural
period at peak response increases to be closer to the range of natural periods of vibration
of isolated buildings
away the pulses diminish until they can be neglected at a distance of 20 km. An appreciation
of the intensity of this effect can be gained from the fact that conventional buildings in the
same period range as seismically isolated buildings, three to five seconds, must be designed
up to 70 per cent stronger than low- to medium-rise buildings.6
If these near-fault effects could possibly apply to a given design, structural engineers
should include several near-fault earthquake records in the suite of records used for design.
According to Peng Pan and his co-authors (2005), many Japanese designers had not explicitly
designed for near-fault pulses. They believed such an event to be very rare and acknowledged
the inevitability of a serious superstructure-perimeter wall collision.7 The authors then
highlight several buildings designed for near-fault pulses where the solution was ‘to place a
very strong vertical load truss system in the first storey above the isolation level in order to
minimize the support points’. Isolators up to 1500 mm diameter sustain the compression
stresses and horizontal deformations of more than 800 mm.
As well as experiencing high energy low frequency pulses, buildings located near an
epicentre may also likely be subject to significant vertical accelerations. Apart from one 3-D
seismically isolated Tokyo building (Figures 5.11 and 5.12), building isolation systems are
ineffective vertically. Fortunately, vertical shaking is usually of far less concern than horizontal
movements but it may induce some unexpected damage to long cantilever beams and
5.11 Chisuikan, Tokyo, Japan, designed by Kozo Keikaku Engineering and completed in
2011, is a housing complex and the first building in the world to be protected by 3-D
seismic isolation
particularly to building contents. The potential damage arising from vertical seismic attack
can be considered a limitation of conventional seismic isolation.
Building height
Higher buildings are more flexible against horizontal forces than shorter buildings. As Benefits and limitations of seismic isolation
buildings increase in height, so do their natural periods of vibration. If we consider the shape
of a response spectrum for firm soils the acceleration response reduces after natural periods
exceed approximately 0.5 seconds (Figure 5.13). To a significant degree then, increasing the
height of a building tends to isolate it from short period seismic vibrations. Then seismic
isolation provides a less dramatic period shift and is less effective as compared to isolating
lower-rise buildings.
Whereas the first generation of seismically isolated buildings were low- to medium-
rise, they are now increasing in height. There are numerous isolated buildings in Japan over
75
20 storeys and this trend continues in other countries. For example, a 25 storey office tower
76
Benefits and limitations of seismic isolation
5.12 The 3-D isolation system of Chisuikan. Located in a basement, vertical isolation is
achieved through the air-springs atop three blue cylinders and silver-coloured dampers.
The bearing above the red beams provides horizontal isolation
has just been completed in Jakarta (Figure 5.14). A fixed-base period of approximately 2.0
seconds was increased by a further 2.5 seconds with isolation. This achieved significant
reductions in inter-storey drifts and forces within the building which translate to overall
improved seismic performance. Immediate occupancy is expected after the design
earthquake.8 The heights of seismically isolated buildings are not as limited as was once
assumed due to advances in bearing manufacture plus designers’ choice for the natural periods
of isolated buildings to be significantly longer than the 2.5–3.0 seconds of the past.
Just as few very tall buildings are seismically isolated, so too very small buildings, such
as houses. Although in principle seismic isolation is applicable to small buildings, and over
4000 seismically isolated houses have been constructed in Japan, a number of factors limit
extensive application of the technology to houses. A study of base isolation of timber-framed
buildings concludes that because the cost of bracing walls is relatively low, isolation may only
Acceleration response
3 storeys
High
8 storeys
14 storeys
Medium
Low
5.13 A response spectrum indicating how the acceleration responses of buildings reduce
as they become higher
be justified where there are insufficient bracing walls. Also, unless isolators are placed under
a heavy (concrete) suspended floor slab, horizontal deflections in wind storms are excessive.9
The provision of such a slab increases construction costs significantly. This is why a special
synthetic sheet slip layer under concrete foundations is attractive, but a combination of high
costs and unconvincing test results is unlikely to see this system applied in practice. Methods
to limit wind movement of lightweight houses by ‘locking’ the isolation system and releasing
it only during an earthquake by some fail-safe system were not considered. The study did
acknowledge that seismic isolation of light-weight houses might be justified where contents
require protection.
Adjacent buildings
The effectiveness and application of seismic isolation can be limited by adjacent buildings.
77
It makes little sense to isolate a building that is close to a more vulnerable building. Damage
78
Benefits and limitations of seismic isolation
5.14 The Gudang Garam Tower, Jakarta, Indonesia. A seismically isolated 25 storey
building
Source: S.M. Hussain
5.15 A 3-D seismically isolated back-up data centre, Victoria University of Wellington. The
main module is behind the self-contained UPS generator to the left
to that building, by way of falling panels or other elements, could directly damage the isolated
building, or could fully or partially destroy the isolation scheme. Debris could fill the
horizontal separation gaps, preventing the superstructure remaining isolated. Even in the
absence of falling debris, the risk of the damaged building collapsing onto the isolated building
might be considered so high as to red placard both buildings until the vulnerable building is
stabilized (Figure 5.17).
Nearby buildings can affect other less damaged buildings even if they are further away,
Site coverage
Modern building codes require that buildings be set-back on their sites to avoid pounding
79
neighbouring buildings. Apart from along street frontages, buildings are to be separated far
80
Benefits and limitations of seismic isolation
5.16 A LoGlider bearing supporting the platform. The red mechanical components provide
vertical isolation
Isolated building
Elevation
5.17 A seismically isolated building should not be built adjacent to a more vulnerable
building. The tank could fall from the modern building to the left, or walls or parapets fall
from the unreinforced masonry building to the right, damaging the isolated building
5.18 Buildings in Christchurch cordoned off in the ‘red zone’ of the city following the 2011
earthquake
enough away from their boundaries so that when they sway sideways they won’t damage what
may lie beyond (Figure 5.19). The width of the separation gap increases both with the height
and flexibility of the building. If a building owner is dismayed at losing so much buildable
area on an expensive site, he or she can request a stiffer structure to reduce the width of the
gap. For a high-rise building, wide gaps at upper storeys may be easily achieved by designing
set-backs. But for seismically isolated buildings almost all of the movement occurs at the plane
Potential cost
block for some building owners. Chapter 9 discusses this topic in detail.
82
Small gap
Benefits and limitations of seismic isolation
Conventional
building
Site boundary
Large gap for
movement
Isolated building
Displaced
position
Elevation
5.19 The need for separation gaps between buildings and boundaries – conventional and
seismically isolated buildings
Maintenance
The structural systems of conventional buildings don’t require special maintenance. Usually,
given the absence of any moving parts or mechanisms, once constructed they can be virtually
forgotten. In contrast, as elaborated upon in Chapter 10, isolation devices and design
requirements such as movement gaps need rigorous periodic checking. Corroded or dirty
slider bearings equate to poor seismic performance, and if seismic gaps are filled by any later
construction or storage of materials, the isolation scheme is ineffective. And, if as expected,
the structural design has been optimized on the basis of an isolated superstructure, unforeseen
severe structural damage is a likely outcome during a large earthquake. Seismic isolation,
such as elevators and other mechanical equipment, is therefore suited only for building
owners willing to commit to regular inspections and maintenance.
Reduced effectiveness in small earthquakes
Seismic isolation is ineffective during very small earthquakes. During low levels of ground
acceleration a seismically isolated building behaves identically to a conventional building.
Only when the static friction within a sliding bearing is exceeded, or the lead plug in a lead-
rubber bearing yields, does the isolation system ‘give’ and begin working. Most recently this
effect has been measured in the Christchurch Women’s Hospital.11 This means that occupants
of seismically isolated buildings feel small sharp earthquake movements just like people in
conventional buildings. Once ground motions increase, movements in the isolated building
are far reduced in amplitude and are more gentle.
Notes
1 Arnold, C. & Reitherman, R., 1982, Building configuration and seismic design, New York: John Wiley.
2 Darama, H. & Zekioglu, A., 2013, ‘Implementation of seismic isolation in Turkey for continued functionality’,
Proceedings of the 13th World Conference on Seismic Isolation, Energy Dissipation and Active Vibration
Control of Structures, Sendai.
3 Cutfield M.R., Ma Q.T. & Ryan, K.L., 2014, ‘Cost-benefit analysis of base-isolated and conventional buildings:
a case study’, Proceedings of the 2014 New Zealand Society for Earthquake Engineering Conference, paper
O45.
4 Ryan, K.L., Erduran, E., Sayani, P.J. & Dhao, N.D., 2010, ‘Comparative seismic response of code designed
conventional and base-isolated buildings to scenario events’, Proceedings of the 9th US National and 10th
Canadian Conference on Earthquake Engineering, Toronto.
5 Whittaker, D. & Jones, L.R., 2013, ‘Design spectra for seismic isolation systems in Christchurch, New Zealand’,
Proceedings of the 2014 New Zealand Society for Earthquake Engineering Conference, Wellington.
6 Standards New Zealand 2004, Structural design actions NZS 1170:5, Standards New Zealand, Wellington, 20–1.
7 Pan, P., Zamfirescu, D., Hakashima, M., Nakayasu, N. et al., 2005, ‘Base-isolation design practice in Japan:
introduction to the post-Kobe approach’, Journal of Earthquake Engineering, vol. 9, no. 1, pp. 147–71, p. 165.
Introduction
I N the previous chapter, one of the pre-earthquake benefits noted was the increased archi-
tectural freedom seismic isolation offers. The architectural benefits of reduced inter-storey
drifts and possible lower levels of structural strength and ductility can be exploited in more
elegant detailing, and more slender and less regular structure. This chapter explores these
architectural benefits, and more, in greater detail. A series of case studies illustrates how
architects are beginning to take advantage of seismic isolation to achieve and reinforce a wide
range of architectural objectives.
The case studies are grouped in four categories of architectural concepts and qualities.
The categories have been chosen from a total of ten categories that encapsulate the most
prevalent architectural concepts and qualities in contemporary architecture.1 The categories
most relevant to the few examples of special seismically isolated architecture are: grounded –
floating, stability – instability, heavy – lightweight and simple – complex. Note that each
category is defined by two extremes within which a work of architecture can be positioned.
Some buildings discussed below reinforce concepts and express architectural qualities from
more than one category. Perhaps they could even be placed in another category altogether.
Our aim in categorizing buildings like this is not to definitively pigeon-hole them, but to facili-
tate an exploration of the variety and richness of architectural form aided by seismic isolation.
Grounded – floating
T HE case studies of this section lie towards the ‘floating’ end of the ‘grounded – floating’
spectrum. This is unsurprising. A building that expresses a sense of being grounded
appears to be strongly anchored or fixed to its site. Like a tree rising from the soil, its vertical
structure is continuous from the foundations upwards. Any visible horizontal seismic isolation
plane breaks that continuity. In many seismically isolated buildings though, the isolation plane
lies beneath ground level, out of sight. In this situation it is possible to have structure read
as grounded, as in the case of the Tod’s building.
At Tod’s Omotesando Building, Tokyo, the deciduous trees along its street frontage are
represented abstractly in the concrete perimeter load-bearing structure (Figures 6.1 and 6.2).
The six-storey superstructure rests on elastomeric bearings above a single basement level.
Wide ‘trunks’ at ground level become finer as they rise towards the roof, reducing in width
to slender ‘branches’. The criss-crossing concrete wall piers form a dramatic surface pattern
both inside and out, with exterior surfaces of the structural members flush with the glazing.
The structural challenges associated with designing such a unique structure in a highly active
seismic zone were lessened by incorporating seismic isolation. Not only could an irregular
structural form be realised without making individual members too massive, but the reduction
of inter-storey drifts allowed for simple flush-glazed detailing.
Architect Toyo Ito & Associates also exploits seismic isolation in the Tama Art
University Library. This building touches the ground lightly rather than appearing grounded
(Figures 6.3 to 6.5). The tiny feet of the arches belie the weight of the steel and concrete
structure and the library books they support. The library building also avoids the straight
lines of Tod’s – instead it celebrates the potential of arches to create architecture. Once again,
concrete is the dominant material, but at the Tama library, centrally located steel plates within
the 200 mm thick arches provide most of the structural strength.2 In structural terms, the
superstructure can be described as two levels of pinned-base steel portal frames, rather than
arches. The thickness of the steel plates increases greatly at the bases of the arches whose
delicate cross-sections are of solid steel. Another unusual feature of the building is its sloping
ground floor in the public areas. To cope with this, as well as a basement over the rear of
the site, the isolation plane steps under the ground floor and basement area. Not only does
the elastomeric isolation system enable a fineness of structure, particularly at the bases of the
arches (or portal frames), but small design-level inter-storey drifts enable flush glazing to
combine with surface structure to achieve very smooth planar and gently curved walls.
Having considered one seismically isolated building that appears as grounded as the
abstract tree forms on its façades, and another that appears to rest lightly upon the ground,
the remaining case-studies in this section unambiguously express how seismic isolation
provides architects with opportunities to ‘float’ buildings above their foundations on isola-
tion bearings. Seismic isolation and architectural form
First we visit the new main building at the Shimizu Institute of Technology campus,
Tokyo (Figures 4.7, 6.6 to 6.8). The four-storey superstructure, supported on six massive
circular concrete piers, is elevated above a ground floor car park. As well as supporting the
entire weight of the building, the piers function as vertical cantilevers in the event of horizontal
seismic forces. Substantial foundation structure beneath the piers withstands the tendency to
make them overturn when horizontal forces are transferred to them through the lead-rubber
bearings mounted at their tops. The corrugated surfaces of the black bearings show the
deformation of the layers of rubber between the internal steel plates. Red steel weldments collect
85
forces from the superstructure and transfer them evenly to the upper surfaces of the bearings.
86
Seismic isolation and architectural form
Because the piers are so widely spaced, four-storey deep steel trusses span between them
and collect forces from beams supporting the suspended floors. The diagonal members of
the perimeter trusses are exposed in places, such as on the front façade of the building. This
language of bracing is adopted by secondary or tertiary structure, such as around the main
elevators. The elevator structure cantilevers vertically downwards from the isolated
superstructure and disappears below ground, separated from the ground by a perimeter moat.
An escape stair, its structure concealed by cladding, also cantilevers down from above. Rather
than plunging into a moat, the stair stops short of the ground to create the necessary horizontal Seismic isolation and architectural form
isolation plane. During the 2011 Tōkohu earthquake a video camera captured the movement
of a bearing.3
For another example of seismic isolation conveying an impression of ‘floating’ we visit
the main building of the Emergency Management Centre, Foligno, Italy (Figure 6.9). Here
seismic isolation is celebrated and elegantly integrated with the overall architecture. The
squatness of the three-storey building plus its distinctive dome shape suggests a rigid building
with a low centre-of-gravity, well suited to its function as an emergency management centre.
The building sits on ten perimeter high-damping rubber bearings. Above each a curved
87
concrete rib rises to connect to a central circular core at roof level. An inner shaft that houses
88
Seismic isolation and architectural form
6.3 Tama Art University Library, Hachioji, Tokyo, Japan, Toyo Ito & Associates. Some of
the two main façades
vertical circulation is suspended beneath the first floor slab and, although completely
structurally separated from ground level, provides the main access to the building. The seismic
isolation system contributes to the experience of entering the building. First, a visitor walks
under a perimeter arch and between two bearings. Their significance is heightened by their
elevation on concrete plinths and how the curves of the arches above the bearings both direct
and force the eye to them. Then, after having passed through the threshold of arches, bearings
and plinths, you are surrounded by them, except for the circular entry core at the centre. The
sense of ‘floatation’ is enhanced by the lesser dimensions of the bearings, even though they
are protected by rounded covers, as compared to the bases of the double concrete arches. To
comply with fire egress requirements an exterior set of fixed-base stairs abut the dome via
cable-stayed bridges without compromising the dome’s seismic isolation.
An architectural reading of a building ‘floating’ is usually made only where the visual
scale of a building far exceeds that of the building’s means of support, such as bearings.
For both the previous examples you can stand under the first floor and see the entire
superstructure supported by bearings. But in many cases only a glimpse is possible. For
example at Te Papa, the National Museum of New Zealand, Wellington, visitors can descend
Seismic isolation and architectural form
6.4 The smoothness of the façade looking towards the front entrance
89
90
Seismic isolation and architectural form
6.5 Interior arches with their fine bases, just inside the main entrance
6.6 Shimizu Institute of Technology, Tokyo, Japan. Some of the piers and isolators
supporting the main building
Seismic isolation and architectural form 91
6.9 Emergency Management Centre, Foligno, Italy. Overall view including stairs
into a small basement viewing room where several lead-bearings are illuminated, and in other
buildings, such as the New Zealand Parliament Building, visitors can take a tour of the
basement where many bearings that were inserted during a seismic retrofit can be observed.
At the Okumura Memorial Museum, Nara, Japan, a taste of flotation is offered to museum
visitors. An elastomeric bearing placed between red painted base-plates is proudly displayed
through exterior glazing (Figure 6.10). Seismic isolation technology is treated as one of the
museum’s precious exhibits.
Stability – instability
Seismic isolation and architectural form
The project site exists only 4.6 km from the Hayward Fault, one of the most active
seismic faults in the world, making the request [for no damage in the Californian
Building Code’s ‘maximum capable’ earthquake] substantially more challenging.
Ground motions expected at the site could result from an earthquake of
magnitude 6.0 or more on the Richter Scale. The approach to design embodies
lightness (both visually and structurally) and luminosity, hence its name, the
Cathedral of Christ the Light. The structure is defined by sacred geometries and
designed for lightness by using seismic isolation to manage ground motions and
reduce the demand on the superstructure during an earthquake.4
For a seismically isolated building that reinforces architectural qualities at the other end
of the stability – instability spectrum we return to the Sendai Mediatheque, Japan (Figures 3.15
and 3.16). Chapter 3 explains how ductile steel moment frames in the basement constitute the
bespoke isolation system. By absorbing much of the earthquake energy in the moment frames
it is possible for the exposed fine and slender structure above ground to avoid seismic damage
while displaying instability (Figure 6.13). In interior spaces, sloping column clusters support
gravity loads. Four larger and braced cores located near the corners of the buildings not only
brace the unstable sloping clusters but resist the seismic forces within the superstructure.
Heavy – lightweight
L IKE the Cathedral of Christ the Light, the Cathedral of Our Lady of the Angels, Los Angeles
also replaced its predecessor, badly damaged in the 1994 Northridge earthquake (Figure
6.14). At the outset of the design the client requested that the cathedral ‘should survive even
a big quake, unscathed’.5 This was a challenge given the proximity of an active fault to the
site. The architectural response was to resort to concrete walls and seismic isolation. Structural
walls dominate the architecture and give a sense of heaviness and solidity. As for most
Seismic isolation and architectural form
6.11 The Cathedral of Christ the Light, Oakland, California, USA, Skidmore Owings &
95
seismically isolated buildings, the isolation system is totally concealed. The high walls are tied
together at roof level by a thick reinforced concrete diaphragm and the whole complex is
supported by almost 200 bearings. Most are high-damping rubber bearings, but slider bearings
are also used. Friction pendulum bearings support the four corners of the bell tower and have
been designed to withstand vertical tension forces resulting from seismic overturning
moments.
While there are indeed some examples of seismically isolated ‘heavy’ architecture as we
have just seen, it is more common to encounter isolated buildings taking advantage of the
lighter and more slender structure that seismic isolation can facilitate. The following four
case-studies illustrate this trend.
The design of Sony City, Tokyo, was led by both the architect’s concept for an expressed
perimeter diagrid structure integrated with a double skin façade, and the client’s request for
‘exceptionally high resilience in strong earthquakes’.6 Both of these requirements were realized
with seismic isolation. Given the inherent stiffness of diagrids against horizontal seismic
forces, without an isolation system the member sizes would have been ‘heavy and un-
economical’. After all, the building is essentially a large cube, with plan dimensions 100 m
by 70 m and earthquake forces are resisted only on the four sides of the building (Figures
6.15 and 6.16). The majority of the gravity loads are carried by internal steel frames. Seismic
Seismic isolation and architectural form
6.13 Sendai Mediatheque, Sendai, Japan, Toyo Ito & Associates. An ‘unstable’ column
97
cluster
98
Seismic isolation and architectural form
6.14 Cathedral of Our Lady of Angels, Los Angeles, USA. Rafael Moneo, architect, 2002.
Some of the heavy interior elements
isolation is provided by 200 elastomeric bearings, the majority of which are high-damping
bearings, and 40 viscous dampers located around the perimeter where they are most effective
in dampening torsional vibrations. Each of the heavily laden interior columns is supported
by four bearings.
Although far smaller than Sony City, the Prada Boutique Aoyama, Tokyo, is another
example of a perimeter diagrid or lattice structure integrated with the façade also resisting
most of the seismic forces (Figures 6.17 and 6.18). It is hard to imagine load-bearing walls
with a greater degree of transparency, yet still able to withstand both vertical and horizontal
forces. Not only does structure allow interior spaces to be flooded with light, the diamond-
shaped structural form synthesises with the five-sided crystalline architectural form. One
reviewer comments that the architects have noted that this was their first building ‘to forge
structure, space and facade as a single unit’.7
According to the structural engineer, the design concept and structural requirements
presented quite a challenge:
1 The diagonal lattice forming the outer lattice should form the structural framework while
being an integral part of the glass façade (avoiding the need for vertical and horizontal
members in the vertical planes).
Seismic isolation and architectural form
99
6.15 Sony City, Tokyo, Japan, 2006. Plantec Architects. The perimeter diagrid bracing
100
Seismic isolation and architectural form
6.16 At the base of the building the diagrid continues just at the corners, but additional
internal bracing is provided
6.17 Prada Boutique Aoyama, Tokyo, Japan, Herzog & de Meuron, 2003. The main entry
and the lower floors of the six storeys above ground level. The black glazing lines
delineate the diagonal structural members while the floor structure is de-emphasised by
a light-colour
2 The diagonal lattice members should be 250 mm in width and 300 mm in depth, includ-
ing the dimensions of the finish materials.
3 The interior vertical shaft should be made as small as possible (without large horizontal
rigidity), and atriums should be provided by floor openings on the 1st, 2nd, 4th and 5th
floors.8 Seismic isolation and architectural form
Of particular concern was the need to brace the building using a system that was not only
stiff and therefore needed to be stronger than usual, but also lacked ductility, leading to even
higher design forces. Seismic isolation was the only way to meet the dimensional limitations
of the architects and to avoid chunky bracing members which would have compromised the
architectural concept. The diagonal members are welded I-sections with varying web and
flange thicknesses. They are welded to cast steel joints which are also connected to perimeter
floor beams. The isolation plane is beneath the basement floor level. It contains 39 bearings
101
– a combination of lead-rubber bearings, rubber bearings and sliding bearings (Figure 6.19).
102
Seismic isolation and architectural form
6.18 The structural diagrid wall is most clearly visible from within the building
Core
Stairs
Plan
Section
movement and counteract seismic forces by sliding back and forth like a pendulum.’10 This
special type of ‘mid-level’ isolation proved effective in decreasing the seismic response of the
building for all the seven earthquake records used in the design of the building. Horizontal
displacements were not a huge issue due to the city council minimum separation between
building and boundary in the order of 500 mm. Without seismic isolation the dimensions of Seismic isolation and architectural form
the structural frames would have had to be significantly larger.
Light-weight is an apt description of the main space of the seismically isolated San
Francisco Airport international terminal. The terminal, underlain by approximately 40 m of
soft soils, is built on concrete piers above an existing road (Figure 6.23).11 Friction pendulum
bearings are placed on top of the piers and below the isolated superstructure. Above the
isolation plane, two floors accommodate arrivals and baggage handling. These lower floors
are braced by conventional structural systems that are protected from damage by the isolators.
Departures occur from the ‘Great Hall’. Its high enclosing glazed walls and 3-D roof trusses
103
that are fabricated from relatively small cross-sectional steel members create an atmosphere
104
Seismic isolation and architectural form
6.20 Nicolas G Hayek Center, Tokyo, Japan. Shigeru Ban Architects. 2006. Front façade
Seismic isolation and architectural form
105
6.23 San Francisco Airport international terminal, USA. A section of the main elevation
of lightness (Figures 6.24 and 6.25). The five lines of trusses are supported by 20 steel canti-
lever columns whose slenderness reflects the lack of roof-top seismic force amplification as
a result of the seismic isolation. The whole structure, including the steel box columns that
are clad to give the appearance of circular columns, was designed to a very high level of seismic
performance. All of the upmost structural elements will remain elastic (no structural damage)
during an earthquake with a 1000 year return period. Seismic isolation achieved the lowest
construction cost while meeting the required seismic performance.
Simple – complex
L IKE most conventional buildings, the majority of seismically isolated buildings possess a
regularity and simplicity of form. There is little point in illustrating such common
architectural forms. So the following case-studies, whichrange between simple and complex
forms, begin with a rectilinear form with a difference. The high narrow building has a
penetrated and stepped façade, injecting variety into its form. The Yoyogi Seminar School
building, also known as the Yozemi Tower, Tokyo, is 26 storeys high above three basement
levels (Figures 4.6, 6.26 and 6.27). Pairs of structural walls coupled by diagonal braces that
6.24 A view along the front of the building showing the cantilever columns
108
Seismic isolation and architectural form
6.25 The three-dimensional steel roof trusses that flood the Departures Hall with light
resist transverse horizontal forces are exposed at each end of the building. Interior braced
frames act in the longitudinal direction. The isolation system consists of bearings and
dampers. Columns bear on a combination of rubber and sliding bearings while damping is
provided by 12 semi-active and 12 passive hydraulic dampers. The semi-active dampers adjust
the level of damping during an earthquake to optimize seismic performance.12
The Inagi Hospital, Tokyo, has a far more complex plan. The irregular tower block form
is butterfly-shaped (Figures 6.28 to 6.30). Podiums at the front and the rear of the two wings
result in a rectangular ground floor plan. The whole plan area is seismically isolated on bearings
below the basement level. If this building were of conventional construction it is likely that
due to its re-entrant corners the tower would have been separated into three structurally
separate buildings – the two wings and the central core. Especially given the large penetrations
for vertical circulation through the central core floor diaphragms, a strategy of separation would
have made far more sense than trying to tie the three blocks together through the weakened
central area. That is the usual structural response to re-entrant corners.13 However, since the
whole hospital is seismically isolated, its inter-storey drifts and accelerations have been reduced
sufficiently for the designers to tie the superstructure together. Rather than vibrating as
three independent structures, it acts as a single unit. One of the big advantages of this approach
has been to avoid seismic separation gaps between the three towers and the subsequent
considerable simplification of the architectural and services detailing across the gaps. For
example, the need for floor, wall and ceiling cover plates are avoided, and services do not
Seismic isolation and architectural form
6.26 Yoyogi Seminar School building, also known as the Yozemi Tower, Tokyo, Japan.
109
require any flexible connections to accommodate the relative horizontal movements between
blocks. Also, the connections between the podiums and the towers have been made a lot
easier. Without isolation, the podiums would have had to be tied to one block and seismically
separated from the others. Alternatively, they could have been designed as free-standing
structures with their own seismic resistance and separated from all three towers. The elegance
of the seismically isolated solution is highlighted when these alternative design approaches
are considered.
In the final example, seismic isolation is the means of achieving architectural complexity.
This quality of the Delegation of the European Union to Japan, Tokyo, is apparent on the
main façade notably its randomly placed and sized openings. The seismic isolation scheme
comprises natural rubber laminated bearings and sliding bearings. Six oil dampers acting in
each orthogonal direction provide restraint against wind, provide damping and control Seismic isolation and architectural form
torsional deformation during an earthquake that arises from the 3-D structural complexity.14
According to T. Mizutani (personal communication, 15 August 2014), due to seismic isolation,
which greatly reduced the seismic design forces to approximately 25 per cent of what would
normally be designed for, it was possible to design and detail the very complex perimeter
and internal reinforced concrete walls satisfactorily. Nonetheless, the highly penetrated
walls and other structural complexities such as some walls being off-set in plan, required
sophisticated computer modelling. The wall reinforcement was able to be detailed bearing
in mind the need for construction practicality.
111
112
Rubber bearing
1 Charleson, A.W., 2014. Structure as architecture: a source book for architects and structural engineers, 2nd edn,
Oxford: Routledge, Chapters 12 and 13.
2 Toyo Ito & Associates, 2007, ‘Tama Art University Library’. Japan Architect, no. 67, pp. 120–29.
3 Shimizu Corporation, ‘Effects of seismic isolation.’ Available from: www.shimz.co.jp/english/theme/
earthquake/effect.html (accessed: 14 August 2014).
4 Sarkisian, M., Lee, P. & Long, E., 2011, ‘A celebration of structure as architecture – The Cathedral of Christ
the Light’, Proceedings of the ASCE, Structures Congress, Buildings section, pp. 971–82, p. 971.
5 Rosta, P. & Post, N., 2001, ‘L.A.’s third millennium Noah’s ark’, Engineering News Record, vol. 42, no. 5,
p. 42.
6 ‘Sony City, Tokyo: a diagrid combined with base isolation’, The Arup Journal, 2009, no. 2, pp. 49–51, p. 51.
7 Herzog & de Meuron 2004, ‘Prada Aoyama Epicenter’, A+U: Architecture and Urbanism, vol. 406, no. 7,
pp. 78–83.
8 Nakai, M., 2008, ‘Unique architectural forms enabled by base-isolation’, Proceedings of the 14th World
Conference on Earthquake Engineering, Beijing.
9 ‘Nicolas G. Hayek Center, Tokyo’, The Arup Journal, 2009, no. 2, pp. 52–4, p. 52.
10 Pollock, N.R., 2008, ‘Shigeru Ban sets showrooms and facades in motion at the Swatch Group’s Nicolas G.
Hayek Center in Tokyo’, Architecture Record, no. 5, pp. 200–05, p. 202.
11 Mokha, A.S. & Lee, P., 1999, ‘Wings of isolation: San Francisco International Airport’s new terminal is protected
by 267 steel seismic isolators’, Modern Steel Construction, October.
12 EERI reconnaissance team 2012, ‘Performance of Engineered Structures in the Mw 9.0 Tohoku, Japan,
Earthquake of March 11, 2011’, EERI special earthquake report, pp. 1–16, p. 12.
13 Charleson, A.W., 2008, Seismic design for architects: outwitting the quake, Oxford: Elsevier, pp. 132–3.
14 Mizutani, T., Hayabe, Y. & Yoshikawa, H., 2013, ‘Large structural wall system with random openings realized
by seismic isolation system’, Proceedings of the 13th World Conference on Seismic Isolation, Energy Dissipation
and Active Vibration Control of Structures, Sendai.
Retrofitting
Introduction
Since this first seismic isolation retrofit project, hundreds of others have been
undertaken in seismically prone areas around the world. Some projects have been even more
impressive in terms of architectural monumentality and scale than the Salt Lake City and
County building. Others have little architectural merit and are retrofitted for other reasons
that are listed in Chapter 2. The two most significant advantages of this retrofitting technology
are first, being able to use the building during the retrofit process, and second, keeping the
superstructure retrofit interventions to a minimum.
It is often inconvenient and costly to vacate a building during a retrofit but, with the
choice of seismic isolation, many buildings can be occupied during construction. In Japan,
approximately 70 per cent of seismically isolated retrofitted buildings continue to be occupied
during the retrofit process.2 Heritage buildings, in particular, benefit from seismic isolation
due to the need for minimal strengthening work to their superstructures. For example, at the
Retrofitting
Salt Lake City and County building, major new structural components were needed only in
the clock tower.
The next section discusses options to be considered when deciding upon the height or
115
level of the isolation plane in a retrofitted building. Then we present several case-studies of
116
completed projects, dividing them into two groups – unreinforced masonry buildings, and
then more recent column and beam framed buildings. The chapter concludes by discussing
Retrofitting
T HE height of the isolation plane above the top of existing foundations depends on
a number of factors, such as the type of construction, and whether or not there is a
basement. Isolation plane heights tend to be lower in unreinforced masonry buildings
compared to frame buildings. Masuzawa and Hisada report that in the 90 isolated retrofitted
buildings they studied, 40 per cent had the isolation plane near the existing foundations,
and in the remainder the isolators were placed in between floors.3 Where located between
floors, 40 per cent were in basements, 40 per cent in ground floors, and the remainder at
higher levels.
In existing load-bearing wall buildings, mainly constructed from unreinforced masonry,
isolation bearings are placed under and within walls. Slots are cut through the walls where
bearings need to be placed. The use of discrete bearings raises two structural challenges.
First, new beams must pick up the vertical loads from the walls above the bearings and trans-
fer them to the bearings. In most cases ‘sandwich’ beams, cast on both sides of the walls
and then clamped together with tensioned rods, as at Salt Lake City, perform this role.
If, for aesthetic or heritage reasons, this solution is inappropriate, it is possible to cast a
new beam, section by section, directly under an existing wall but this procedure requires
extensive propping of the wall and is to be avoided if possible. Just as beams transfer forces
from above into the bearings, similar beams are usually required under the bearings to safely
distribute the point loads from individual bearings to the lengths of foundation walls
underneath them.
A basement makes the isolation of a load-bearing wall building much more feasible.
As shown in Figure 7.2, if a building is founded on shallow foundations it might be necessary
to raise the level of the ground floor to accommodate the depth of the upper sandwich beams
(approximately 1.0 m) and the height of the bearings (between 300–600 mm) as well as create
a crawl space for construction and maintenance access. If the foundations are deeper, the
bearings and the two layers of sandwich beams can be constructed below the existing ground
floor but this requires a large volume of excavation. With a basement, the construction of
the beams, especially alongside interior walls, is so much easier. The upper sandwich beams
are normally cast near the top of the basement walls. This minimizes bending of the upper
walls when the bearings are displaced normal to the walls. New concrete wall jackets may be
required either side of the basement walls to withstand far greater bending from the same
horizontal forces acting normal to their lengths. Existing internal masonry or concrete
columns in the basement will have bearings placed above them and they will also require
New retaining wall
for horizontal clearance Existing floor and beam
New floor
Original Basement
Ground level floor level
Bearing
Section Section
(a) (b)
7.2 The advantage of a basement in isolating an existing building is due to the extra
height in which to create the isolation plane. In (a) construction conditions are very
cramped, while in (b) the basement makes construction easier
jacketing and possible enlargement of their footings to resist the bending at their bases.
Figures 7.3 and 7.4 show typical construction details.
Retrofitting concrete or steel frame buildings by seismic isolation is easier than for load-
bearing wall buildings. There is far less structure to isolate and there is usually no need for
beams to transfer loads from masonry into and out of bearings. First, a decision must be
made concerning the level of the isolation plane. We need to bear in mind the horizontal
plus or minus 300 to 500 mm design movement in the isolation plane and its implication for
elements, such as elevators, stairs, non-structural walls and services, crossing it. More
information about treating these elements is given in Chapter 8. If we assume the isolation
plane will be somewhere within the ground floor level there are three options (Figure 7.5).
Insertion of bearings at column bases is less common due to the high degree of bending in
the column (and beam) at first floor level. In many situations this bearing placement option
requires significant structural strengthening of the first floor structure. Mid-height isolation
is more common. The existing column, beam and foundation strengths may be sufficiently
strong. If not, bearings might be best placed just below the first floor beams. Now the bending
Retrofitting
demands on the first floor beams and columns are as low as possible. However, the ground
floor columns and their foundations may need upgrading to withstand the greater bending
at foundation level. The construction sequence for isolating an existing frame building is
117
Masonry
Masonry
Retrofitting
Foundation footing
Section Elevation
(a)
Bearing
Section Elevation
(b)
7.3 Typical construction details when inserting isolation bearings in a load-bearing wall
without a basement. (a) New sandwich beams are constructed and an opening for the
bearing is made, and in (b) the bearing is inserted, pre-loaded by a flat jack to carry the
loads from the wall
New or strengthened Temporary steel collar
floor system clamped to pier
Flat jack
Bearing
Strengthened
pier base
Temporary prop
and foundation
Section Section
(a) (b)
7.4 Typical construction details when inserting isolation bearings under a masonry column
or pier. After strengthening the pier base and forming a strong slab-and-beam floor (a), the
pier load is transferred through temporary props so the pier can be cut and a bearing
inserted and then loaded with a flat jack (b)
RC frame
Bearing
Floor level
Elevation
7.5 Three potential levels of bearings and isolation planes for an isolated retrofitted
moment frame building
Retrofitting
119
120
This was the first unreinforced masonry building in Japan to be retrofitted using seismic
isolation (Figure 7.6). The main body of the building, constructed from load-bearing brick
walls and interior steel posts and beams, was completed in 1906. In 1929 it was extended
using reinforced concrete frames with brick infill.4 2002 marked the completion of the most
recent extension, refurbishment and retrofit. The renovation of this designated historic
building required upgrading of its seismic, fire and egress performance. Due to its heritage
value it was also necessary to preserve existing exterior and interior architectural finishes
and features. Seismic isolation was the solution with least impact upon the existing fabric.
As well as attending to preservation and refurbishment, Tadao Ando Architect and Associ-
ates introduced several unmistakably modern glazed volumes that contrast with the heavy
existing construction. These include the new glazed entry and a light-filled multi-storey rear
extension between two concrete cores (Figures 7.6 and 7.7). These additions not only satisfy
7.6 The International Library of Children’s Literature, Tokyo, Japan. The main façade with
the new glazed entry
7.7 The new rear façade of the library that comprises two reinforced concrete cores and
Retrofitting
glazed walls
121
122
Retrofitting
7.8 A circulation and gathering space between the rear existing façade and the new
glazed wall
seismic and fire requirements but create attractive spaces contributing to improved building
and functionality.
All existing and new construction is isolated by natural rubber bearings and separate
lead dampers. The isolation plane generally is located below the ground floor beams but also
steps down under the bases of the two concrete cores (Figure 7.9).5 The construction details
are typical of retrofitted load-bearing wall isolated buildings as discussed above. Due to the
two different structural materials and systems in the existing building, and the addition
of two concrete cores, a new reinforced concrete ground floor, and slabs at other levels, create
strong diaphragms that tie the different structural systems together. The building reacts
against earthquake forces as a single unit.
Concrete core
Glazed entry
Isolation plane
Retrofitting
7.9 A simplified section through the retrofitted building showing the varied isolation plane
levels
123
124
The historic City Hall was constructed during the 1810s in a neo-classical Viennese style
(Figures 7.10 and 7.11).6 It was damaged during the 1977 Bucharest earthquake, even though
the intensity of shaking at Iasi was quite low. Analyses indicated the likelihood of severe
damage during the design earthquake, as defined by the current Romanian seismic code.
Typical of monumental buildings of its period, it is built of many interconnected unreinforced
masonry walls on stone masonry foundations that support the reinforced concrete floor slabs
that have replaced the original wooden flooring. The walls are very thick. Exterior walls vary
in thickness from 1.0 m to 1.7 m, while interior walls reach 0.6 m thick.
Rather than attempting to retrofit the building using conventional techniques, like
inserting new structural elements such as structural walls, it was decided to seismically isolate
the building. This solution has avoided the need for any additional superstructure streng-
thening which would have greatly affected the heritage values of the building. By founding
the building on a combination of lead-rubber and sliding bearings, the seismic accelerations
experienced in the superstructure walls are reduced enough to prevent damage during the
design earthquake.
Figure 7.12 shows a foundation plan with the many new bearings positioned along walls
and at wall intersections. The horizontal isolation plane is just below ground floor level. To
distribute gravity forces from the walls into the bearings and then from the bearings into the
stone masonry foundation walls below, pairs of permanent shoring (sandwich) beams were
placed above and below the isolation plane (Figure 7.13). After casting the beams, which are
clamped to the walls by post-tensioned steel rods, openings were made for the bearings
and then a horizontal isolation plane gap created along the remaining lengths of wall. A new
reinforced concrete diaphragm slab was cast at ground level under the building to tie the
foundation walls together and to prevent the new perimeter retaining walls from sliding into
the basement space.
Retrofitting
7.12 A foundation plan showing the large numbers of walls and the placement of isolating
bearings
125
7.13 Details of the pairs of beams required above and below the bearings
Source: Miyamoto International
As Ron Mayes explains, there were several noteworthy aspects to this project completed
in 1991:
It was the first base-isolated building retrofit in California, and also the first office
building in the USA to be seismically rehabilitated using base isolation while being
completely operational. It was also the first framed structure to be retrofitted with
base isolators at the mid-height of the bottom-story columns, which avoided
disruption to the building occupants during the construction phase.7
The original eight-storey building was built in the mid-1960s. Waffle slab floors
are supported by reinforced concrete columns and perimeter moment frames. A 1982
structural review of the non-ductile building, confirmed by other reviews five years later,
indicated the building possessed only one quarter of the seismic strength required by the then
current codes.
Four retrofit schemes were ultimately selected for analytical comparison: (1) base
isolation plus exterior diagonal bracing; (2) conventional diagonal braced frames
on the exterior; (3) exterior shear walls in the perimeter frames; and (4) jacketing
of the non-ductile concrete beams and columns.8
The base isolation plus strengthening scheme was eventually selected. Although it was more
expensive than the other schemes it met the owner’s directive: ‘The building must survive an
expected major earthquake with no downtime for the building systems, contents and
occupants [. . .]. Structural damage must be limited, and minor in nature. The elevators must
resume operation promptly.’9
Prior to retrofitting, the recommended exterior diagonal bracing was replaced by
stepping reinforced concrete moment frames that far more satisfactorily integrate with the
original architecture of the building. In fact, they possibly enhance it. And due to the way
they express the need for increased seismic strength towards the base of a building, they are
a good example of ‘earthquake architecture’.10
Normally, the isolating bearings would have been inserted into the basement columns.
However, sensitive equipment at that level could not be disturbed, so the isolators were placed
in the ground floor. Steel yokes were clamped to the existing columns and hydraulic jacks
supported the loads while blocks were cut and removed, and new lead-rubber bearings
inserted and fixed in place. The new moment frames were connected to the existing structure
by thousands of epoxy-grouted reinforcing bars, the holes of which were drilled at night. Now
these moment frames on all four sides of the building above the bearings resist the vast
majority of the seismic forces that pass through the base isolation system. The columns
beneath the bearings extend down to basement level and are supported by cast-in-place piles.
Since the isolation plane is above ground level it proved a challenge for the two elevator
cores to service the basement while passing through the isolation plane and coping with
displacements within it. The solution was to support the elevators at first floor level and
provide movement gaps at ground and basement level. Vertical support was provided at the
base of the elevator pit by Teflon bearings. Now, during an earthquake, the elevators move
freely with the superstructure relative to the ground and basement floors.
This building houses the Victoria University of Wellington library collection (Figure 7.14).
It was constructed in the early 1960s in reinforced concrete. Positioned on a sloping site, it
has two basement levels over some of its length, a two-storey podium and an eight-storey
Retrofitting
tower. Its floors consist of precast waffle slabs post-tensioned in two directions through
columns to create two-way moment frames. The three ribs that meet the column faces com-
prise the moment frame beams. There are eight transverse frames and two longitudinal
127
Prior to retrofitting, a review of the building’s seismic vulnerability concluded that after
a design-level earthquake demolition would be more economic than repair. The university
was also aware of how its whole viability depended on the continuous operation of the library,
the possibility of a major earthquake in the Wellington area, and how insurance premiums
were increasing sharply.11 After considering the cost and benefits of different retrofit strategies,
seismic isolation was adopted. Its two most attractive features were that no retrofitting work
was required above the basement levels, and that the library could continue to operate without
significant disruption throughout the duration of the retrofit. The client was also able to
negotiate significant earthquake insurance premium deductions that covered the cost of the
retrofitting after a relatively short time-frame.
The retrofit project followed the normal steps necessary for isolating an existing
frame building. During the project, the lowest basement area was extended the whole length
of the building. This gave access to the base of the columns, which were previously part of
the foundations, and provided sought-after additional floor space. It meant that the southern
area of the podium required temporary support (Figure 7.16). One by one, primary columns
were propped to completely remove any compression force at the lowest basement level.
7.15 A cross-section through the Rankine Brown building
Then a wire-saw, with diamonds embedded in the continuous wire, made two cuts through
the concrete and the reinforcing steel of the columns. Each cut took approximately four hours.
The concrete blocks were removed and new lead-rubber bearings inserted, grouted and bolted
to the columns (Figures 7.17 and 7.18). It was fortunate that due to the combination of the
Retrofitting
basement low inter-storey height and the strength of the existing columns and foundations,
these structural elements did not require any additional strengthening.
Lead-rubber bearings were placed only under the 16 primary columns. Smaller concrete
129
columns supporting the podium at the lowest basement level were pinned top and bottom.
130
Retrofitting
7.16 The propping of the podium during the excavation to existing foundation level
They will tilt or rotate when the separated superstructure above them moves up to the
maximum displacement of plus or minus 600 mm. Existing floor-to-retaining wall
connections were replaced by gaps covered by steel plates able to slide above the retaining
walls. At the main entry level these plates are covered by paving stones. Evidence of the
isolation system is generally hidden, but in other areas the cover plates are exposed and
the isolation plane is visible on close inspection (Figures 7.19 and 7.20).
Even though it opened as late as 1959 this building is of considerable historic significance,
being one of the works of architect Le Corbusier (Figure 7.21). In part, due to the application
of some of his ‘five points of a new architecture’ such as pilotis and the free plan,12 seismic
weaknesses became apparent and were addressed by a seismic retrofitting programme that
was completed in 1997. Obviously, it was necessary for any retrofit solution to have minimal
impact upon the existing architecture, and so a seismic isolation scheme was adopted.
The original building is of reinforced concrete construction. It consists of three floors
above ground and a partial basement. It is square in plan with seven rows of seven columns
at 6.3 m centres (Figure 7.22). Unlike the first two examples of retrofitting a frame building,
the isolation scheme was implemented by excavating beneath each column to cast a new
foundation pad and insert a bearing between the new and existing pads (Figure 7.23). A high-
damping rubber bearing is located under each column. Because of seismic isolation only a
Retrofitting
7.17 Three props on each side carry the column load as a new bearing is inserted into the
131
7.20 The isolation plane, only 20 mm thick, runs along the bottom of the lower black
coloured wall band. Vertical sacrificial flashing near the corner of the wall hides the rattle
space
small amount of structural improvement was necessary to the superstructure. The ground
floor columns were increased in diameter from 530 mm to 600 mm, and those on first floor
level required similar strengthening to achieve the desired level of seismic performance. New
retaining walls were constructed around the site and adequate separation gaps provided
between the main superstructure and adjoining buildings, including the external stairs
adjacent to the main façade. The separation details are so well resolved that visitors to the
museum have no idea that the building is seismically isolated (Figure 7.24).
7.21 The main entrance to the National Museum of Western Art, Tokyo, Japan
This final case-study illustrates a very innovative form of seismic retrofit where a building is
retrofitted by adding a two-storey isolated block on top of an existing building.
The client wanted to expand a 1988 three-storey office building but the existing structure
could only accommodate a single-storey vertical addition without seismic retrofit. The
solution consisted of introducing seismic isolators at the level of the existing roof and then
another two storeys of office accommodation was constructed above them (Figures 7.25 and
7.26). These two additional floors above the isolators provide a mass damping effect due to
the lower and the isolated storeys vibrating out-of-phase. This sufficiently reduces the seismic
demand in the existing structure so that the need for extensive strengthening of the occupied
building was avoided. China Basin Landing is the first building in the USA to have a
Retrofitting
7.22 Several interior columns near the perimeter of the building in a gallery
7.23 A cross-sectional model of the isolated museum displayed in the museum. Note the
varied levels of the isolation plane
7.24 A seismic separation gap was created between the external stairs and the main
building
Limitations
retrofitting by seismic isolation. It’s not surprising then that designers and contractors are
always trying to develop easier and more economic construction methods. In one new approach
developed in Italy, a new bottom foundation concrete slab is cast around the existing foun-
137
dations. Then another slab, which is strongly connected to the foundations, is cast on top.
138
Retrofitting
7.25 China Basin Landing office building, San Francisco, USA. The existing three-storey
1980s building was retrofitted by placing isolators at roof level and then constructing
another two storeys
Source: Dynamic Isolation Systems Inc.
The foundations are cut just below the upper slab which is then jacked up so that isolating
bearings can be inserted between the two slabs.13
Although a retrofitting scheme comprising seismic isolation may be the preferred
method of retrofitting, it may not be feasible due to the need for wide seismic separation gaps
between the retrofitted building and its neighbours. If a building intended to be seismically
isolated is built close to or against its site boundary then seismic isolation is probably not a
feasible solution. Imagine how unrealistic it would be to demolish say a 400 mm wide
perimeter strip from the floor plates of a building. The task would be daunting even for a
building like the National Museum of Western Art with its perimeter columns set in from
the façades. The need for wide separation gaps between adjacent buildings is therefore one
of the most common factors preventing widespread adoption of seismic isolation as a seismic
retrofitting strategy.
7.26 China Basin Landing office building during construction
Source: Dynamic Isolation Systems Inc.
Notes
1 Allen, E.W. & Bailey, J.S., 1988, ‘Seismic rehabilitation of the Salt Lake City & County building using base
isolation’, Proceedings of the 9th World Conference on Earthquake Engineering, Tokyo-Kyoto, vol. 5, paper
7-9-5, pp. 639–44.
2 Masuzawa, Y. & Hisada., Y. 2012, ‘Current state of retrofitting buildings by seismic isolation in Japan’,
Proceedings of the 15th World Conference on Earthquake Engineering, Lisbon, p. 3.
3 Ibid. p. 4
4 National Diet Library 2006, ‘Architecture of the International Library of Children’s Literature: preservation
and renovation of an old brick building from the 1900s’, Conference of Directors of National Libraries in Asia
and Oceania (CDNLAO) Newsletter, no. 55, p. 1.
5 Tasaka, M., Mori, N., Yamamoto, H. et al., 2008, ‘Applying seismic isolation to buildings in Japan – retrofitting
Retrofitting
and middle-story isolation’. Proceedings of the 18th Analysis and Computation Speciality Conference, ASCE,
Vancouver, p. 4.
6 Glilani, A.S. & Miyamoto, K.H., 2102, ‘Base isolation retrofit challenges in a historical monumental building
139
7 Mayes, R.L., 2007, ‘R.L. Mayes’, Structural Design of Tall and Special Buildings, vol. 16, no. 1, pp. 3–36, p. 22.
8 Gates, W.E., Nester, M.R. & Whitby, T.R., 1992, ‘Managing seismic risk: a case history of seismic retrofit for
Retrofitting
a non-ductile reinforced concrete frame high rise office building’, Proceedings of the 10th World Conference
on Earthquake Engineering, Rotterdam: Balkema, pp. 5261–6, p. 5263.
9 Dooley, C.T. & Robison, R., 1990, ‘Seismic surgery’, Civil Engineering, ASCE, vol. 60 no. 9, pp. 72–5.
10 Charleson, A.W., 2008, Seismic design for architects: outwitting the quake, Amsterdam: Elsevier, p. 251.
11 Clark, W.D. & Mason, J.E., 2004, ‘Base isolation of an existing 10-storey building to enhance earthquake
resistance’, Proceedings of the New Zealand Society for Earthquake Engineering Technical Conference,
Rotorua, paper 10, p. 1.
12 The National Museum of Western Art, ‘Discover Architecture Map’. Available from: www.nmwa.go.jp/jp/about/
pdf/discoverarchitecturemap_en.pdf (accessed: 22 September 2014).
13 Briseghella, B., Zordan, T., Zambianchi, L. et al., 2012, ‘Lift-up and base isolation as a retrofit technique for
R.C. existing building’, Proceedings of the 15th World Conference on Earthquake Engineering, Lisbon.
– 8 –
Design aspects
Introduction
T HIS chapter focusses upon seismically isolated buildings with an emphasis upon their
unique design requirements as compared to the design of conventional buildings. We
assume that architects are responsible for architectural form which is developed in
collaboration with structural engineers for both types of buildings. After discussing the special
requirements for the composition of and the effectiveness of the design team, such as the
need for more intensive communication and collaboration, we turn our attention to how
design for seismic isolation impacts structural engineers. Then, in the major section of the
chapter, we discuss the full range of architectural design implications of a seismically isolated
building. These cover conceptual architectural design to construction detailing.
Design team
T HE successful introduction of any new technology into a design and construction project
necessitates additional communication and a higher level of collaboration between
members of the design team. Seismic isolation is no exception. Even before making a final
decision to isolate, numerous meetings and briefing sessions involving client, architect and
structural engineer are to be expected. In the design of a conventional fixed-base building,
codes of practice are followed to ensure a ‘code compliant’ design. Implicit assumptions
embedded in codes regarding seismic risk and performance are usually accepted uncritically
by the design team. However, if seismic isolation is a possibility, engineers need to brief the
Design aspects
architect and the client on matters that are mainly only discussed by committee members
developing seismic-related codes.
Once a decision to seismically isolate is made, the intensity of collaboration increases.
Seismic isolation brings with it additional design and detailing issues needing resolution. For
141
example, architect and structural engineer together have to decide how and where to isolate.
142
Will the isolation plane be above ground or within a basement? Then, after running numerous
dynamic analyses, the engineer needs to provide information about the maximum horizontal
Design aspects
displacements along the isolation plane to be designed and detailed for. This information is
additional to values of superstructure inter-storey deflections that are always needed to provide
adequate clearances between structure, such as columns, and non-structural components, such
as partition walls. Estimates of residual displacements of the isolation system after a major
earthquake, which may affect various details such as cover plates over seismic separation gaps,
also need to be determined by the engineer and communicated to the architect. All this
additional complexity requires greater interaction within the design team, as well as
determination within it to work together for the sake of a successful project.
• meeting special structural design requirements and using sophisticated, infrequently used
and time-consuming design tools such as inelastic time-history analyses;
• participation in independent design reviews;
• designing additional components such as the bearings, their fixings, suspended lift-pits,
movement and separation details, and retaining (moat) walls;
• interaction with specialist consultants such as seismologists, geotechnical engineers
and bearing manufacturers;
• greater involvement with services engineers and architects in the isolation plane area;
• reviewing bearing performance test results and possibly observing bearing tests; and
• on-going monitoring of the isolation system as part of its maintenance schedule (see
Chapter 10).
Due to the extra work undertaken by the structural engineer it is reasonable to expect a
commensurate increase in professional fees.
One of the most important items of information the structural engineer provides to the
architect is the width of the seismic or movement gap. Its purpose is to allow horizontal
movement between the isolated superstructure and the foundations, retaining structures or
adjacent buildings above the level of isolation plane. Cutfield and Ma conclude their cost-
benefit study of an isolated and a conventional building by noting:
Overall, the performance of the isolated building models was far superior to the
conventional building model; however, this performance degraded somewhat in
the unlikely event of structural pounding against the building’s moat wall. Hence,
considered and conservative selection of the building seismic gap is important
for achieving the best performance from a base isolated building.8
walls might be a source of pounding, they can also prevent elastomeric bearing instability.
Wide gaps also mean greater clearance around the perimeter of the building. This might
reduce useable floor area and increase the possibly undesirable visual impact of architectural
details covering the gaps.
143
144
Architectural design
Design aspects
S EISMIC isolation increases the architect’s as well as the structural engineer’s workload. At
both conceptual and final detailing levels of design, many design scenarios require
addressing that are irrelevant for conventional buildings. Yet, while detailing for the large
movements across an isolation plane can be challenging, the significantly reduced movements
within an isolated superstructure can simplify detailing elsewhere.
As mentioned at the start of this chapter, early collaboration between architect and
engineer is especially important for the design of a seismically isolated building. Many
conceptual aspects of the isolation scheme require discussion and resolution in those first
few design team meetings. The decision regarding the height of the isolation plane has already
been mentioned, but there are many others. For example, should a bearing be placed under
each column, or are there advantages in providing transfer beams to reduce the numbers
of bearings? How might that decision help or hinder the provision of the minimum
recommended 1.2 m crawl space height for inspection and maintenance underneath the
beams protruding into that space (Figure 8.1). Or, if there are several multi-storey blocks to
be isolated, should they have separate isolation systems, or might founding them on a common
isolated slab be a better solution (Figure 8.2)? Also at the conceptual stage of design,
preliminary ideas of key details benefit from a collaborative approach. Strategies for achieving
the horizontal clearances for movement along the isolation plane and up the perimeter of
a building need resolution, and these will ultimately affect the built floor area (Figures 8.3
to 8.5).
In Japan, the recommended clearance to protect mainenance staff from being crushed
during an earthquake is 200 mm wider than the design displacement unless the area immed-
iately adjacent to the structure is a walkway. In this case the additional clearance should be
800 mm to protect the public. Signage is sometimes provided to warn people of the danger
of the gap narrowing during an earthquake.12
Later in the design process detailed design is undertaken, but before considering typical
architectural details, observations after the 2011 Tōhoku earthquake, Japan, indicate that their
performance left a lot to be desired. Although in principle the provision for unrestrained and
damage-free movement is straightforward, it is more difficult to achieve in practice. Saiki and
others’ survey of over 300 isolated buildings revealed that 30 per cent of the buildings
experienced damage to movement joints.13 Even though the movements were in most cases
far less than the maximum design displacements, joints did not function as intended by their
designers. Defects were observed due to the location of joints, obstructions in their immediate
vicinity and lack of maintenance. The authors suggest a pragmatic and sensible approach to
specifying adequate performance of movement details. They accept minor damage in locations
other than evacuation routes and acknowledge different degrees of damage are appropriate
for different earthquake intensities. After defining four levels of damage (Table 8.1), they
propose movement joints are classified and specified by a three level Performance Index (Table
8.2). These tables are based on those proposed in their paper.
(a)
Transfer
beam
Transfer
beam
Design aspects
(b)
8.1 A bearing is usually placed under each column (a), but a transfer beam reduces the
numbers of bearings while maintaining a crawl space (b)
145
146
Design aspects
(a)
(b)
8.2 Several separate buildings can be individually isolated (a), or share a single isolated
slab (b)
Site boundary
Vertical
clearance
Crawl
space
Moat
Table 8.2 Categories of movement joint performance for two earthquake intensities,
different joint locations and performance verification methods
These details apply around the perimeter of seismically isolated buildings. As their names
suggest, they allow unrestrained movement of the isolated superstructure relative to the
ground in any direction. As shown in Figures 8.6 and 8.7, the main requirement is to achieve
the primary horizontal movement gap and then to cover it, to allow access in some areas,
and to prevent ingress of water, rubbish or snow. The width of the horizontal gap, which is
provided by the structural engineer, is dependent upon many factors. Typically, its width is
in the order of 300–500 mm, but can become as wide as 900 mm.15 The moat cover can
consist of a cantilever slab, hinged slabs or steel plates. It usually connects to the isolated
structure and rests on the top of the retaining wall, simply sliding along it and over it (Figures
8.8 to 8.10). Steel plates are particularly common moat covers at entrances of buildings and
across vehicle ramps (Figures 8.11 to 8.13).
There are many instances where moat covers or movement joints abut exterior paving
where the level of paving and the surface of the isolated ground floor align. In these cases a
typical detail employs angled sliding surfaces that prevent compression occurring in the cover
plate when the building moves towards the paving (Figures 8.14 to 8.18). Several companies
manufacture and supply cover plates for floors, both exterior and interior. Information
published on the internet can be consulted. The wide range of cover plates includes those
that self-centre (Figures 8.19 and 8.20). When installed correctly and paved, they can be almost
invisible (Figure 8.21), although it can be convenient to use cover plates that allow water to
enter a drain that is incorporated into or near a moat (Figure 8.22). Internal cover plates
utilize the same principles as those discussed above. They are also available in many different
forms, ranging from simple highly visible surface-mounted plates to those that are more
Design aspects
concealed (Figures 8.23 and 8.24). They may incorporate additional complexity to provide
fire resistance across the gap.
In some projects the expense of specifying relatively sophisticated and expensive cover
plates may be queried due to the low risk of damage and the probable ease of post-earthquake
151
repair. Some clients might find it acceptable to use very simple cover plates with minimal
152
Design aspects
8.6 Horizontal clearance between an external stair to the left and an isolated building to
the right, Tokyo, Japan
Steel cover plate
Drainage
Moat
8.7 A section through a typical moat, a perimeter column and bearing where the moat is
covered by a cover plate connected to the isolated structure and able to slide on the
retaining wall beneath it
movement allowance, but protected against falling with restraint wires in the event of large
displacements (Figure 8.25).
As well as moat covers and other flooring movement joints, capacity for seismic movement is
also required where an isolated building connects to another building, be it isolated or not, at
the roof, up walls and along and across ceilings. Just as proprietary covers are available for
moat and floor seismic gaps, so it is with ceilings, roofs and walls. However, it is relatively easy
Design aspects
to create movement joints for ceilings using negative joints, and allowing ceiling planes at
different levels to slide along and across each other (refer to Figure 8.25). Movement gaps at
roof level can be accommodated by using a ‘first-principles’ approach. A solid cover is attached
on one side only, and then covered by some type of flexible metal flashing. Wall joints are
153
more difficult. While allowing for inwards and outwards movement in the plane of a wall is
154
Design aspects
8.8 A cantilever slab moat cover that can move into and away from a garden area, Tokyo,
Japan
8.9 A cantilever slab moat cover moving into and away from a garden on one side and
forming a step at a paved area on the other, Tokyo, Japan
Design aspects
155
156
Design aspects
8.10 A steel plate moat cover attached to the isolated building on the right and able to
slide over the retaining wall and paving to the left, Wellington, New Zealand
straightforward, displacements normal to the walls are far more difficult to achieve while
simultaneously coping with wind pressure and other weather-proofing requirements (Figure
8.26). One approach is to specify a complex proprietary system consisting of magnets,
springs and a hinge which would not be expected to be damaged even during extreme move-
ment (Figure 8.27). Alternatively, small movements may be allowed for by minimal movement
provision and by elastic bending of wall linings and some members designed to be sacrificial
in a large event, perhaps sacrificial steel work and frangible lining panels. This was the detailing
philosophy adopted for the connection between the buildings shown in Figure 8.28.
The challenge of designing movement joints that meet performance objectives
acceptable to the client should not be underestimated. It is likely that several iterations of
8.11 At a main entrance the cover plate is attached to the isolated superstructure on the
left and able to slide over the paving on the right. The fixings of the vertical steel plate to
the left of the handrail are designed not to impede the movement of the stair towards it,
San Francisco, USA
design development will be required. Designers should consider constructing small physical
models to both develop the design and help communicate it, possibly to the client, but
definitely to the contractor. Assume that a contractor has never had to allow for such large
seismic movements before.
Detailing for movement is necessary whenever a building element crosses an isolation plane.
Stairs often fall into this category. To prevent them reducing the effectiveness of the seismic
Design aspects
isolation system and to avoid damage to the stairs themselves during seismic movement,
separation is needed. Both the stair structure and handrails require separation (Figures 8.29
and 8.30).
Elevator shafts often cross isolation planes. The usual strategy is to fix the elevator
157
shaft to the isolated structure and suspend it so that the isolation plane passes beneath it.
158
Design aspects
8.12 A steel cover plate at a secondary entrance slides over the exterior concrete paving.
The chain linking the building to the handrail is considered sacrificial. It will break and be
replaced after an earthquake, Wanganui, New Zealand
Fixing Paving
Moat
Design aspects
Section
8.15 A detail of an angled sliding interface between a moat cover and paving to avoid the
159
8.16 45 degree sliding surfaces between concrete paving and a steel grating cover plate,
Wellington, New Zealand
8.17 Detailing of a movement joint between isolated and fixed paving that utilizes angled
sliding planes, Tokyo, Japan
8.18 A precast moat cover where movement parallel to the entrance is accommodated by
the nearer angled sliding joint. Movement at right angles occurs when the precast slabs
Design aspects
slide along the steel support rails that are attached to the isolated structure, Tokyo, Japan
161
162
Hinge
Design aspects
Movement gap
Centering bar
angled in plan
Hinge can slide
and rotate in plan Grout
Concrete anchor
Movement gap
Gap
8.19 Three types of cover plates. All plates allow horizontal movement in any direction
8.20 The underside of a self-centring floor cover plate. Diagonal bars that centre and
support the plate are visible as well as a flexible elastomeric wall joint, Wanganui, New
Zealand
Design aspects
163
164
Design aspects
8.21 A cover plate allowing a full range of movement is almost indistinguishable from the
surrounding area of paving. The plate is fixed to the building and due to its 45 degree
sliding surface will slide across and along the moat gap below, Tokyo, Japan
The elevator shaft needs to be braced so that it can resist horizontal inertia forces as it
cantilevers from the floor above (Figure 8.31). A cover plate over the clearance gap must be
provided at elevator doors and where otherwise visible.
If the isolation plane is located mid-storey then the partition walls of that storey must
be detailed for movement. As shown in Figure 8.32, movement along the isolation plane must
not be obstructed, so the walls require all their structural support from (usually) the floor
below. If large movements are to be accommodated without damaging the wall, it should be
physically separated from the floor above. Its strength against wind pressure and other
horizontal forces can be provided by cantilevering it from its base using specifically designed
vertical wall structure rigidly fixed to the floor beneath, or else using return walls to brace it.
The gap at the top may need to be filled with soft fire-proof material and covered by a scotia
which may be sacrificial. If some wall damage is acceptable the wall need not cantilever but
can be fixed normally at its base and connected to the floor soffit or beam above in a way
that allows sliding movement along its length.
8.22 A steel grill cover plate where drainage is integrated with a moat space, Tokyo,
Japan
Although the architect will not personally detail movement capacity of pipes and wiring
crossing an isolation plane, some extra provision of space to accommodate it may be necessary.
Design aspects
The principle is straightforward. Pipes need to be flexible or pin jointed to undergo the design
displacement without damage (Figures 8.33 to 8.35). Wiring should be provided with
sufficient slack or excess length. Because flexible or multi-pinned pipe joints are expensive,
cost savings are possible by connecting pipes together in the isolated superstructure in order
165
8.23 A visible surface (and wall) mounted interior cover plate, San Francisco, USA
Ceiling
hanger
Section
8.25 An example of a low-cost floor cover plate which will need repair following a
moderate earthquake. Movement at ceiling level is accommodated by a negative detail
Design aspects
167
168
Wall original
position
Design aspects
Gap
Range of wall
in-plane movement
Range in movement
normal to wall
8.26 Accommodating movement of two walls towards and away from each other is easily
achievable, but movement normal to them is more challenging
Aluminium frame
Caulking
Metal spring
Magnet
Steel plate
8.27 A sophisticated wall movement joint that allows the gap width to change, movement
along the plate and the two walls to move relative to each other at right angles to the wall
panel
Design aspects
169
170
Design aspects
8.28 The vertical black painted thin metal covering is part of the wall movement joint
between the seismically isolated building to the left and the fixed-base building to the
right. The choice of materials and colour achieve an elegant transition. Sacrificial steel
elements and frangible interior linings will need replacement after a moderate to large
earthquake, Wellington, New Zealand
8.29 The stair stringer and balusters are cut to allow movement in all horizontal directions.
The handrail is cut (not shown) and detailed in a similar manner, Wellington, New Zealand
Design aspects
171
172
Design aspects
8.30 An external secondary stair. The stair landing cantilevers from the isolated building to
the right and slides on the supporting pier. The handrails are separated to allow
movement, Lower Hutt, New Zealand
Elevator shaft
Gap Gap
(a)
Elevator shaft
First floor
Isolation plane
Door opening
Partition wall
Cover plate
Ground floor
Gap
Gap Gap
Design aspects
(b)
8.31 Sections through an elevator shaft showing how it hangs from the superstructure and
is braced back to it. It is separated from any other elements that might prevent its
movement with the isolated superstructure. In (a) the isolation plane is at the base of the
173
Movement of
floor above
Design aspects
Gap
Scotia
Bearing Scotia not
shown Wall
Clearance
Elevation Section
8.32 The need to provide movement between walls and upper floor where the isolation
plane is between floors. If a wall is self-supporting normal to its length, then scotias must
be weakly attached to allow movement and be reinstated after an earthquake. If the wall
relies on its connection to the structure above for its stability, the scotias must provide that
restraint
Floor
Hanger
Pipe Pipe
Basement floor
Ball and
Flexible pipe socket joints
Gap
(a) (b)
8.33 Pipework needs flexibility (a), or several short pin-jointed lengths to allow damage-
free seismic displacement (b)
8.34 Movement in a pipe connecting an isolated and fixed-base building is achieved by
several pipe lengths with ball and socket joints that allow rotation between pipes, San
Francisco, USA
Design aspects
175
176
Design aspects
1 ASCE 7 2009: Minimum design loads for buildings and other structures, American Society of Civil Engineers,
Reston VA., Chapter 17 Seismic design requirements for seismically isolated structures.
2 Booth, E., 2014, Earthquake design practice for buildings, 3rd edn, London: ICE Publishing, p. 312.
3 Pietra, D., Pampanin, S., Mayes, R.L. et al., 2014, ‘Design of base-isolated buildings: an overview of international
codes’, Proceedings of the New Zealand Society for Earthquake Engineering Conference, Wellington, paper O47.
4 Martelli, A., Clemente, P., De Stefano, A., Forni, M. et al., 2013, ‘Development and application of seismic
isolation energy dissipation and other vibration control techniques in Italy for the protection of civil structures,
cultural heritage and industrial plants’, Proceedings of the 13th World Conference on Seismic Isolation, Energy
Dissipation and Active Vibration Control of Structures, Sendai, p. 1.
5 Becker, T.C., Furukawa, S., Mahin, S.A. & Nakashima, M., 2010, ‘Comparison of US and Japanese codes and
practices for seismically isolated buildings’, Proceedings of the 2010 ASCE Structures Congress, pp. 2330–8,
p. 2331.
6 Pan, P., Zamfirescu, D., Nakashima, M. et al., 2005, ‘Base-isolation design practice in Japan: introduction to
the post-Kobe approach’, Journal of Earthquake Engineering, vol. 9, no. 1, pp. 147–71, p. 162.
7 Feng, D., Miyama, T., Liu, W. & Chan, T., 2012, ‘A new design procedure for seismically isolated buildings
based on seismic isolation codes worldwide’, Proceedings of the 15th World Conference on Earthquake
Engineering, Lisbon.
8 Cutfield, M.R., Ma, Q.T. & Ryan, K.L., 2014, ‘Cost-benefit analysis of base isolated and conventional buildings:
a case study’, Proceedings of the New Zealand Society for Earthquake Engineering Conference, Wellington,
paper O46, p. 7.
9 ASCE 7 2009: Minimum design loads for buildings and other structures, American Society of Civil Engineers,
Reston VA., Chapter 17 Seismic design requirements for seismically isolated structures.
10 Ozdemir, G. & Akyuz, U., 2010, ‘Response of isolated RC buildings under bi-directional near-fault ground
motions’, Proceedings of the 9th US National and 10th Canadian Conference on Earthquake Engineering,
Toronto, paper 460.
11 Kircher, C.A., 2006, ‘Chapter 11 seismically isolated structures’, in Building Seismic Safety Council, NEHRP
recommended provisions: design examples - FEMA 451, National Institute of Building Sciences, Washington,
D.C., p. 11–10.
12 How to plan and implement seismic isolation for buildings, 2013, The Japan Society of Seismic Isolation, Tokyo:
Ohmsha, p. 46.
13 Saiki, K., Kitamura, Y., & Kani, N., 2013 ‘Damage of expansion-joints for seismically isolated buildings and
countermeasures’, Proceedings of the 13th World Conference on Seismic Isolation, Energy Dissipation and
Active Vibration Control of Structures, Sendai.
14 ASCE 7 2009: Minimum design loads for buildings and other structures, American Society of Civil Engineers,
Design aspects
Reston VA., Proposed 2014 changes to ‘Chapter 17 Seismic design requirements for seismically isolated
structures’, C 17.2.6.
15 Ko, E., Morgan, T., Bello, M., Bailey, R. et al., 2010, ‘Base isolated structure – the new San Francisco general
hospital & trauma center’, Proceedings of the 9th US National and 10th Canadian Conference on Earthquake
177
Introduction
I N the usually cost-driven construction sector, it is particularly important for cost differences
associated with seismic isolation to be defined. Unfortunately, as we explain below, it is
impossible to give readers an appreciation of how the cost of a seismically isolated building
differs from that of a conventional building with a high degree of accuracy due to the large
number of factors at play. However, first we must check we are comparing like with like.
Comparing the construction costs of these two types of buildings is like comparing the cost
of an ordinary motor vehicle to one with advanced safety features. One vehicle has an im-
proved braking system to avoid damage in the first place, and then, in the event of a crash,
energy-absorbing crush zones in the chassis and air bags to increase protection. There is no
doubt about the improved safety performance for which we should expect an increased price.
Few studies have compared the initial costs of seismically isolated buildings to those of
conventional buildings with the same seismic performance. Unfortunately, special buildings,
often housing essential facilities, rather than typical buildings are those that are reported upon.
Furthermore, the studies don’t capture the reduction in downtime and other benefits seismic
isolation provides. Nevertheless, Ron Mayes reports how the construction cost of a nuclear
power plant was 2 per cent less expensive, and a fire command and control facility 6 per cent
cheaper with seismic isolation.1 Seismic isolation led to lower costs to brace piping and
electrical and mechanical equipment which lowered the overall cost. In a more recent example,
Ko reports that a major new seismically isolated hospital was cheaper and easier to build than
a conventional hospital building. Because the code places limits on the inter-storey drifts of
hospitals, or in other words, how much further one storey moves horizontally compared to
the storey below, the savings of steel framing more than offset the costs of isolation.2 Other
advantages he mentions are the reduced floor accelerations and therefore less damage to
contents, and shallower beams and smaller columns that improve space planning.
Other reasons for a lack of definitive seismic isolation cost information is that
construction economics vary from country to country, and that most seismically isolated
buildings are different. Perhaps site seismicity or soil conditions vary, or the isolation plane
is in a different position. In spite of all these differing factors some boundaries can be drawn
around cost variations between conventional and isolated buildings. Drawing from eight case-
studies mainly in the USA, Mayes reports that the cost of construction increases up to 5 per
cent and savings of up to 3 per cent where using seismic isolation.3 Japanese experience is
summarized: ‘Generally, for a building with less than about ten stories, the initial construction
cost is several per cent higher than for the building without isolation, but for structures more
than ten stories, there is almost no difference in construction cost.’ 4 This 2013 information
suggests there has been a reduction in seismic isolation costs since 1999. Then the cost
premium for seismic isolation was estimated at 7.6 per cent on the basis of survey responses
of 25 Japanese designers and 24 owners and researchers.5 It is likely that the costs of seismic
isolation will continue to fall as more bearing manufacturers enter the market and regulatory
requirements are rationalized. In New Zealand, a study of four isolated hospitals that were
completed between 2005 and 2013 revealed that the total additional cost of seismic isolation
was 3 per cent of their construction costs.6 Since hospitals are heavily serviced and therefore
cost more per square metre, the additional isolation costs for most other building types
can be expected to be slightly greater. In a European example involving the fast-track build
of 4500 apartments after the damaging 2009 L’Aquila earthquake, Calvi reports that seismic
isolation represented 2 per cent of the total cost.7 We stress once again that these cost com-
parisons neglect completely the potentially enormous cost benefits that seismic isolation
delivers after a major earthquake by reducing injuries and downtime.
I F there are additional costs of protecting a building and its contents by seismic isolation,
how do they arise? Almost all of the additional costs are incurred in the vicinity of the
isolation plane. This means the cost of isolation per square metre of construction is reduced
both by reducing the ground floor plan area and increasing the numbers of storeys. Additional
costs arise from a possible additional suspended floor, isolation devices such as bearings,
provision of moats or rattle-space including retaining walls, and moat covers which are
usually required unless the isolation plane is above or at ground level. Movement joints
between adjacent buildings add to the costs, as well as detailing of flexible electrical and other
Economics of seismic isolation
services, stairs and elevators that cross the isolation plane, and increased design and peer
review fees. To these costs could be added the loss of income from being unable to develop
as much of the site area as normal. This is due to wider-than-normal movement gaps above
the isolation plane adjacent to neighbouring buildings and site boundaries, and is rarely
acknowledged.
Of all these costs, the most significant is the cost of an additional suspended floor
immediately above the isolation plane. Ryan and others highlight this in their breakdown of
179
the additional costs of seismic isolation for a three-storey steel framed US building (Table 9.1).8
Table 9.1 Detailed additional seismic isolation construction costs
180
Excavation 8
Retaining wall and moat cover 11
Isolator pedestals 1
Isolation devices 27
Level 1 floor and framing 36
Flexible connections 6
Crawl space drainage and lighting 6
Suspended elevator shafts 5
In the study of the four New Zealand seismically isolated hospitals mentioned previously, the
costs of isolators ranged from 1–2 per cent of the construction cost including the cost of
isolator prototype tests, and all other costs, including the rattle space and retaining walls lay
within the same range.9 In all cases, basements at least under part of the ground floors of the
hospitals are used for mechanical services or car parking.
The additional costs of seismic isolation discussed above are inevitable. Of course, the
percentage contribution each item adds to the final cost varies from project to project.
However, as already mentioned, there are examples where seismic isolation is cost-neutral or
even cheaper than conventional construction. These situations occur most commonly in
essential facilities such as hospitals, civil defence offices and fire stations – operations that must
function post-earthquake without interruption. The conventional approach to the design of
these facilities is to increase their structural strength and stiffness. At least in New Zealand,
such buildings are almost twice as strong as usual, their structural members significantly larger
and more expensive. Since seismic isolation may meet and exceed the required seismic
performance without the need of larger structure, some cost savings are expected.
Even for seismically isolated buildings with normal occupancy, such as apartment or
office buildings, some structural savings will partially off-set the additional costs of seismic
isolation. Structural components, such as beams, are not required to withstand seismic
overload and absorb earthquake energy by yielding of either structural steel sections or re-
inforcing steel embedded in concrete. Columns may be able to be reduced in strength and
structural detailing reduced and simplified. This means fewer reinforcing ties in reinforced
concrete columns and beams, or simpler beam to column connections in structural steel
frames, speeding up construction and lowering cost.
Cost savings other than to primary structural framing are likely as well. In isolated
buildings, mechanical and electrical plant and architectural elements, such as suspended
ceilings, require less bracing in order to prevent overturning, sliding or just resist their own
earthquake inertia forces. Also, because horizontal superstructure deflections are reduced by
seismic isolation, savings arise from simpler and smaller movement details. According to a
structural engineer of a new office building in Christchurch, seismic isolation ‘had a significant
benefit when we were detailing the cladding system and attachments for the curved façades.
If we had used a conventional structure design, then the inter-storey drifts would have made
the curved glass façades cost-prohibitive’(Figure 9.1).10
Pre-earthquake economics
Economics of seismic isolation
E VEN before the benefits from seismic isolation are realized after an earthquake, seismic
isolation can positively affect a building owner’s income. For example, Japanese apartment
dwellers are prepared to pay higher rentals for the security and peace of mind resulting from
superior seismic performance. The other possible on-going financial advantage of seismic
isolation is earthquake insurance. Not only is it easier to obtain insurance cover in a market
insurers may be withdrawing from, such as post-earthquake Christchurch and other higher
risk regions, but significant cost savings are also possible. Armed with an engineering report
of his or her isolated building, an owner might succeed in negotiating a significant reduction
to the annual earthquake insurance premium. Alternatively, an owner can self-insure. It is
most unlikely the damage to an isolated building would exceed a typical insurance deductible
of approximately 5 per cent of the sum insured (building and contents). The deductible and
savings on annual premiums will more than pay for any additional cost of seismic isolation.
Earthquake insurance premiums fluctuate considerably. They depend upon many factors
including the seismicity of a site and the state of the insurance industry at that particular
time, but when they are of the order of 1 per cent of the sum insured, self-insurance becomes
very attractive. One significant impediment to self-insurance is the typical requirement of
financial lenders, such as banks, to require full insurance cover, but there is at least one US
precedent of a lender agreeing to waive earthquake insurance on the basis of the enhanced
seismic performance provided by seismic isolation.11
Post-earthquake economics
T HE most significant economic benefits of seismic isolation occur during and immediately
after a major earthquake. Occupants’ personal losses are reduced, together with costs
arising from potential deaths and injuries. Provided that the building is not damaged by
adjacent buildings, is accessible and inhabitable due to continued supply of water, sewerage
and energy, then losses of rent, disruption of production and loss of market share are
minimized to the greatest possible extent. Although insurance companies offer business
disruption insurance, seismic isolation provides by far the most elegant solution to the
problem, through avoidance.
Owners of conventional buildings must assess these possible losses for themselves. They
are often huge and difficult to predict accurately. For example, how much time might be
spent accessing, planning and repairing a building in the aftermath of a major earthquake?
Given the extent of post-earthquake community disruption, very little works as normal.
Everything is in a state of flux, including local authorities. In Cutfield and Ma’s study, business
interruption or downtime cost was estimated at five to ten times the cost of building damage,
unless the owner was able to re-locate quickly and efficiently.12 Based on his experience
following the 2010–11 Canterbury earthquake sequence, which caused so much destruction
in Christchurch, David Whittaker calculates the cost of business disruption as the rental
income lost for 3 years – equivalent to losing 30 per cent of the building value.13 For high-
tech manufacturers or research establishments, far higher levels of business disruption losses
can be imagined, possibly leading to businesses failing.
Life-cycle analyses
W E all know the pitfalls associated with choices made solely on the basis of first or initial
costs. The cheapest cladding system may look fine during installation, but what’s it
like five or ten years later? Life-cycle analyses are particularly useful and revealing for simple
and relatively quantifiable situations like this. But when a client asks about the life-cycle costs
or savings from seismic isolation, the calculations are not straight forward. Even if
sophisticated Net Present Value methods are adopted, the answers depend upon fundamental
assumptions that are subject to large variability. However, all published life-cycle analyses for
buildings located in high seismic hazard areas of western North America and New Zealand
consistently demonstrate the cash flow benefits of seismic isolation.14, 15, 16, 17 For example,
Terzic and others calculate a minimum 3 per cent return on investment over a 50-year time
frame.18 They exclude potential insurance savings and acknowledge even more attractive
returns for seismic isolation if downtime estimates are refined.
Most life-cycle analyses neglect the costs and benefits of earthquake insurance. However,
in his simplified analysis based upon annualised costs, Whittaker makes a compelling case
for owners of seismically isolated buildings to self-insure.19 He shows that the annual costs
of earthquake damage plus business interruption, insurance premium assuming no reduction
for the isolated building, and annualised deductible are similar for owners of conventional
and seismically isolated buildings. The spike in insurance premiums immediately after the
Canterbury earthquake has reduced dramatically from approximately 1 per cent to 0.25 per
cent. But an update of Whittaker’s spreadsheet shows that the annualised cost of an uninsured
(self-insured) isolated building is only 7 per cent of an insured conventional building.
Economics of seismic isolation
Notes
1 Mayes, R.L., 1990, ‘The economics of seismic isolation in buildings’, Earthquake Spectra, vol. 6, no. 2, pp.
245–63.
2 Ko, E., Morgan, T., Bello, M., Bailey, R. et al., 2010, ‘Base isolated structure – the new San Francisco general
hospital & trauma center’, Proceedings of the 9th US National and 10th Canadian Conference on Earthquake
183
3 Mayes, R.L., 1990, ‘The economics of seismic isolation in buildings’, Earthquake Spectra, vol. 6, no. 2, p. 252.
4 The Japan Society of Seismic Isolation (ed.) 2013, How to plan and implement seismic isolation for buildings,
Economics of seismic isolation
Introduction
Construction
S INCE most contractors haven’t built a seismically isolated building before, its designers
need to explain how the isolation system works, and its construction implications.
Although some of this communication can be achieved through contract documentation,
Construction and maintenance
there is no substitute for designers requesting special meetings to outline the unique aspects
of the isolation system. During the construction of an isolated building, the contractor has
to act against his or her instincts. A contractor is used to fixing or joining elements together
rather than letting them slide, and is used to filling gaps rather than leaving them open. Unless
explained clearly and perhaps repeatedly, habitual ways of building, like anything else in life,
are difficult to change. For this reason thorough construction inspections are recommended,
185
lated buildings they inspected after the 2011 Tōhoku earthquake suffered damage to moat
covers and other components of the isolation systems.2 Some of this damage was due to faulty
or careless construction. During post-earthquake inspections of the Christchurch Women’s
Hospital in 2010 and 2011, buckling damage to a length of seismic moat cover at ground
level, and damage to an air bridge connecting the hospital to another building was evidence
of how the effectiveness of the isolation system was reduced.3, 4
Another aspect of seismic isolation that is new to most contractors is the procurement
process of the isolation devices. Early ordering of isolation devices to avoid installation delays
and the need for temporary inserts is highly recommended. Between three to six months
should be allowed for device manufacture, excluding transportation time. The client may
wish to order and purchase devices directly and have them delivered to site in order to avoid
the contractor’s margin. The structural engineer plays important roles during procurement,
depending on which procurement strategy is adopted. Does the engineer specify a comply-
ing system, specify performance requirements with device vendors undertaking the design,
or is a complying system specified along with its performance requirements that can then be
matched by other devices?5 The final decision has to balance the need for optimal design,
to include all potential bidders, and finally, to reduce the difficulty and effort of checking
large numbers of analyses as part of the bid evaluation process.
During the final weeks of construction, work should commence on communicating to
all relevant people the fact that a building is seismically isolated. Permanent notices warning
against placing objects that might prevent unobstructed movement between isolated and non-
isolated structure need to be posted. One notice should be placed at the entrance to the crawl
space (if provided), and other notices placed around the building perimeter in an attempt to
prevent any obstructions reducing the effectiveness of the seismic movement gap. Also at this
time, a Building Seismic Isolation Manual should be prepared jointly by the structural engineer
and architect. Written for the building owner, its purpose is to describe the isolation system,
explain how it is expected to perform, and document all the seismic isolation details. Since
the manual needs to be referred to during regular maintenance inspections, it should include
design details and as-built drawings as necessary. All structural and non-structural isolation
details, such as movement gap details, should be included to help future inspectors with no
first-hand knowledge of the building.
Maintenance
P HYSICAL systems deteriorate with age. The oxygen and water that sustain us oxidize and
corrode construction materials, including those of seismic isolation systems. The thin
outer layer of a rubber bearing will suffer a loss of elasticity, while the surfaces of sliding
joints are prone to sticking, or even locking up in the presence of moisture or surface debris.
However, seismic isolation systems face a possibly more severe danger during their lifetimes
– inadvertent human interference with seismic movement gaps. A gap filled in over just part
of its length could cause severe structural damage to an isolated building during an earthquake.
If the isolated structure can’t move freely within its enclosing retaining walls then it is no
longer isolated. It will be subject to inertia forces far larger than those designed for.
A maintenance programme therefore needs to address both the natural and human
threats to which seismic isolation is vulnerable. The 2012 Japanese Society of Seismic Isola-
tion maintenance standard specifies the frequency and nature of inspections of seismic
isolation devices, such as bearings, sliders and dampers; the seismic isolation plane including
the perimeter details of seismic clearance gaps and movement joints; and the flexible services
pipes and wiring entering the building.6 The standard also defines five types of inspection:
After the 2011 Tōhoku earthquake the lead dampers in at least one building cracked and
needed replacement.7
Personal experience of one of the authors of this book has also shown the need for this
type of maintenance regime. Before the mistake was rectified, an isolated building being
refurbished and subject to landscaping changes was observed to be strongly tethered to the
surrounding ground by horizontal steel bars. The bars supported a fence that was part of a
landscaping project abutting the isolated building.
Kelly and others cite a similar but more serious example: Construction and maintenance
At the LA County Fire Command Center, the contractor had poured a reinforced
concrete slab under the floor tiles at the main entrance to the buildings, preventing
free movement in the E-W direction. Apparently the reinforcing was added after
the contractor had replaced the tiles several times after minor earthquakes and
did not realize that this separation was designed to occur.8
Examples of the importance of periodic inspections include the need to replace the rubber
187
bearings of what some regard as the first base-isolated building in the world. The 40-year-
188
old rubber block bearings (without the benefit of the steel confining plates of modern
bearings) had cracked and deformed.9 Twenty-four years after their installation, the lead-
Construction and maintenance
extrusion dampers at the Wellington Central Police station required maintenance. Tests
showed that an undesirable bond between the lead and the steel cylinders had formed.
The interface between the materials needed re-lubricating to maintain the safety of the
building.10
The maintenance programme for a seismically isolated building should ideally be
established and agreed to by the client before construction is complete. We believe that the
programme should be as legally binding as those for elevators and fire alarm systems that are
regularly inspected and maintained. In New Zealand, most non-residential buildings require
an annual Building Warrant of Fitness. It consists of a statement supplied by the owner to
the local council stating that the building systems listed in a compliance schedule have been
checked and maintained in accordance with the schedule for the previous 12 months. The
compliance schedule, which should include the seismic isolation system, is agreed to by the
council at the time the building owner applies for Building Consent to begin construction.11
If other countries do not have such a system, it is the responsibility of the structural engineer
to set up a fail-safe inspection and maintenance system. It needs to outlive the structural
engineer and the existence of his or her firm for the entire duration of the life of the building.
Anything less cannot guarantee the effectiveness and safety of a seismic isolation system.
Notes
1 The Japan Society of Seismic Isolation (ed.) 2013, How to plan and implement seismic isolation for buildings,
Tokyo: Ohmsha, p. 62.
2 Kani, N., Ogino, N., Kitamura, Y. & Kitamura, H., 2012, ‘Effects of seismically isolated buildings during the
huge 2011 earthquake in Japan’, Proceedings of the 15th World Conference on Earthquake Engineering, Lisbon.
p. 1.
3 Gavin, H.P. & Wilkinson, G., 2010, ‘Preliminary observations of the effects of the 2010 Darfield earthquake
on the base-isolated Christchurch Women’s Hospital’, Bulletin of the New Zealand Society for Earthquake
Engineering, vol. 43, no. 4, pp. 360–7.
4 Kuang, A., Sridhar, A., Gavin, H. & Gutschmidt, S., 2013, ‘Analysis of the seismic response of the Christchurch
Women’s Hospital’, Proceedings of the 2013 New Zealand Society for Earthquake Engineering Conference, p. 1.
5 Kelly, T.E., Skinner, R.I. & Robinson, W.H., 2010, Seismic isolation for designers and structural engineers,
Kanpur: NICEE, p. 115.
6 Kani, N., 2013, ‘Maintenance standards for seismic-isolation buildings’, Proceedings of the 13th World
Conference on Seismic Isolation, Energy Dissipation and Active Vibration Control of Structures, Sendai,
Japan, p. 2.
7 Motosaka, M. & Mitsuji, K., 2012, ‘Building damage during the 2011 Pacific Coast Tohoku earthquake’,
The Japanese Geotechnical Society Soils and Foundations, vol. 52, no. 5, pp. 929–44, p. 931.
8 Kelly, T.E., Skinner, R.I. & Robinson, W.H., 2010, Seismic isolation for designers and structural engineers,
Kanpur: NICEE, p. 327.
9 Gjorgjieiv, I. & Garevski, M., 2012, ‘Replacement of the old rubber bearings of the first base isolated building
in the world’, Proceedings of the 15th World Conference on Earthquake Engineering, Lisbon.
10 Smart, C.R., 2013, ‘Wellington Central Police Station base isolation maintenance’, Bulletin of the New Zealand
Society for Earthquake Engineering, vol. 46, no. 3, pp. 141–56.
11 Wellington City Council 2005, Building warrant of fitness, viewed 15 December 2014, http://wellington.
govt.nz/services/consents-and-licences/building-consents/building-warrant-of-fitness
Conclusions
O VER the last five years hundreds of seismically isolated buildings have performed very
well in damaging earthquakes. This has confirmed the results of thousands of computer
analyses and numerous full-scale physical earthquake simulations of isolated buildings and
isolation devices. As a consequence of positive comparable results from field performance,
laboratory testing and computer modelling, we now can have unprecedented confidence in
the improved performance seismic isolation can provide.
T HE potential benefits of seismic isolation are enormous. The inclusion of the word ‘poten-
tial’ acknowledges the likelihood, in the range of 80–90 per cent, that the design earth-
quake will not occur during the life of a building. However, during and after the design
earthquake striking an isolated building, the building owner will benefit in many different ways
from the protection provided by seismic isolation. It is not too dramatic to say that lives, jobs
and businesses could be saved. And even before the ‘big one’ strikes, if ever, financial, aesthetic
and functional benefits arise from seismic isolation.
T HE case studies in Chapter 6 illustrate how seismic isolation can offer architects
opportunities to achieve and reinforce certain design concepts and architectural qualities.
As designers, and especially structural engineers, become even more confident in the
technology, seismic isolation may encourage structural forms in seismic zones that are more
irregular than those recommended for conventional construction. In spite of this freedom,
the discipline of careful architectural detailing, particularly around movement gaps, needs
constant attention.
Just as personal human freedom has its limits, so too does seismic isolation. Some seismo-
logical and geological conditions, and urban building sites are unsuitable for seismic isolation.
Uncertainties
S EISMIC isolation is a relatively new technology. Its 35 years of age represents one third of
the era of modern earthquake engineering. As such, it is not surprising some technical
aspects, such as the width of moat gaps, are still being debated. Although the science of seismic
isolation is now well understood, some uncertainties linger related to the nature of and the
characteristics of earthquake shaking. We are designing for a most unpredictable hazard. Also,
post-construction uncertainties, such as how environmental conditions affect isolation devices
and how building use, like the inadvertent blocking of movement gaps, might negate seismic
isolation systems need to be reduced, or better, eliminated, by rigorous inspection and
maintenance regimes.
Looking ahead
T HIS book presents a state-of-the-art summary of the theory and practice of seismic
191
isolation from an architectural perspective. Given its sole emphasis on recent and current
192
practice, it is now appropriate to consider the future of seismic isolation. How might it
develop in the coming years? Will it become more widely adopted and what might be the
Conclusions
National Museum of Western Art 130, 134–7 seismic isolation: 3–D 74–6, 77, 79; architectural form
Nicholas G. Hayek Center 102–6 84–113; benefits 65–72, 190; confidence in 190;
Index